QWERTY, path dependency, Microsoft, linux and oh yeah, evolution.

[John Bianco]

.
> Maybe linux is to Microsoft what the Japanese car makers
>were to Detroit.

As a side note, the Japanese did not just capture a big chunk by soley
makeing high quality cars. The Japanese had ultra agressive and predatory
business practices that untill the mid-late 80s often sold their cars at a
loss to the US market to get market share. Most Japanese business' also
shared this practive of being very agressive, while often outright stealing
US and European companies inventions.

People now currently talk about Microsofts and Intels business'
practices, on how they do not play fair, and how they try, and often suceed
in putting their competitors out of business. Well, it seems that they have
indeed learned well from the Japanese.

The newly evolved species, able to perfect it's
>superiority in an isolated little environment till it can come forth.
>A few years ago, when I told people I used linux I got blank stares.
>Now, very often, people have at least heard of it.
> I actually wonder if linux (and/or Free-Net-OpenBSD and/or HURD, sorry
>guys for not mentioning you sooner) might not be a new economic
>phenonemon made possible only in this information age and a reaction
>to the hobbles of 'intellectual property'? In the past, if somebody
>came up say with a cool way to soup up his hot rod, he could tell his
>friends, maybe visit their garages and help them tinker with their
>engines, or publish his idea in a magazine, but he couldn't just give
>his souped up engines to people for free. You can do that with
>software though, and it is done, often.
>
>------remove "UhUh" and "Spam" to get my real email address---

[cpw...@spamrahul.net]

<NOTE, a much shorter version of this article, with the same
subject, is being posted to sci.bio.evolution>

I started reading sci.econ and sci.bio.evolution a little
while ago because I had a thought and wanted to post something about
it, but, I thought I should lurk awhile first and I'm glad I did. The
thought sprang from reading an article in the New Yorker while
in a waiting room that mentioned Microsoft and a paper written by
Brian Arthur about how a company can 'lock in' a market. It struck a
chord with me and it also struck me that there was a similarity to a
biological phenomenon I had read about in a book by Stephen J. Gould,
"Wonderful Life" which talked about the Cambrian explosion and how
certain phyla of life had mysteriously become extinct while others, which
in that time, seemed no more likely to flourish, such as the
progenitors of the insects, had in fact done so. There was
speculation that some temporary advantage at a critical moment had
become permanent and I thought aha, there was some quality of systems,
whether biological or economic, at work. Someone else was bound to
have noticed and I wanted to ask who.
Well, reading about QWERTY and Microsoft in here has given me
pause. Apparently there's not so much evidence to support the
contention of what's being called 'path dependence' as I thought, and
the QWERTY phenomenon, which I'd seen mentioned many times and
believed, is an urban legend. Dvorak and other style keyboards had a
fair trial against QWERTY and were not shown to make typists much more
productive. Other examples of supposed inferior products locking out
better products, like VHS triumphing over Beta, also don't bear up to
close inspection. I think there's a strong force driving these
legends though, based upon peoples experiences and perceptions of the
market.
Back in the 50s and 60s people complained about Detroit's
"we build 'em and you buy 'em" attitude towards customers. Finally
Detroit faced competition from foreigners. Good thing the world was a
big enough place that competition could develop elsewhere. (Again, I
find myself harking back to a meme from biological evolution, the
hypothesis that evolution happens in isolated communities where the
mutations aren't swamped by the general population. Japan, in a
sense, was an isolated community able to work out it's different
approaches to manufacturing).
Of course, the Big 3 or Big 4 of Detroit were not a monopoly,
they were an oligopoly.
I'm not hear to advocate anything, or to try to persuade or
change peoples minds (I know better than that, I've been reading
Usenet a long time). My interest is in a socratic argument to try to
find the truth, really. But I'll be up front about the fact that I
have very strong prejudices, and I admit they are prejudices. I'm a
computer programmer who has been in the biz since 1978, and been
working with Unix since 1982, and I hate it when I have to come near
Microsoft stuff. But it's a prejudice because I honestly haven't used
Microsoft stuff enough to give it a fair trial. (What am I gonna do?
Go out and buy their stuff just to see if I like it? I got Windows
3.1 shoved down my throat twice when I bought PCs, putting money in
Bill Gates pocket for something I didn't want.)
I was working under contract at a small outfit in the summer
of '96. This company started out making their application for use
under X-windows, but their plan was to move everything to windows or
windows NT. I mentioned to the owner/president/entrepreneur one time
that I hoped to be able to retire before Microsoft took over the
world. He said I had about 2 years (maybe it was 3, I don't
remember). All of his clients wanted Windows NT. I had never met his
clients, but I've got my prejudices to fall back on, I imagined them
as Dilbert style managers, nervous nellies whose thought process went
something like "Microsoft big, Microsoft must be good" but that's path
dependence, isn't it?
So I have prejudices, and ya'll are invited to confirm those
prejudices or help show me the true nature of things. But I'm not the
only one with these prejudices. If you look in the advocacy groups,
(I occasionally subscribe for awhile to comp.os.linux.advocacy, then
get out after awhile, the repetitiveness and rhetoric get to be too
much, but I'm resubscribing now for awhile to chase followups to
this.) In comp.os.linux.advocacy you'll see variations on a theme,
("Linux is great, NT stinks, I know I've used both", "No, NT is great,
Linux stinks, I know, I've used both). Actually, since I only ever
subscribe to the linux advocacy group, the response is more like, "NT
is so reliable, and the support is so good, and if you've had training
it can too be flexible, etc." Usually the NT advocates will admit
that Linux makes more efficient use of computer resources than NT
though.
Linux folks reassure themselves that Gates can't drive them
out of business like he's driven so much competition out of business
because it's not a business. Linux is free and anyone can get the
code and support it. (Is it a prejudice or unfounded urban legend
that Microsoft has driven worthy companies out of business? An
economist would say, cite examples, and some people have cited
examples in the newsgroups but I don't know how scientific the info
is) Some linuxers worry that Gates will find SOME way to destroy
linux, that he'll concoct a claim that linux is violating copyrights
or something. I'm not sure how paranoid the average Linuxer is
though, that may just be the fringe element talking. Isn't Andy Grove
at Intel supposed to be a big believer in paranoia?
Now there's a familiar tone to some of this. I mentioned the
perception of the automobile industry earlier. There's also IBM.
From what I understand, IBM was the biggest name in pre-computer style
business machines, sorters, card-readers, that kind of thing. It was
led by Thomas J. Watson, a man with a background in selling.
(Whenever I worked for a really small company, the head and founder of
the company, even if he had a technical background, always regarded
himself as a salesman more than anything else.) And IBM was supposed
to have a really topnotch sales force. Men with WASP type backgrounds
who would be reassuring to the corporate managers they were selling
to. Watson planned to have his son follow him in the business and the
son happened to see some of these new fangled computer things in
action and told the old man IBM needed to get into that business. His
son was one person Watson senior would listen to so he said OK go
ahead. IBM built computers that used the same kinds of punch cards as
their existing equipment and they had that great sales force so they
soon dominated the industry.
The guys who actually invented the computer, Eckert and
Mauchly, had a lot of bad luck. For instance, just as the New World
got named after Amerigo Vespucci instead of Christopher Columbus, so
computer scientists speak of the "Von Neumann" model computer instead
of the "Eckert-Mauchly" computer. But the troubles that happened to
Eckert and Mauchly might be considered Brian Arthur style historical
small events, there's that notion of path dependence coming up again.
My historical notions come from the book "Bit By Bit, An Illustrated
History of Computers" by Stan Augarten, Ticknor & Fields, New York,
1984.
I didn't work with IBM computers in the old days (before they
got into Unix type systems) either, but I spoke to a lot of people who
had. From what they said, everything else was better but "nobody ever
got fired for buying IBM", those damned Dilbert-style no-nothing
managers again. Then there's this other super-successful giant, Intel.
I was an assembly language programmer for years and people hated the
instruction set of the 8x86 style computers, but they were forced down
our throats, that's what the customer wanted. Presumably because IBM
used them in their PCs, IBM who just sort of decided to dabble in PCs
and then all the clueless layman public rushed out to buy them just
because they had IBM's name on them.
So that's some of the lore, and I've deliberately tried to
relay it as lore rather than make a strong argumentative case for it
because this is about the background of the prejudices, and I want to
be clear that I'm not calling them anything but prejudices, even if
they are prejudices which I most sincerely share in, against
Microsoft. People like me can't understand how these miserable
companies can be so successful and have been searching for an
explanation. So we figure they are big because they came on the scene
early, just like we thought the QWERTY keyboard succeeded when it shouldn't
havem, and I was ready to embrace this new economic theory of
"path dependency". OK, if these explanations are no good, then what
does explain it?
Are there compelling reasons that us poor old engineers just
don't understand, mysterious marketing and business advantages that
Microsoft exploits, and that IBM exploited before them, that make theirs
truly the superior products. "Ease of use" for instance. Linux is supposed
to be hard to install and learn compared to Windows95 or Windows NT. If
that's so, then I think a lot of it is due to "path dependency". When
somebody builds a new scanner, or video card or sound card, they as a
matter of course, write a software driver for Microsoft to go with it
and include the diskette in the package. They don't do that for
linux. Somebody has to find the specs (if they can, sometimes these
are considered proprietary secrets) and write a driver for linux
themselves. That's not innate superiority of Microsoft, that's just
being there first and already established, path dependency.
If somebody came to me 5 years ago and said, "I'm gonna bet
the farm on this new company I'm starting, what kind of computer
system should I use?" I would probably have said, go with Microsoft
and intel style PCs, anything else is too risky if you're betting the
farm, much as I would have hated mouthing those words. But isn't that
a manifestation of this discredited path dependence eonomic theory?
Now, (Oh Joy!) I might, with a clear conscience, convince the person to go
with linux.

Maybe linux is to Microsoft what the Japanese car makers

were to Detroit. The newly evolved species, able to perfect it's

[Andrew Pimlott]

On 23 Feb 1998 02:21:14 GMT, cpw...@Spamrahul.net <cpw...@Spamrahul.net>
wrote:

[snip]

>thought sprang from reading an article in the New Yorker while in a waiting
>room that mentioned Microsoft and a paper written by Brian Arthur about how
>a company can 'lock in' a market. It struck a chord with me and it also
>struck me that there was a similarity to a biological phenomenon I had read
>about in a book by Stephen J. Gould, "Wonderful Life" which talked about
>the Cambrian explosion and how certain phyla of life had mysteriously
>become extinct while others, which in that time, seemed no more likely to
>flourish, such as the progenitors of the insects, had in fact done so.
>There was speculation that some temporary advantage at a critical moment
>had become permanent and I thought aha, there was some quality of systems,
>whether biological or economic, at work. Someone else was bound to have
>noticed and I wanted to ask who.

Steven Gould is a bright guy, and AFAIK his speculation that the Cambrian
explosion may have been a path-dependant "lottery" is well respected by
biologists (I took a course from him last spring :) ). Yet you seem to
believe that path dependance has been discredited in the economic sphere:

> Well, reading about QWERTY and Microsoft in here has given me
>pause. Apparently there's not so much evidence to support the
>contention of what's being called 'path dependence' as I thought,

People like to say this, but lets compare this situation with that of
Gould's theory. Is there any direct evidence that the selection of the
organisms that prospered in the Cambrian was by-and-large a random process,
not correlated with the fitness of the organisms? Of course
not!--prehistory only happened once, and though we have a decent record of
it, the most we can say is that we can't find any strong trends in the
characteristics of the organisms that succeeded. So we conclude that it is
likely that evolution is a highly contincent process.

The "problem" is that neither evolution nor economics is a primarily
experimental science (economics probably less so than evolution). We don't
get a chance to repeat events with slightly different initial conditions to
see whether they come out differently. Nor do we get to observe systems
with the same precision as in physics. So, it would be almost impossible to
prove that a specific outcome arose by path dependance; but it's very
possible to look at a set of outcomes and say that it's very likely that
some were path dependant.

The people who led you to believe that path dependance is a failed economic
theory have tricked you into taking a very narrow view of the evidence. By
a broader view (which allows for common-sense interpretation, by the way),
path dependance is much more plausible.

>and
>the QWERTY phenomenon, which I'd seen mentioned many times and
>believed, is an urban legend. Dvorak and other style keyboards had a
>fair trial against QWERTY and were not shown to make typists much more
>productive.

This argument has been advanced by those who need it for their pet theory.
If you really believe that the Dvorak keyboard is anything but clearly
superior, read http://www.ccsi.com/~mbrooks/dvorak/dissent.html . The
Dvorak layout has objectively quantifiable advantages, and the nay-sayers
have even less hard evidence to go on than the proponents. Also, see the
paper by Stan Liebowitz that (I believe) started the Dvorak debate (it's
referenced by the above URL); read it closely and you'll be surprised at how
poor his logic is (I don't really feel like going into it now, but I can if
pressed).

> If somebody came to me 5 years ago and said, "I'm gonna bet
>the farm on this new company I'm starting, what kind of computer
>system should I use?" I would probably have said, go with Microsoft
>and intel style PCs, anything else is too risky if you're betting the
>farm, much as I would have hated mouthing those words. But isn't that
>a manifestation of this discredited path dependence eonomic theory?
>Now, (Oh Joy!) I might, with a clear conscience, convince the person to go
>with linux.

Your own experience leads you to believe that events are path-dependant, but
you reject the notion because someone else tells you it's discredited? Go
look for their hard evidence that Microsoft's (or QWERTY's) success is based
on fitness. It isn't there either. So rely on your own judgement.

> I actually wonder if linux (and/or Free-Net-OpenBSD and/or HURD, sorry
>guys for not mentioning you sooner) might not be a new economic
>phenonemon made possible only in this information age and a reaction
>to the hobbles of 'intellectual property'?

This seems to be a new topic, but free software absolutely makes
intellectual property look silly, or at least awkward. As the free software
movement comes into contact with the mainstream software industry, this will
become more and more clear. For example, I recently heard the author of a
proprietary program express dismay at a free software project to duplicate
his program's functionality. He had worked under the assumption that his
competiton would be from other proprietary programs, not from hackers who
share their work freely.

>In the past, if somebody
>came up say with a cool way to soup up his hot rod, he could tell his
>friends, maybe visit their garages and help them tinker with their
>engines, or publish his idea in a magazine, but he couldn't just give
>his souped up engines to people for free. You can do that with
>software though, and it is done, often.

Andrew

--
A lover's a liar,
To himself he lies.
The truthful are loveless,
Like oysters their eyes!

[SUSUPPLY]

Andrew Pimlott wrote:

" see the paper by Stan Liebowitz that (I believe) started the Dvorak debate
(it's referenced by the above URL); read it closely and you'll be surprised at
how poor his logic is (I don't really feel like going into it now, but I can if
pressed)."

Consider yourself pressed, Andrew.

Patrick

[cpw...@spamrahul.net]

"John Bianco" <RFte...@worldnet.att.net> wrote:

>As a side note, the Japanese did not just capture a big chunk by soley
>makeing high quality cars. The Japanese had ultra agressive and predatory
>business practices that untill the mid-late 80s often sold their cars at a
>loss to the US market to get market share. Most Japanese business' also
>shared this practive of being very agressive, while often outright stealing

Thanks for adding insight on this. I would say that it underscores
how difficult it is for newcomers to break in to a market. No American
seemed to be able to topple the Detriot Hegemony, not Bricklin or DeLorean
or that guy Jeff Bridges played in the movie (or was that Bricklin?) If
one relates this to my biological meme (I know, my posting is carrying a
lot of different concepts for one thread) it harks to the idea that new
evolution has to occur in protected environments.

and...@pimlott.student.harvard.edu (Andrew Pimlott) wrote:

|-On 23 Feb 1998 02:21:14 GMT, cpw...@Spamrahul.net <cpw...@Spamrahul.net>
|-wrote:
|-
|-[snip] <That's Andrew's snip from my posting, but I'm snipping more -cpw>
|-
|-Steven Gould is a bright guy, and AFAIK his speculation that the Cambrian
|-explosion may have been a path-dependant "lottery" is well respected by
|-biologists (I took a course from him last spring :) ). Yet you seem to
|-believe that path dependance has been discredited in the economic sphere:
Dammit Andrew I'm a programmer not an economist, I don't know if
it's been discredited or not. Maybe what's happened is that the mechanisms
proposed for path dependence, as in the case of Brian Arthur, are being
challenged. Reminds me of how American Geologists didn't believe in
Continental drift until a mechanism, plate tectonics, was identified. (I'm
glad somebody picked up on the biological side of things as my post to
sci.bio.evolution doesn't seem to have made it.)
<here's more of Andrew's post including some of his quotes from me>
|-
|-> Well, reading about QWERTY and Microsoft in here has given me
|->pause. Apparently there's not so much evidence to support the
|->contention of what's being called 'path dependence' as I thought,
|- <snip>
|-
|-The people who led you to believe that path dependance is a failed economic
|-theory have tricked you into taking a very narrow view of the evidence. By
|-a broader view (which allows for common-sense interpretation, by the way),
|-path dependance is much more plausible.
Actually I was playing dumb. I'm not convinced that path
dependence doesn't work but I read some of the stuff in those web sites
people cited in that other thread:
"Myths of Economics: QWERTY: Fact of <sic> Myth???" where they seemed so
sure of themselves, and maybe what they wrote is legit but I just
posted something in that thread myself with a real blast from the past.

Marc Fearby <ho...@tpgi.com.au> wrote:

>=My oh my! Where to start? I'll keep this brief, as I'm sure some
>=die-hard Linux user will elaborate for me.
Remember, I'm a Linux user and I said my prejudice was against
Microsoft. That was a sneaky trick. I figured if I said I was prejudiced,
maybe some people from the other side might admit they were prejudiced too.
>=
>=Microsoft is where they are today because they are good at marketing
>=their crap! Microsoft aren't really innovators - they are poachers that
>=copy other people's work and sell it for mega-bucks! It's a pity that
>=complete idiots can get managerial positions and choose wintel - simply
>=becuase it's the "popular" and "easier" choice. I'm sure if they were
>=independantly assessed, they would have to admit that Microsloth
>=products not only cost a fortune, but are buggy and cause more problems
>=than they solve!
<rest snipped>
OK you folks in sci.econ, that's an example of
comp.os.linux.advocacy. I agree with Marc, I share his feelings, but I'm
not sure his rhetoric persuades the unconverted. We can all make claims,
but proving them is hard. People try to prove path-dependency with the
examples of Dvorak keyboards and Beta video as 'better' products that
fail and the initial assumption that they were 'better' gets challenged.
We say Microsoft is lousy but how do you prove it to someone else?
Especially in a way that economists would accept and use to construct or
revise their theories? People can only prove it to themselves, it's so
subjective.
You know what? I thought about this post today, riding MUNI
to work, and I scribble down a few things. But then I had to work at
my job, and I came home and started slogging through these and other
posts, and my daughter needed help with her homework and now it's
late. I had, earlier in the day, an idea of how to tie some of the
disparate concepts of this together a little better, but it'll have to
wait Here's what I scribble down though:
3 questions:
Is MicroSoft so successful because they're good at providing what
the market wants or because of path dependency? (If there's no such
thing as path dependency could there be some other explanation for
it's success or does it's success absolutely prove it's good?)
If a monopoly exists can other companies topple that monopoly
or do you have to wait till their economic niche erodes? (OK, I'll
try to expand on this even now, this evening. Very simply, I think IBM
was a monopoly or monopoly like entity in the 60s, "IBM and the 7
dwarfs" they said. They were finally humbled by DEC but that's
because DEC occupied a niche, minicomputers, that IBM didn't, and that
niche widened into a very significant segment. Then DEC got humbled
by the next niche, microprocessors, that suggests to me that a
monopoly can last a very long time if it's market lasts a long time.
What's scary to me is, I think IBM only achieved it's clout in
mainframes as a path-dependency because of it's position in
pre-computer business machines, and Microsoft achieved it's clout
largely because it rode IBM's coattails into the PC market. (I
concede that not just any company that had formed a deal to provide an
OS for IBM in 1980 would have gotten where Microsoft is today, but, to
paraphrase Mae West, "Goodness had nothing to do with it"). So here
we have (I suspect) a path dependency going way back that perservered
through one change in technology (old style business machines to
mainframe computers) and leapfrogged another change in technology
(mini-computers). The implication is that we are just plain stuck
with Microsoft for as long as this technology rules. Anti-Trust
legislation may not help because it's like natural law (part of why I
bring in the biological comparisons). And what about the next
technological breakthrough. Will Bill Gates sit idle when nanotech or
whatever comes along, or will he exploit his path dependency advantage
to lay his heavy hand on THAT technology too?

----------Remove "UhUH" and "Spam" to get my real email address.----

[Grinch]

cpw...@Spamrahul.net wrote:
>
> I started reading sci.econ and sci.bio.evolution a little
>while ago because I had a thought and wanted to post something about
>it, but, I thought I should lurk awhile first and I'm glad I did. The
>thought sprang from reading an article in the New Yorker while
>in a waiting room that mentioned Microsoft and a paper written by
>Brian Arthur about how a company can 'lock in' a market. It struck a
>chord with me and it also struck me that there was a similarity to a
>biological phenomenon I had read about in a book by Stephen J. Gould,
>"Wonderful Life" which talked about the Cambrian explosion and how
>certain phyla of life had mysteriously become extinct while others, which
>in that time, seemed no more likely to flourish, such as the
>progenitors of the insects, had in fact done so. There was
>speculation that some temporary advantage at a critical moment had
>become permanent and I thought aha, there was some quality of systems,
>whether biological or economic, at work. Someone else was bound to
>have noticed and I wanted to ask who.

A talk "What Economists Can Learn From Evolutionary Theorists", given
by a noted economist to the European Association for Evolutionary
Political Economy, is at http://web.mit.edu/krugman/www/evolute.html

Teaser: It says some nasty things about Stephen J. Gould.

Regards

[Bas van Gils]

On 23 Feb 1998, SUSUPPLY wrote:

# Andrew Pimlott wrote:
#
# " see the paper by Stan Liebowitz that (I believe) started the Dvorak debate
# (it's referenced by the above URL); read it closely and you'll be surprised at
# how poor his logic is (I don't really feel like going into it now, but I can if
# pressed)."
#
# Consider yourself pressed, Andrew.

hmmm... at our FIRST YEARS UNDERGRADUATE class of microeconomics we
concluded the same thing :-) Many thanks to Dr. Martin van Tuijl :-)

grtz
_ _ _ _ _
>(')____, >(')____, >(')____, >(')____, >(') ___,
(` =~~/ (` =~~/ (` =~~/ (` =~~/ (` =~~/
~^~^`---'~^~^~^`---'~^~^~^`---'~^~^~^`---'~^~^~^`---'~^~^~
Bas van Gils b.va...@kub.nl
http://stuwww.kub.nl/people/b.vangils
Student Informationmanagement and Technology
Tilburg University, The Netherlands

[SUSUPPLY]

>hmmm... at our FIRST YEARS UNDERGRADUATE class of microeconomics we
>concluded the same thing :-) Many thanks to Dr. Martin van Tuijl :-)
>
>grtz
> _ _ _ _ _
> >(')____, >(')____, >(')____, >(')____, >(') ___,
> (` =~~/ (` =~~/ (` =~~/ (` =~~/ (` =~~/
> ~^~^`---'~^~^~^`---'~^~^~^`---'~^~^~^`---'~^~^~^`---'~^~^~
> Bas van Gils b.va...@kub.nl
> http://stuwww.kub.nl/people/b.vangils
> Student Informationmanagement and Technology
> Tilburg University, The Netherlands
>
>
>
>
>
>

I am also interested in your reasoning Bas, please elaborate.

Patrick

[Grinch]

and...@pimlott.student.harvard.edu (Andrew Pimlott) wrote:

>On 23 Feb 1998 02:21:14 GMT, cpw...@Spamrahul.net <cpw...@Spamrahul.net>
>wrote:
>
>[snip]
>>thought sprang from reading an article in the New Yorker while in a waiting
>>room that mentioned Microsoft and a paper written by Brian Arthur about how
>>a company can 'lock in' a market.

There's a rather tough critique of that article at Slate.
http://www.slate.com/Dismal/98-01-14/Dismal.asp

And here's the brouhaha that followed.
http://www.slate.com/Features/Krugman/Krugman.asp

>> It struck a chord with me and it also
>>struck me that there was a similarity to a biological phenomenon I had read
>>about in a book by Stephen J. Gould, "Wonderful Life" which talked about
>>the Cambrian explosion and how certain phyla of life had mysteriously
>>become extinct while others, which in that time, seemed no more likely to
>>flourish, such as the progenitors of the insects, had in fact done so.
>>There was speculation that some temporary advantage at a critical moment
>>had become permanent and I thought aha, there was some quality of systems,
>>whether biological or economic, at work. Someone else was bound to have
>>noticed and I wanted to ask who.

The similarities between economics and biology were noted a long time
ago. Darwin got his initial idea for how natural selection works from
reading Malthus and Smith. But the differences between the two fields
were noted a long time ago too.

There's no similarity at all between "path dependence" in economics
and the "lottery of life" in evolution. The essence of "path
dependence" in economics is *market failure*.
I read _Wonderful Life_ and it was interesting enough, but I don't
remember Gould writing anything in it about market failure in the
Cambrian explosion. Economics and biology are taught in different
academic departments for a reason.

>Steven Gould is a bright guy, and AFAIK his speculation that the Cambrian
>explosion may have been a path-dependant "lottery" is well respected by
>biologists (I took a course from him last spring :) ). Yet you seem to
>believe that path dependance has been discredited in the economic sphere:

Unless Prof. Gould's course covered such topics as positive returns,
network externalities, market failure and monopoly profits, it is
wholly irrelevant as to path dependence in economics. So what do your
class notes say about monopoly profits in the Cambrian?
Naively transplanting ideas from biology into economics without
understanding the differences in the fields is a very dubious sort of
pop intellectualism. I suspect that as far as their application to
economics is concerned, you are making much more of Gould's ideas than
Gould does.

>> Well, reading about QWERTY and Microsoft in here has given me
>>pause. Apparently there's not so much evidence to support the
>>contention of what's being called 'path dependence' as I thought,
>
>People like to say this, but lets compare this situation with that of
>Gould's theory. Is there any direct evidence that the selection of the
>organisms that prospered in the Cambrian was by-and-large a random process,
>not correlated with the fitness of the organisms? Of course
>not!--prehistory only happened once, and though we have a decent record of
>it, the most we can say is that we can't find any strong trends in the
>characteristics of the organisms that succeeded. So we conclude that it is
>likely that evolution is a highly contincent process.

Yes. So what?
What does this have to do with path dependence in economics?

>The "problem" is that neither evolution nor economics is a primarily
>experimental science (economics probably less so than evolution). We don't
>get a chance to repeat events with slightly different initial conditions to
>see whether they come out differently. Nor do we get to observe systems
>with the same precision as in physics. So, it would be almost impossible to
>prove that a specific outcome arose by path dependance; but it's very
>possible to look at a set of outcomes and say that it's very likely that
>some were path dependant.

That's not a problem, it's a trivial observation.
The idea that slight, random changes in initial conditions have
unpredictable large-scale consequences was examined in economics by
Alfred Marshall 100 years ago when he was doing his work on positive
returns.
Let's not confuse century-old knowledge with some clever new
insight.

The point of the "path dependence" argument in economics is *not* that
some slight historical accident can have major consequences later,
"locking in" the behavior of millions of people in regard to, say,
QWERTY keyboards. That's obvious.

The point of the economic path dependence argument is that the "lock
in" force is so strong that it somehow prevents economic actors from
improving their welfare through normal market activity -- that they
are somehow "locked in" to a *poorer* state of being and *cannot
escape*. In other words, market failure.

For anyone who's seriously interested in this subject, here's a good
professional article explaining it: "Path Dependence, Lock-In, and
History" from the Journal of Law, Economics and Organization,
http://wwwpub.utdallas.edu/~liebowit/paths.html (It contains a nice
analysis of the VHS/Beta story.)
You'll notice the lack of references to Gould or other
biological-science sorts of fellows. You'll also note that while it
is easy to show the "small cause/big consequence" effect in economics,
it's hard to show that it has much economic significance.

For a more simplistic example, try this:
In the US today everybody drives on the right hand side of the
road. In the UK people drive on the left. One can easily imagine that
but for some odd historical accident 100 years ago, when cars were
first hitting the road, people might be driving on the left today in
the US.
If so, then some minor, arbitrary event 100 years ago has "locked
in" the behavior of millions of people today as they drive about in
their cars, and also has "locked in" the design of cars and many of
the actions of the huge corporations that build them, and so on. So a
minor historical event has had major lasting consequences.
Is this an example of economic path dependence? Not at all,
because it makes *no economic difference* whether cars drive on the
right or the left. It's entirely arbitrary. If that's the
biologist's definition of "path dependence" when looking at economics,
it is entirely uninteresting and pointless.

The path dependence argument in economics says this:
Assume that in the US today a new technology is developed that
enables cars to get 30% more miles per gallon, but *only if* they are
driven on the left side of the road. These savings *can't* be
realized because people are "locked in" to driving on the right side
of the road. The first person to switch to driving on the left will
die in a head-on, and we all know it. So nobody will ever switch.
Thus we are all *poorer* than we should be, by dint of historical
accident.

Of course, if in reality such savings could be attained through a
new generation of cars, you'd see traffic rules change very fast. It
costs no more to drive an old car on the left than the right, so
there'd be no economic reason not to switch. There'd be big monetary
incentives to switch. (And energy-saving, environmental, and
profits-for-carmakers reasons as well.) The only delaying factor would
be some people not wanting to change their way -- but they wouldn't
hold up the majority for long.
In actuality countries *have* changed their traffic rules to
switch the side of the road they drive on, for much less reason than
this. (I remember pictures of all the traffic stopped when the last
one did so -- Sweden?)
And people have changed all kinds of other general standards they
once followed, as soon as they had a reason to do so -- CP/M for DOS,
and you can think of a million other examples.
But the "path dependence" argument says these changes happen much
more slowly than they "should", or don't happen at all when they
should, and that we are all the *poorer* for it. Without the alleged
economic cost, there's nothing at all to it as far as economists are
concerned.

The naive believers in path dependence (QWERTY-myth believers)
make the mistake of assuming that because the trivial type of "lock
in" is easily seen, it is economically significant. And they overlook
the obvious fact that if history should "lock in" the economy to a
really costly standard or practice, then really *strong* incentives
are created to break it. The more costly the "error" of history, the
*greater* is the reward for fixing it.
That's why old standards and practices are abandoned all the
time, and why it's so hard to find much *measurable* economic cost of
path dependence -- in spite of the "new" insights of biology students.

>The people who led you to believe that path dependance is a failed economic
>theory have tricked you into taking a very narrow view of the evidence. By
>a broader view (which allows for common-sense interpretation, by the way),
>path dependance is much more plausible.

Do please explain the broader "common sense" interpretation.

>>the QWERTY phenomenon, which I'd seen mentioned many times and
>>believed, is an urban legend. Dvorak and other style keyboards had a
>>fair trial against QWERTY and were not shown to make typists much more
>>productive.
>
>This argument has been advanced by those who need it for their pet theory.
>If you really believe that the Dvorak keyboard is anything but clearly
>superior, read http://www.ccsi.com/~mbrooks/dvorak/dissent.html.

Oh, sheesh. This link leads to one howler after another. I'll point
out a few at the end of this post.

>The
>Dvorak layout has objectively quantifiable advantages, and the nay-sayers
>have even less hard evidence to go on than the proponents.

Except for several controlled studies (referenced below) that found
Dvorak to be worth zilch, compared to zero controlled studies finding
Dvorak to be worth anything.

>Also, see the
>paper by Stan Liebowitz that (I believe) started the Dvorak debate (it's
>referenced by the above URL); read it closely and you'll be surprised at how
>poor his logic is (I don't really feel like going into it now, but I can if
>pressed).

Here it is, in the Journal of Law & Economics vol.. XXXIII (April
1990) http://wwwpub.utdallas.edu/~liebowit/keys1.html

Please do compare the quality of the logic therein to the silliness on
the Web page you cite.

>> If somebody came to me 5 years ago and said, "I'm gonna bet
>>the farm on this new company I'm starting, what kind of computer
>>system should I use?" I would probably have said, go with Microsoft
>>and intel style PCs, anything else is too risky if you're betting the
>>farm, much as I would have hated mouthing those words. But isn't that
>>a manifestation of this discredited path dependence eonomic theory?

No.
Wouldn't the computer system you choose depend on your needs? For
generic business word processing, spreadsheets, etc, maybe Windows.
But if you're in the publishing business, Apple's still got the market
lead, better graphics. For a corporate server, Sun's got the market
share lead -- but you might go with something else if someone else
offers an even better product for your particluar needs at a better
price.
What's path dependence got to do with that? Unless you say that
*every* company that has *any* kind of competitive advantage anywhere
is benefiting from path dependence, which would be trivializing the
whole idea out of existence.

>>Now, (Oh Joy!) I might, with a clear conscience, convince the person to go
>>with linux.

If it's best for their needs, sure!

>Your own experience leads you to believe that events are path-dependant, but
>you reject the notion because someone else tells you it's discredited? Go
>look for their hard evidence that Microsoft's (or QWERTY's) success is based
>on fitness. It isn't there either. So rely on your own judgement.

Again missing the fundamental point about QWERTY and standards. Some
arbitrary keyboard layout must naturally exist as a standard. It might
as well be QWERTY as any other, such as IXGLOT or PORKME, so long as
no other is demonstrated to be *measurably better*.
The fact that QWERTY itself may be no better than many other
possible layouts does *not* support the argument for path dependence
in economics, because as long as no other standard is measurably
better than it, the QWERTY standard carries no economic cost.

Now, let's take a brief look at the recommended pro-Dvorak "Brooks"
site, http://www.ccsi.com/~mbrooks/dvorak/dissent.html

Remember, the whole point of the QWERTY/Dvorak "path dependence"
argument is that *Dvorak is measurably better* -- not as a matter of
taste in that some people may prefer one keyboard over another, but as
a matter of basic principle -- and that "path dependence" makes us all
poorer by stopping us all from switching to Dvorak.

Thus, if QWERTY and Dvorak have similar merits, Dvorak *loses* -- and
so does "path dependence".

So, here's the number of studies mentioned in the Brooks site that
document superior performance for Dvorak: ZERO.

However, the site does say this:
> But what about studies showing Dvorak is worse than QWERTY? There aren't any!
Duh.

That should end the argument right there. But I'll mention a couple
more points to illustrate the incoherence of the Brooks-site "defense"
of Dvorak.

The Liebowitz/Margolis paper points out that the reputable supporters
of Dvorak have *over and over* cited exactly *one* study showing its
alleged superior performance relative to QWERTY. This was a Navy
study conducted in the 1940s which generated a truly impressive result
-- retraining in Dvorak paid for itself within 10 days of the *start*
of retraining. That's like a 2000% annual return on investment!
However, L&M found a couple of interesting things regarding this
study:
First -- Nobody who had cited it had actually read it! L&M were
the first ones to track down a copy of it. All the others' retellings
of it had been, well, urban legend.
Second -- When L&M found and read the Navy study, they discovered
it had been conducted by ..... Lt. Cmmdr. August Dvorak, holder of the
patent on the Dvorak keyboard! Moreover, he was in the process of
receiving $130,000 in grants for his work on it (a *lot* of money in
the 1940s). And it took him *three* studies to get the results he
wanted.

How does the Brooks page resond to these two findings?
It completely misrepresents the first one.....

>The "Fable" article belabors the point that the Navy cannot produce a
>copy of its study.... but keep in mind we are speaking of a Government
>agency in a chaotic time (World War II). By coincidence, some time ago
>I asked the Navy for information about an airplane they evaluated in
>the early '40s (the XNL Langley Twin); they could find no record except
>the plane's name.
Logic here? How does this answer the fact that nobody who cited the
study in support of Dvorak had read it?

As to the second one....

>What about Dvorak's patent? The article's authors might like us to
>think anybody who holds a patent is a fraud. I doubt it.... On the
>face of it, that does show vested interest, but it also shows that
>Dorak believed in his invention!

Ah, "vested interest" indeed. I'm sure Dvorak *did* believe in his
invention. And when Consumer Reports starts letting product makers
evaluate their own products for it, this answer may begin to make some
sense.

If a Harvard student really thinks this kind of Web page illogic
refutes L&M in the Journal of Law & Economics, then what so many say
about the decline of American education may be true.

I leave you now by reposting the results of some controlled studies
that *have* been done on Dvorak, and which L&M mentioned, but which
the Brooks pro-Dvorak web page *somehow* managed to ignore -- as have
all retellers of the QWERTY/Dvorak myth.

[]A 1953 Australian Post Office study failed to show any benefit
versus QWERTY.
[]A 1978 study at Oregon State University indicated that after 100
hours of training, Dvorak typists were up to only 97.6% of their old
QWERTY speed.
[]A 1973 study at Western Electric found that after 104 hours of
training on Dvorak, typists were all of 2.6% faster than they had been
on QWERTY. (Of course, after 100+ hours of training you'd expect
QWERTY typists to be a lot better at QWERTY.)
[]A. Miller and J. Thomas, of the IBM Research Laboratory, writing
in the International Journal of Man-Machine Studies, concluded that
"no alternative has shown a realistically significant advantage over
the QWERTY for general purpose typing."

The Urban Legend of QWERTY/Dvorak no doubt will live on in the
wider world. But as far as this newsgroup is concerned, let's bury
once and for all the myth that Dvorak is a superior keyboard that has
been "suppressed" by QWERTY through "path dependence".

BTW, if you like the Dvorak keyboard, go out and buy one. Nothing's
stopping you. ;-)

Regards

[Grinch]

oldn...@mindspring.com (Grinch) wrote:

>For anyone who's seriously interested in this subject, here's a good
>professional article explaining it: "Path Dependence, Lock-In, and
>History" from the Journal of Law, Economics and Organization,
>http://wwwpub.utdallas.edu/~liebowit/paths.html (It contains a nice
>analysis of the VHS/Beta story.)
> You'll notice the lack of references to Gould or other
>biological-science sorts of fellows.

Except for footnote 14:
"We should note that both Gould and Mokyr warn of the misapplication
of biological evolution to social systems."

My mistake. ;-)

Regards

[Gary Forbis]

Grinch wrote:

> For anyone who's seriously interested in this subject, here's a good
> professional article explaining it: "Path Dependence, Lock-In, and
> History" from the Journal of Law, Economics and Organization,
> http://wwwpub.utdallas.edu/~liebowit/paths.html (It contains a nice
> analysis of the VHS/Beta story.)

Thank you for yet another interesting link.

I'm a bit concerned by the arguments about third degree path dependency.

I remember a story told several years ago about how one city's bus system
was primarily electric trollies on rail but was bought out by GM and Shell
whereupon they replaced the trollies with busses and ripped out the rails.
Whether or not this story is true, isn't it possible for a concerted effort
on the part of parties with deep pockets to establish a less than optimal
standard which then behaves as if a second degree path dependency?
If such a situation is found to exist would it not fit the author's definition of
a third degree class dependency?

--
--gary
for...@accessone.com
http://www.accessone.com/~forbis

[r.e.b...@usa.net]

In article <6cqmeq$mea$1...@news1.sirius.com>,

cpw...@Spamrahul.net wrote:
> <NOTE, a much shorter version of this article, with the same
> subject, is being posted to sci.bio.evolution>

[snip]

> There was
> speculation that some temporary advantage at a critical moment had
> become permanent and I thought aha, there was some quality of systems,
> whether biological or economic, at work. Someone else was bound to
> have noticed and I wanted to ask who.

There is some truth to this. Very often, throughout history, great
innovations have been stifled and even "dead and buried" only to surface
at a "critical moment".

UNIX and the Internet were a very small portion of the market while IBM
dominated the Server market with it's MVS mainframes. It also dominated
the PC market with it's "IBM & True Compatibles".

There were two triggering events that moved IBM from leader to follower
in a very short period of time. IBM decided to switch from the ISA bus
to the Microchannel bus and refused to disclose the intended user of the
"reserved" pins. The existing user base (about 80 million) didn't want
to be told their machines were obsolete, and the card vendors didn't want
to end up with obsolete cards when IBM finally released the design that
used all the pins. The result was a revolt by both consumers and 3rd
party providers that left Microsoft as the dominant force in the industry
and gave Compaq, Dell, and others a much larger share.

During this same period, IBM released MVS 4.0 which, combined with all of
the corequisite upgrades totalled a price that was more than the depreciated
price of the hardware, on company was quoted $30 million for software
upgrades to equipment valued at $10 million.

The backlash was that VPs, faced with the choice of layoffs, poor profit
figures, or other budget cuts, were suddenly willing to listen to the
advocates of Open Systems, particularly the UNIX people. When these VPs
realised that they could delay or eliminate the need for the upgrades by
using UNIX systems from SUN, HP, or even IBM to offload some of the
work from the mainframe, a market shifted overnight.

The shift from LU 2 to TCP/IP opened the door for corporate "intranets"
between sites using frame-relay or ATM. When MCI started offering access
to the NSF archives and other government databases via the "internet"
(little internet), it created a market for Publishers who were able
to reach thousands of business users for a fraction of the price of
"paper copy" information. In addition, the information could be more
current, making it more attractive to the Business user. At the same
time, dial-up users started accessing this information via independent
services providers.

Though it looks like the "Internet" (big internet) was an "overnight success",
it was actually something that had been waiting in the wings "eating the
eggs of the dinosaurs" until the dinosaurs decided they didn't need to
protect their eggs.

Fred Smith had approached the U.S. Postal Service AND U.P.S. with his
proposal for 1 day package delivery. During a period of complacency
in the package delivery business, the industry was "ripe" for Federal
Express.

The inventors of the Quartz watch tried for years to persuade the Swiss
Watchmakers that their invention would be a hit in the marketplace. When
they changed tactics and talked to Japaneze watch-makers, Seiko and quartz
watches became the de-facto standard of the industry.

> Well, reading about QWERTY and Microsoft in here has given me
> pause. Apparently there's not so much evidence to support the
> contention of what's being called 'path dependence' as I thought, and
> the QWERTY phenomenon, which I'd seen mentioned many times and
> believed, is an urban legend.

I have written several articles on this. The QWERTY layout is less
efficient, but was heavily advocated by typing schools and typewriter
manufacturers.

> Dvorak and other style keyboards had a
> fair trial against QWERTY and were not shown to make typists much more
> productive.

Ironically, the breakthrough proof came through two forms. The Navy started
using Dvorak Simplified Keyboard (DSK), for the conversion data between
systems. For several years, the incompatibilities between computers made
it necessary to have typists take printouts from one system and manually key
it into the other system. The DOD discovered that they were saving 20%
by switching to the DSK. Part of this was because the the work involved
other activities such as manipulating papers and transposition of forms.

The real proof was developed by Computer Consoles of Rochester, New York.
They had developed the first fully computerised directory assistance system
and to get around union rules, installed DSK consoles instead of the QWERTY
consoles. The immediate gain was that they could take lower paid operators
and train them to type very quickly in a very short amount of time. CCI
also kept extremely accurate statistics on keying time. During a CWA strike,
managers and secretaries had to be provided with QWERTY keyboards. The
statistics sytem indicated that keying time had increased by 50-80%.
Eventually the CWA returned and not only did the management return the
DSK keyboards, but they also began paying MORE for the DSK typists even
though they were more easily trained.

These discoveries led to the including of DSK options on the Apple IIc,
the Mac, Windows 3.1, and X11 servers. Most DSK users do NOT rearrange
the keycaps, but leave them as is in case they need to look at the QWERTY
layout. Some people even become "bimodal" easily switching between
keyboards. I started using the DSK in 1984, and have used it whenever
possible since. I personally don't type that much faster, but often find
that I can quickly create a document, especially a news posting or
an informal report, in a few minutes.

> Other examples of supposed inferior products locking out
> better products, like VHS triumphing over Beta, also don't bear up to
> close inspection. I think there's a strong force driving these
> legends though, based upon peoples experiences and perceptions of the
> market.

Beta was superior for editors and news rooms because the tape was usually
being quickly manipulated for editing purposes. VHS was much less expensive
and much more reliable for the viewing of rented movies. Those who purchased
the inventory and replenished the damaged tapes quickly developed a
preference for VHS.

> Back in the 50s and 60s people complained about Detroit's
> "we build 'em and you buy 'em" attitude towards customers. Finally
> Detroit faced competition from foreigners.

The advantage of an oligopoly or monopoly is that you have economies of
scale. In other words, you can strong-arm your suppliers and dealers
into giving you deals that they would never agree to in a competitive
marketplace. Foreign competition not only gave the dealers new product
variety, but also gave them bargaining leverage with the Oligopoly.

> Of course, the Big 3 or Big 4 of Detroit were not a monopoly,
> they were an oligopoly.

The american car industry once consisted of over a dozen car makers.
Companies like Hudson, Studebaker, Nash, Rambler, Jeep, International,
Cord, Pontiac, Chevrolet. Eventually modular construction enabled
mergers and aquisitions and enabled the delivery of various options
that could be selected at the show room floor. Even today, many
dealers make a substantial portion of their profit by installing
the radio, air-conditioner, sun roof, and various other customisations
which are applied "in house".

> I'm not hear to advocate anything, or to try to persuade or
> change peoples minds (I know better than that, I've been reading
> Usenet a long time). My interest is in a socratic argument to try to
> find the truth, really.

The truth is that computers are just millions switches flipping on and off
at 100 million to 600 million times per second. The rest is
interpretation :-).

Seriously, there are often many ways to "get from point a to point b" on
the grid. Some just take you through rougher neighborhoods. Programming
is both a science AND an art. Microsoft is weak on science and very good
on art. Even from the first copies of MS-DOS, Microsoft and IBM made it
a point to distinguish their products from CP/M and UNIX by concentrating
on graphics, colors, high resolution, and "presentation".

Until Linux, the "UNIX" world was heavily focused on science and minimally
focused on art. Even with the advent of X11, it was very common for UNIX
gurus to choose TWM because it was fast, even if it was ugly. Because UNIX
people were almost entirely focused on reliability and performance, there
wasn't much market for pretty software. The UNIX programmer would whip out
an ad-hoc program to parse and catalog 25,000 documents in less than an hour.
Often, it ran as a cron job and was never heard from again. A few
applications did eventually become popular on DOS and became pretty.

The classic example is this. Unix comes with two utilities, compress and arc.
To compress all of the files in a directory and save them into a single file,
the unix programer could type "compress *; ar cv rex.ar *" and have
the entire directory in a compact library.

In the PC market, the two were combined into something called arc. PKWare
created a competitor called pkarc. When the developer of arc sued, PKWare
switched to the GNU compression "zip" (I believe they contributed it).
PKarc became pkzip. When windows came out, PKWare created "Winzip".

In other words, a 1 line command evolved on MS-DOS and Windows into a
huge program costing $40.

> But I'll be up front about the fact that I
> have very strong prejudices, and I admit they are prejudices.

Welcome to the club.

> I'm a
> computer programmer who has been in the biz since 1978, and been
> working with Unix since 1982, and I hate it when I have to come near
> Microsoft stuff. But it's a prejudice because I honestly haven't used
> Microsoft stuff enough to give it a fair trial.

I have. I have used both Microsoft products and Unix variants since
1981. I actually started in computers back in 1976 and remember when
Microsoft BASIC was sold on Paper Tape for $50 (Microsoft wanted $500).

Microsoft has thrived by setting standards and getting everybody to
agree to them. The Commodore Pet, the TRS-80, and the IBM PC all
sported BASIC-IN-ROM written by Microsoft. IBM's muscle created the
agreement required to make MS-DOS (IBM preferred to call it PC-DOS)
the industry standard.

The flip side of this is that Microsoft maintains it's revenue stream by
staging "Revolutions" every 2-3 years. Until Microsoft releases the next
coup, it rigidly enforces it's current standard, often by publicly
discrediting anyone who doesn't measure up, especially competitors.
Just as the competition heats up, Microsoft will release it's next
revolution, which breaks most of the old software but that's OK because
Microsoft has new software to replace what was broken.

> (What am I gonna do?
> Go out and buy their stuff just to see if I like it? I got Windows
> 3.1 shoved down my throat twice when I bought PCs, putting money in
> Bill Gates pocket for something I didn't want.)

People use Windows 3.1 out of ignorance (they don't know any better and
are unwilling to find out). They use Windows 95 out of convenience (it's
what came with the computer and they don't know or care how much it cost).
They use Windows NT out of arrogance and "status" (NT is Microsoft's "Top
of the Line" - like a Lincoln Town car).

> I was working under contract at a small outfit in the summer
> of '96. This company started out making their application for use
> under X-windows, but their plan was to move everything to windows or
> windows NT. I mentioned to the owner/president/entrepreneur one time
> that I hoped to be able to retire before Microsoft took over the
> world. He said I had about 2 years (maybe it was 3, I don't
> remember).

Many great leaders have tried to take over the world. Julius Ceaser,
Napoleon, Hitler, Stalin, J.P. Morgan, and Bill Gates. The methods and
tactics varied, but each tried to determine what resources needed to be
controlled to control "the world", and sought to control those resources.

Each also had someone with a plan to stop them. Ceaser had Butus,
Nepoleon had Wellington, Hitler had Churchill, Stalin had Eisenhower,
Morgan had Roosevelt, and Gates has Linus (or me :-). The opposition
in each case appeared to be overwhelmed and the conquerer destined to
succeed, but the tide turned, quickly and radically.

> All of his clients wanted Windows NT. I had never met his
> clients, but I've got my prejudices to fall back on, I imagined them
> as Dilbert style managers, nervous nellies whose thought process went
> something like "Microsoft big, Microsoft must be good" but that's path
> dependence, isn't it?

I probably knew his clients. They wore Ghermani Suits that cost $3000
each, Rolex Watches that cost $5000, drove Lexus or Mercedes Benz cars
costing $50,000, and Lived in East Side apartments that rented for
$5,000/month or sold for $1 million. The suits tore, the watch stopped,
the car was "in the shop", and the apartment was "on the market" for
over 2 years. They also made sales presentations on Windows 3.1 Laptops
that crashed in the middle of the show and didn't notice until after
NT Workstation was installed on dozens of machines that the $50 CAL was
only for the ACCESS to the NT server. He was good buddies with the
CIO and so he didn't lose his job when he had to add $2 million in
royalties to his original budget. They didn't even raise an eyebrow
when the Servers would crash every week. They just added 4 more admins
to keep up with the load.

Meanwhile, back at the ranch, a guy in $200 suits (when he wore them),
who wore a $50 Citizen watch, drove a Mazda Protege (when he drove),
lived in a huge $500/month apartment in a great neighborhood, quietly
ran 5 UNIX boxes that ran for 6-12 months at a time between service
calls and kept systems going that provided 80% of the division's revenue,
over $1 million/day. Although he had a top of the line Windows Workstation,
he was able to quickly produce very effective solutions and create new
business products very quickly on his Linux PC, a box he had rescued from
the Dumpster.

One day, the Microsoft guy came to the UNIX guy and told him to publicly
endorse NT and promise never to say anything bad about NT again. The UNIX
guy left, and the Microsoft guy suddenly found that he was spending almost
as much as he was making. Fortunately a friend of the UNIX guy was able to
step in and keep it all working. Until they switched to NT.

> So I have prejudices, and ya'll are invited to confirm those
> prejudices or help show me the true nature of things. But I'm not the
> only one with these prejudices. If you look in the advocacy groups,
> (I occasionally subscribe for awhile to comp.os.linux.advocacy, then
> get out after awhile, the repetitiveness and rhetoric get to be too
> much, but I'm resubscribing now for awhile to chase followups to
> this.)

c.o.l.a. is a very closely watched group. If you want to know what Microsoft
is doing, you go to the Microsoft Web site. If you want to find out what the
Linux community is doing, the best place to look is c.o.l.a.

> In comp.os.linux.advocacy you'll see variations on a theme,
> ("Linux is great, NT stinks, I know I've used both", "No, NT is great,
> Linux stinks, I know, I've used both). Actually, since I only ever
> subscribe to the linux advocacy group, the response is more like, "NT
> is so reliable, and the support is so good, and if you've had training
> it can too be flexible, etc." Usually the NT advocates will admit
> that Linux makes more efficient use of computer resources than NT
> though.

Linux is a rapidly evolving platform. It's strengths are:
Performance (ability to get high throughput from low-cost hardware,
and higher throughput from high-cost hardware).
Reliability - Linux has gained a reputation for going 3-6 months without
requiring a reboot or having an unrecoverable failure. Stories
of Linux boxes that have run for a year or more are common.
One admin claims that he has been running a machine
continuously since 1983.
Low TCO - Linux can easily be configured to be self-maintaining. In
the event that service is required, remote login and access
provides the capability to manage remote systems. This
also means that consultants can often service multiple small
companies and fewer admins are needed for large companies.
Many Linux admins double as programmers or mentors as well.

NT also has a few strenghts:

Ease of use - The NT user interface and the Microsoft applications are
generally very easy to learn and very consistent. There is
also a much lower initial training cost. Linux has a
"staircase" learning curve. People learn the interface
and can do what they do with NT. Then they learn more tools
and become more productive. It's important to balance
learning with productivity.

Installation - NT usually comes preinstalled and the manufacturer has taken
the steps necessary to make sure that NT will detect the
hardware configuration.
Linux must be installed by the user and will usually detect
the configuration correctly. The exceptions can be painful.
Some vendors such as D-Link are now putting "linux compatible"
on their boxes. 3Com has always supported Linux and is a
pretty safe bet. Some video cards such as Trident 9685 chips
are just not supported. Cards using the S3 chip provide
excellent service.

Push-Button Administration - NT proponents point to the GUI administration
tools that are available. Most resources are configurable
through the various tools in the Control Panel. The remainder
can be configured through the registry editor. There is also
the option of editing an ini file and rebooting.
The down side is that most configuration changes require
a reboot of the entire system. Linux can be managed with
almost no rebooting, the exceptions are changing the kernel,
changing ISA cards, or changing the interface.
Linux has some great GUI tools as well. The Red Hat control
panel is very nice. Caldera also provides a pretty friendly
panel. Slackware provides a "CHUI" (graphical Character User
interface), it's not pretty, but it is easy to use.

> Linux folks reassure themselves that Gates can't drive them
> out of business like he's driven so much competition out of business
> because it's not a business. Linux is free and anyone can get the
> code and support it.

This is only partially true. While Linux itself is not owned by a
single corporation, there are several tightly held companies that
are relying on Linux for their economic survival. Caldera and Red Hat
are totally dependent on the success of Linux for their continued growth
and survival. The good news is that they are being well supported. Debian,
Slackware and FreeBSD are also supported, but the companies that back them
are not totally dependent on Linux. Another company that gets very little
press is WorkGroup Solutions. They make a very good distribution with
some good commercial-grade applications. In fact, WGS actually was a
pioneer in the Linux-based Integrated WorkGroup concept.

> (Is it a prejudice or unfounded urban legend
> that Microsoft has driven worthy companies out of business?

I worked for a company called Data Law in 1981. They had built
a Legal billing system around Microsoft Basic based on the understanding
that the run-time was free. When we upgraded to the next version,
Microsoft called up and demanded that we pay $400/seat for each of
our 10,000 installed seats. The company went chapter 11 and was taken
over by a larger firm after Data Law had laid off almost all of their
programming staff, including me - who had reccomended Microsoft in the
first place.

> An economist would say, cite examples, and some people have cited
> examples in the newsgroups but I don't know how scientific the info
> is)

Let's just put it this way. A company that gets too friendly with Microsoft
makes a good takover candidate. Very few companies go out of business by
going chapter 13, liquidating their assets, and closing their doors. Most
try to find someone with deeper pockets, better negotiators, and better
lawyers. There are usually layoffs and reorganizations in the process.

> Some linuxers worry that Gates will find SOME way to destroy
> linux, that he'll concoct a claim that linux is violating copyrights
> or something. I'm not sure how paranoid the average Linuxer is
> though, that may just be the fringe element talking. Isn't Andy Grove
> at Intel supposed to be a big believer in paranoia?

There is little doubt that Gates would like to do everything he can to
make sure that Linux dies, grows as slowly as possible, or becomes
associated with some radical fringe element of psychopathic sociopathic
cyber-terrorists. Groups like http://www.ultraviolet.org can be painted
this way. Linus isn't exactly a "suit", Pat Volkerding has hair down to
his shoulders and wears it in a pony tail (or used to anyway), RMS looks
like Jesus after the crucifixion, and there's "Maddog" Hall. I may be
one of the few linux users with a "Corporate" image (even I'm letting my
hair grow a little).

On the flip side, Gates doesn't want any martyrs. Microsoft isn't likely
to take on a direct assault. That would be very bad PR, and Microsoft has
enough problems already. The tactics are more subtle such as creating
deals with MCI to provide the NT servers to handle their Dial-up ports, and
then not following the public standards. The net effect is that Linux users
cannot log in and only Win32, Windows 95, and Windows NT users can access
the internet via MCI. Since MCI is not only the largest backbone provider
on the Internet, but is also the leader in dial-up POPs, it means that
Microsoft has pretty tight control, including the ability to monitor
traffic, check cookies, and claim statistical advantages.

[IBM switched from punched card polling machines to computers]

IBM had a good sales force, good relationships with the business community,
and a good legal team. Microsoft had a good sales force (OEMs), good
relationships with the business community (they have direct sales forces to
the fortune 500). Microsoft also has good negotiators and a good legal
department. Microsoft is very careful to not cross the line from unethical
to illegal.

> The guys who actually invented the computer, Eckert and
> Mauchly, had a lot of bad luck. For instance, just as the New World
> got named after Amerigo Vespucci instead of Christopher Columbus, so
> computer scientists speak of the "Von Neumann" model computer instead
> of the "Eckert-Mauchly" computer.

Linux is loosely based on Tannenbaum's MINIX. At the same time there
was GNU HURD, and Mark Williams' Coherant. All of these were candidates
to create the operating system that could "Beat Microsoft". Linus was
actually elevated by a few promoters :-). Much of the Linux distribution
comes from Richard Stallman's FSF, and Becker did the TCP/IP port from
his KA9Q package.

> But the troubles that happened to
> Eckert and Mauchly might be considered Brian Arthur style historical
> small events, there's that notion of path dependence coming up again.
> My historical notions come from the book "Bit By Bit, An Illustrated
> History of Computers" by Stan Augarten, Ticknor & Fields, New York,
> 1984.

Some call it "being in the right place at the right time". Arthur
Nightengale put in nicely. "Luck is when preparedness meets opportunity,
and Opportunity is there all the time". Bill Gates had learned enough
about computers to be able to write a BASIC interpreter (he learned by
studying a listing he found in a dumpster) - the real opportunity that
created Microsoft was that a 15 year old kid (Gates) was able to find
pure gold in a dipsy-dumpster. By the time the Popular Science article
on the MITS came out, Gates was prepared to write the BASIC interpreter.
Microsoft has often given DEC preferential treatment as a result.

> I didn't work with IBM computers in the old days (before they
> got into Unix type systems) either, but I spoke to a lot of people who
> had. From what they said, everything else was better but "nobody ever
> got fired for buying IBM", those damned Dilbert-style no-nothing
> managers again.

That really wasn't true, and hasn't been since the mid 1970's when
IBM switched from the BSC to the SNA protocols. Vendors that supported
IBM with "plug compatible peripherals" suddenly found themselves with
no market and no product. Admins who had purchased leased lines and
hardware to support BSC suddenly found themselves doing tap-dances to
justify replacing equipment that was, in some cases, less than a year old.

Those who purchased the "Series 1" instead of the DEC PDP-11 also found
themselves with orphan hardware excessive upgrade costs, and no marketplace.
IBM had ignored UNIX and DEC had begun exploiting it heavily.

Those who purchased 3090-600J computers in 1990 were suddenly facing
upgrade costs as IBM switched to ES-9000 and the ESA architecture along
with the shift from 370 to 390. The double whammy left many admins and
even many IBM employees hitting the pavement hard.

> Then there's this other super-successful giant, Intel.
> I was an assembly language programmer for years and people hated the
> instruction set of the 8x86 style computers, but they were forced down
> our throats, that's what the customer wanted.

Prior to the 8x86, your choices were the 808x, the Z80, the 6502, and
the 6809. The day Intel unveiled the 8088 and the 8086, people looked
at the 16 bit segmentation and dubbed it the biggest mistake intel ever
made. As dumb as it was, developers wrote code that was so tightly
dependent on it that neither the 80286 nor the 80386 could break the code
loose. Even today "thunk mode" and software emulation are used to support
the 8086 dependent segmentation.

> Presumably because IBM
> used them in their PCs, IBM who just sort of decided to dabble in PCs
> and then all the clueless layman public rushed out to buy them just
> because they had IBM's name on them.

IBM didn't just decide to dabble in PCs. IBM clearly saw that the Z80
based Cromemco Z-DOS (unix-like) and MP/M were a direct threat to IBM's
bread-and-butter mainframe business, not to mention it's 3270 terminal
business. What IBM wanted was a brain-damaged Personal computer that
was just good enough to keep people from buying CP/M-86 or a 68000 based
*NIX box.

By the time IBM came out with the PC/XT, the competition
was threatening to come out with UNIX systems based on the Z8000, 68010,
80286, and some other minor chips. The divestature of AT&T actually helped
IBM because AT&T took ownership and tied UNIX up in legal battles with
BSD. AT&T wanted $50,000 or more for UNIX source code, and $700/user
for the run-time binaries to the "Brain Damaged" System III. Eventually,
AT&T settled with BSD, trading rights to V7 code and utilities for
BSD 4.x code and utilities. It took AT&T almost 4 years to come up with
a version of UNIX SysVr4 that met the expectations of BSD 4.2 users.

Just as UNIX looked like it was about to get very hot, Microsoft sold
it's interest in Xenix to SCO and decided to dedicate all of it's resources
into a competing product (Windows). The Mac became the machine to beat.
IBM sighed with relief. Even IBM had created a version of UNIX for the
PC/XT called PC/IX. The emphasis on graphics gave them time to adjust.

About this time, MIT and Project Athena (who had been funded by IBM, DEC,
and HP), was working on a graphical interface. By 1989, X11R3 was stable
and working. Ironically, Windows 386 was still unmarketable.

By 1991, Windows 3.0 was out, but so was X11R4, Motif, and SunView. Gates
realized he had made a big mistake (much of what he needed to make Windows
work was available in X11/Motif. The PC makers feared that people would
buy X-Terminals instead of the more expensive PCs. Microsoft's
implementation prevented the use of Windows over remote lines.

When Linux started to surface, it went from being almost invisible in
1992 (probably less than 10,000 users in March), to a blip in 1993
(about 100,000 users by December), to a wave in 1994 (nearly 500,000 users).

Gates saw that Linux was a big threat in early 1994. Linux was already
capturing much of the server market he had hoped to get with NT 3.51 Server.
Even worse, people were upgrading from Linux on a 386 to a Sun SparcServer.
Microsoft promoted Windows 95 or "Chicago" like it was the last operating
system Microsoft would ever make. It almost was the last, running over
budget, 6 months late, and suffering from strained relations with vendors,
Microsoft barely delivered Windows 95 in Microsoft fiscal year 1995 (The
year ended August 31 and 95 was released August (15? or 29?).

It was a race Microsoft absolutely had to win. They limped across the finish
line inches ahead of Linux. It wasn't enough. Linux sales were still
increasing at a rate of 15%/month - Microsoft's indicator was stand-alone
copies of MS-DOS 6.22, which users would install in "Bare Bones" machines
to create the Boot floppies for Linux and to drive the "Dos Emulator".
Linux was still small, probably less than 1 million users, but it was
emerging. Microsoft announced NT 4.0 Workstation and Server. By the
time NT 4.0 was released, Linux had over 3 million users and was still
growing at a rate of nearly 15%/month, with surges following the
announcement of new distribution releases.

Microsoft used every chip it could. It even resorted to selling an
"NT 4.0 Academic Edition" to jobbers who would sell them at Computer fairs
for $120/copy. By the end of 1997, Microsoft wasn't even selling "standard
edition" Windows 95. To make things worse, Microsoft was loosing bargaining
leverage with the OEMs and many of the Fortune 500 companies.

In the last month, Microsoft has seen the beginning of the defections. A
new computer maker is creating a computer called the VarStation. The
low-end machine with Pentium 266 runs $1600 with Linux pre-installed. For
another $200, you can buy Windows NT as well. Meanwhile, Microsoft has
tried to bolster it's bargaining leverage by filing piracy lawsuits against
vendors of "Acedemic Edition", "OEM Edition", and "Upgrade Edition" versions
of NT 4.0 and 95. They are trying to protect their OEM market and prevent
further erosion of the price structure that OEMs pay to Microsoft.

> So that's some of the lore, and I've deliberately tried to
> relay it as lore rather than make a strong argumentative case for it
> because this is about the background of the prejudices,

Prejudice is when you attempt to make a judgment prior to investigation.
Extreme prejudice is "Contempt Prior to invstigation".

If, after weighing the available information, you form a strong opinion,
that is merely an informed opinion or an informed judgement. To act
consistant with the judgments you have made based on as much information
you can gather in a practical manner, is wisdom. To act consistent with
the evidence this far, and to be open to new information is also wisdom.

To act inconsistantly with your informed judgement is foolishness. To
refuse to consider new information before or after making an informed
judgement is ignorance. To be indecisive until you have considered
all of the evidence is dangerous. Stand between the to rails of the train
tracks until the train comes. Don't make any snap decisions until you
are absolutely certain that there really is a train, it really is
travelling down the tracks you are standing on, and that it will not stop
before it hits you. The only way you can be certain of all of these
things is to actually stand on the tracks until the train hits you.

Now, try the same exercise - with your eyes closed. Have someone else
tell you every reason why you should not get off the tracks. Since he
has nothing no loose (and possibly something to gain) if the train hits
you, what are the chances that he will be a reliable source of information?

Many managers in the corporate world are like the blind man standing between
the tracks, listining to the man who will be the beneficiary of his life
insurance policy telling him to stand right where he is. NT administrators
are very good politicians. When the manager gets canned because results
don't get produced, the NT administrator is promoted to the manager's
position.

> and I want to
> be clear that I'm not calling them anything but prejudices, even if
> they are prejudices which I most sincerely share in, against
> Microsoft.

If you have no evidence other than that provided by the Microsoft advocates
in your organization, and that provided by the Microsoft Sales Rep, and
that provided by media that relies on Microsoft for 90% of it's funding
(Microsoft ads and Microsoft Co-OP), you are like the blind man on the
tracks. If you only use UNIX and only read c.o.l.a. you are also the
blind man. Get and use both Linux and NT for at least 30 days each.
THEN you will be capable of generating an informed judgement.

I have used both, each for over a year now. Today, I know on which
side of the tracks I prefer to stand. Put simply, if my manager
said "Start an NT box and a Linux box, choose which one you think
will run the longest before failing, if you are wrong, you will lose
your job" - I choose Linux.

On the flip side. If my boss told me "Set up two workstations with
Office Suites, Linux and NT, choose which one your director will use
to communicate with his VP. At this moment, I'd still fudge for NT.
Only because there is a good chance that the VP might send my director
Microsoft Word97 document via Lotus Notes. My director would never
get the mail.

If I could choose for the Director AND the VP, I'd give them both Netscape
Communicator, StarOffice, and a dual-boot NT/Linux machine set to run Linux
by default.

If they asked me which I would prefer for the people working at the customer
service terminals, I would definitely choose Linux. Have you ever called
a travel agent or service rep and been told they can't help you because
their computer is down?

> People like me can't understand how these miserable
> companies can be so successful and have been searching for an
> explanation. So we figure they are big because they came on the scene
> early, just like we thought the QWERTY keyboard succeeded when it shouldn't
> havem, and I was ready to embrace this new economic theory of
> "path dependency". OK, if these explanations are no good, then what
> does explain it?

People resist change. They are addicted to being right, they fear the
unknown, and they remember all the "new things" they tried that didn't
work out. To overcome that resistance, there must be a payoff. Directory
Assistance switched to Dvorak because the CWA wanted an extra $1/hour
for QWERTY trained typists. By taking untrained typists and putting
them on DVORAK, they discovered superior performance.

The day a big insurance company like Prudential or Aetna decides that
Dvorak might lower the incedence of RSI or aid in healing those with
RSIs' and decides to reduse the health insurance rates by $2/year/person
if they use the Dvorak instead of QWERTY, or the day someone wins an
RSI lawsuit against their employer because their employer didn't tell
them about Dvorak, potentially reducing their risk of injury, is the
day that every company in America will put out a memo describing how
to switch from QWERTY to Dvorak using the Windows Control Panel/Keyboard
select Input Locales tab and select Add then select "English (United States)
Dvorak"... the 10 second process will become a standard notice.

Many will ignore the notice - which is fine because that absolves the
company of liability for RSIs related to typing. Others will observe
the notice, type faster, and be more productive with less stress.
Even if they do get RSIs, the overall expense is lower because there
was more work produced in less time. In addition, only the minority
will actually make the change. Even if 30% make the change, that
reduces the liability profile from 100% to 30%.

> Are there compelling reasons that us poor old engineers just
> don't understand, mysterious marketing and business advantages that
> Microsoft exploits, and that IBM exploited before them, that make theirs
> truly the superior products.

Much of American business depends not only on product quality but also on
the relationship one has established with the customer. If your broker
has established a certain level of trust with you and guided you into some
really good investments, he can call you with a hot tip and you won't check
the facts, you'll just take the tip. Even if he gives you a few turkeys,
all he really has to do is make sure that you win more often than you loose.
Brokers use this to cover high-risk positions. If the deal goes south,
the customer may loose a little money. If the deal goes through your still
a satisfied customer.

IBM was excellent at delivering what they said they would. When they
damaged their customer relations with Microchannel, OS/2 2.0, and
MVS 4.0, all in a period of less than 2 years, it nearly cost them
the company. The customers were still polite and interested, but
were actually looking elsewhere for alternatives.

Microsoft is in about the same bag. Windows 3.1 stayed around way too
long and people got really sick of GPFs and Reboots. NT 3.51 was too
little, too late. Windows 95 was too big, unreliable, and expensive.
Windows NT 4.0 had a very rough start. NT 4.3 (service release 3) was
the best OS Microsoft has ever produced. To those whose sole experience
is with Microsoft are pleased as punch and ready to put it into roles
as production servers. Those who have used UNIX are having nightmeres
imagining their mission critical servers that cost millions of dollars
to go down 3 minutes/year being turned over to a system that is down
for 4-8 minutes/week.

> "Ease of use" for instance. Linux is supposed
> to be hard to install and learn compared to Windows95 or Windows NT.
> If that's so, then I think a lot of it is due to "path dependency".

True. The "familiar path" is that OEMs send the drivers to Microsoft
who debugs them and installs them on the distribution CD-ROMs. The
core drivers ared created for the chip makers (just like Linux), and
then "branded" for the individual brands. For example, there are
about 9 cards based on the S3-Virge chip. Microsoft will "create"
branded drivers for the Stealth, Zoom, Accelerator...whatever.

Microsoft recognizes the importance that others place on their branding.
Linux actually got open support from 3Com because Becker described the
cards using their 3Com card numbers instead of the chip numbers. The
same thing has happened with the NE2000 compatibles in RH 5.0.

Linux has become a valuable trademark in it's own right. In 1996, people
didn't want to be associated with the "Hacker Operating System". Today,
Red Hat and Caldera are starting to create the brand Identification and
quality that others want to be associated with. I've seen a few peripheral
card that are listing "Linux Compatible" as one of their features. It
has become very important for the sale of after-market cards.

The funny joke is that many video card vendors use Linux and Xfree to
debug their card drivers before porting to NT and 95. There is enough
"common code" between the two to make it practical.

> When
> somebody builds a new scanner, or video card or sound card, they as a
> matter of course, write a software driver for Microsoft to go with it
> and include the diskette in the package. They don't do that for
> linux.

Again, much of this is a branding issue. Many of these scanners can use
the ppa SCSI driver module that is flagged as the "Zip Drive" interface.
Someone sent me a list of drivers for scanners and many were simply
the ppa interface with a different driver.

There seems to be a push among those who distribute Linux pre-installed
workstations (there are about 3 companies that I know of) to configure
with SCSI cards and drives. Linux can take advantage of the performance
features of SCSI, actually looking even faster than NT on the same hardware
(almost 10x faster).

> Somebody has to find the specs (if they can, sometimes these
> are considered proprietary secrets) and write a driver for linux
> themselves.

Again, there is a shift. Chip and Card makers have been pushed around
and given the runaround by Microsoft for several years now. It's refreshing
to have an outlet that can simply use the existing cards. Some card
makers may start creating and shipping binary Linux "modules", which
do not have to be compiled into the core Linux kernel.

> That's not innate superiority of Microsoft, that's just
> being there first and already established, path dependency.

Again, Microsoft has shifted from the "Cute Little Nerd with the
nifty operating system" to "The Blood-sucking vampire hell-bent on
driving us out of business". The average price of a PC has dropped
from $5000 in 1994 to under $10000 in 1997. In that time, Microsoft's
percentage (via license royalties) has actually increased from
$60 in 1993 to $200 in 1994. The total package has increased from
nearly $120 to $500, on a $1200 computer.

> If somebody came to me 5 years ago and said, "I'm gonna bet
> the farm on this new company I'm starting, what kind of computer
> system should I use?" I would probably have said, go with Microsoft
> and intel style PCs, anything else is too risky if you're betting the
> farm, much as I would have hated mouthing those words.

Actually, in 1993, I was asked by several hundred publishers what system
to use for getting on the Internet (it wasn't even "The Web" back then).
At the time, I suggested Slackware Linux. I told them that they could
get started with a very inexpensive machine and when they outgrew it,
they would have enough revenue coming in to upgrade to a Sun server.

Nearly every publisher who took my advice has become incredibly successful.
This includes several companies that did also start using NT. This
included the New York Times, Wired magazine, Yahoo, a couple hundred ISPs,
InJersey, and so on. In many cases, shortly after they went public
Microsoft purchased enough stock to influence the board into a forced
switch to NT (Microsoft has a major portion of it's assets in equities).
Yahoo and Lycos actually put both servers on-line and eventually pulled
NT 3.51 servers because they were so poor. Many use NT as the "front end",
and use UNIX as the back-end. Nasdaq for example uses several NT front-ends
to claim "NT powered", but the NTs which are non-critical and redundantly
routed, connect to a back-end UNIX system for the actual mission critical
data.

Several ISPs offer hosting on both NT and Linux. The prices for the Linux
system are often around $100/month (1 year minimum). The prices for NT
are over $500/month (1 year minimum). Some actually want the NT bill
"up front" because they have so many who decide to switch back to Linux.

> But isn't that
> a manifestation of this discredited path dependence eonomic theory?
> Now, (Oh Joy!) I might, with a clear conscience, convince the person to go
> with linux.

YESSSSSSS!

> Maybe linux is to Microsoft what the Japanese car makers
> were to Detroit. The newly evolved species, able to perfect it's
> superiority in an isolated little environment till it can come forth.
> A few years ago, when I told people I used linux I got blank stares.
> Now, very often, people have at least heard of it.

Remember, in Japan, cars were such a luxury that you often drove a motorcycle
for years before owning a car. Honda put the motorcycle engine into a very
light 4-wheel frame and created a car that was much easier to afford (still
a big luxury in most of Japan).

Linux went directly from being a laboratory exeriment into one of the most
demanding environments imaginable - Connected directly to the Internet
via a T1 frame-relay link. The early versions were buggy, but the bugs
were very quickly fixed and Linux developed a reputation as the "90 day
wonder" because it would run without stopping 24/7 handling traffic
from the U.S., Asia, Australia, India, and Europe, without a hitch,
until it would be stopped for about 15-20 minute for the next software
upgrade.

> I actually wonder if linux (and/or Free-Net-OpenBSD and/or HURD, sorry
> guys for not mentioning you sooner) might not be a new economic
> phenonemon made possible only in this information age and a reaction
> to the hobbles of 'intellectual property'?

Fundamentally, there has been a big shift. Effectively, Bill Gates, and
nearly every corporate Microsoft Advocate, was telling the most brilliant
programmers around the world "My bugware is worth $80 billion, and your
rock-solid superior software isn't worth the 40 minutes it would take to
load it into my PC".

There is a trend toward getting rid of corporate programmers entirely
and having full-time "Managers" simply be wined and dined by consulting
firms and independent software vendors, only to choose Microsoft software,
because that's the "company policy". More and more, managers expect
to be entertained and amused - a trip to Orlando Florida in the Winter,
or a trip to Vermont or Aspen in the summer, to "evaluate" technology,
only to come back reccomending the Microsoft product because it's
politically correct. Microsoft often asks companies to send the best
people to Seattle for a week or two. The companies think that they are
going to get a "scoop", some insider information that will give them a
competitive edge. What is really happening is Microsoft is looking for
the people who will be given the "grand tour" at the end of the week.
The people Microsoft plans to lure into saying "I wish I could work here".
The people Microsoft will hire, but only after getting as much information
as possible from their employers. In many cases, they are looking for
consultants. Not the folks who make megabucks, but the guys who will
work 80 hours/week for 8-14 months, only to be assigned to "hell" at
the end of the project.

> In the past, if somebody
> came up say with a cool way to soup up his hot rod, he could tell his
> friends, maybe visit their garages and help them tinker with their
> engines, or publish his idea in a magazine, but he couldn't just give
> his souped up engines to people for free. You can do that with
> software though, and it is done, often.

Don't you love running a "hot Rod PC"?

> ------remove "UhUh" and "Spam" to get my real email address---

Check out my "Linux advocate's Handbook"
http://www.access.digex.net/~rballard/cola

Rex Ballard
http://www.access.digex.net/~rballard

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading

[William T Wilson]

On 23 Feb 1998 cpw...@Spamrahul.net wrote:

> pause. Apparently there's not so much evidence to support the
> contention of what's being called 'path dependence' as I thought, and

I simply think it's that no one thinks about path dependence that much.

> the QWERTY phenomenon, which I'd seen mentioned many times and
> believed, is an urban legend. Dvorak and other style keyboards had a
> fair trial against QWERTY and were not shown to make typists much more

Then why are all the world's typing records held by Dvorak typists? Why
are the differences in speed between Qwerty and Dvorak statistically
significant? You might say that it's simply that the best typists prefer
Dvorak, rather than Dvorak making them better typists; but why would they
choose it if it weren't better? I doubt from simple elitism.

> productive. Other examples of supposed inferior products locking out
> better products, like VHS triumphing over Beta, also don't bear up to

This was an attempt by Sony to wield more marketing power than they had.
Apple was guilty of the same thing with the Macintosh. VHS, which is
(marginally) inferior than Beta, triumphed because Sony wouldn't license
their Beta technology. When JVC, Phillips, Mitsubishi, Sanyo, and
everyone else is making VHS VCR's and only one company is making Beta,
sheer numbers will squash Beta unless it's EXTREMELY superior. MS-DOS,
which is (significantly) inferior than the Macintosh (or at least, it was
in the mid-1980's) failed (or failed to become dominant) because of
exactly the same reasons. Look at IDE vs. SCSI. The computer industry
now has "ultra-DMA" IDE controllers, which are almost as fast as SCSI
cards, lack most of SCSI's flexibility, still limited to two devices per
interface and no simultaneous commands, and COST MORE than modern SCSI
adaptors. But because "IDE is the hard drive standard" computer
manufacturers feel compelled to force it to do something which it is
decidedly unable to do properly, because ten years ago cost was much more
important in a drive than performance (as everything was single-tasking
then anyway).

> that I hoped to be able to retire before Microsoft took over the
> world. He said I had about 2 years (maybe it was 3, I don't
> remember). All of his clients wanted Windows NT. I had never met his

Clients only want Windows NT because of that 'path dependency' you
mention. It's notoriously unreliable over the long term, however. IBM is
long gone from having any influence in anything except their legacy
systems (and the inexhorable march of time is scheduled to take an
unusually active role in derailing them further, in a little under two
years).

> clients, but I've got my prejudices to fall back on, I imagined them
> as Dilbert style managers, nervous nellies whose thought process went
> something like "Microsoft big, Microsoft must be good" but that's path

The actual thought process is remarkably similar to this.

> code and support it. (Is it a prejudice or unfounded urban legend
> that Microsoft has driven worthy companies out of business? An

This is true. Although in most cases they simply buy small companies who
have a product that suits their needs, the case of DR-DOS vs. Windows 3.1
comes immediately to mind. Microsoft inserted code (this code has been
located and identified by competent third parties) which detected whether
the system was running anything other than MS-DOS, and refused to run if
the computer was running DR-DOS. Now Microsoft can exterminate the
obviously superior DR-DOS.

Also consider the case of Intuit, whose Quicken product is clearly
superior to MS Money. Microsoft wished to simply buy Intuit and be rid of
them. The courts saved Intuit from MS strongarm tactics, and Quicken
remains much more popular than Money.

> linux, that he'll concoct a claim that linux is violating copyrights
> or something. I'm not sure how paranoid the average Linuxer is

Regardless of paranoia, it's false. MS has no legal standing to challenge
any product which does not use MS code. MS has no legal standing, in
fact, for most of the license terms you see on their packages. But they
put them there anyway, figuring they can get away with it.

> From what I understand, IBM was the biggest name in pre-computer style
> business machines, sorters, card-readers, that kind of thing. It was

As well as the rest of the computer industry, until Unix at the high end
and clones at the PC end showed IBM that quality really is important.
Some of IBM's modern systems (AIX) run rings around the AS/400's and
mainframes that IBM used to concentrate on, in terms of design. They did
an admirable job with Deep Blue. But the fact is, "If it's small enough
to fit in one room, IBM doesn't understand it." People know that now, and
that's why Windows rules the world instead of OS/2...

> managers again. Then there's this other super-successful giant, Intel.

I have different attitudes toward Intel than Microsoft or IBM. Intel is a
'kindler, gentler monopoly'. In the case of Intel, unlike MS or IBM, it
is not clearly demonstrable that the competition is superior. Yes, Intel
does a lot of stupid monopolistic things. But their chips are also in a
number of respects superior than Cyrix or AMD. Their chips are NOT
superior to the PowerPC or Alpha. But they are not far behind.

Intel stays on top by innovating and continually producing better
products. Bill Gates can go on and on all he wants about how MS can't be
all bad if computers double in capacity every year, but Intel made it all
possible...

> I was an assembly language programmer for years and people hated the
> instruction set of the 8x86 style computers, but they were forced down

x86 simply inherits a lot of garbage. If you compare the 8086 to the
other chips of its day (68000, Z80) it's not any worse. Very comparable,
in many ways. Intel is simply stuck with an antique design (which they
are working as hard as they can to get rid of).

Yes, the PowerPC throws it all away and starts over. (Just like Linux
does, in a number of cases). Yet Apple isn't gaining any ground on
Wintel. Why? Because IBM and Apple are the ones selling the PowerPC
chip, and Apple saddles it with their crappy MacOS and system designs.
What good is a PowerPC that's 50% faster than a comparable pentium, when
Apple puts it in a computer with 80ns RAM, an 8 MHz bus, and an OS that
doesn't have protected memory or preemptive multitasking, and then charges
more for it? No wonder only the True Believers buy it.

[cpw...@spamrahul.net]

I'm responding to stuff Grinch wrote. Grinch never mentions
Microsoft, but I'm going to mention it a lot. WAIT! DON'T QUIT OUT Of
this thread yet! I'm not going to try waste bandwidth actually trying
to rehash all the old arguments about why Microsoft is so bad. And I see
Grinch in my mind's eye saying to himself this person is arguing from
invalid assumptions, so his conclusions will be useless so I won't waste
my valuable time trying to set him straight anymore. But that's my
mind's eye's view of the Grinch, who knows what the real Grinch will do.
If I did somehow convince Grinch that Microsoft's stuff
was lousy, then Grinch might consider them an example of market failure.
If he convinced me that Microsoft wasn't lousy, then I wouldn't be scratching
my head over why they are so successful.
I think Microsoft didn't get where it is in an honest way. By that
I mean that I think that if people had had more information about what was
going on, they wouldn't have flocked to Microsoft. Which would mean
the market failed, and I'm trying to figure out what went wrong.
I'm trying to step back from the single instance of Microsoft to see
if there's a more general kind of phenomenon at work that will help me
understand. The debate over QWERTY vs dvorak for instance seems like it
might be another instance, and that's what got my interest.
In my initial posting, I tried to combine several issues in one
thread, speculation about biological and economic phenomena being explainable
by some meta systemic theory, where the actions were isomorphic, and
questions about whether Microsoft's continued success was due to path
dependence (the new buzz word I'd picked up).

oldn...@mindspring.com (Grinch) wrote:

|-and...@pimlott.student.harvard.edu (Andrew Pimlott) wrote:
|-
|->On 23 Feb 1998 02:21:14 GMT, cpw...@Spamrahul.net <cpw...@Spamrahul.net>
|->wrote:
|->
|->[snip]
<the "|->>" lines are some of my words from my original post>
|->>thought sprang from reading an article in the New Yorker while in a waiting
|->>room that mentioned Microsoft and a paper written by Brian Arthur about how
|->>a company can 'lock in' a market.
|-
|-There's a rather tough critique of that article at Slate.
|-http://www.slate.com/Dismal/98-01-14/Dismal.asp
|-
|-And here's the brouhaha that followed.
|-http://www.slate.com/Features/Krugman/Krugman.asp
...<snip>...
|-
|-There's no similarity at all between "path dependence" in economics
|-and the "lottery of life" in evolution. The essence of "path
|-dependence" in economics is *market failure*.
|- I read _Wonderful Life_ and it was interesting enough, but I don't
|-remember Gould writing anything in it about market failure in the
|-Cambrian explosion. Economics and biology are taught in different
|-academic departments for a reason.
|-
...<snip>...
|- The idea that slight, random changes in initial conditions have
|-unpredictable large-scale consequences was examined in economics by
|-Alfred Marshall 100 years ago when he was doing his work on positive
|-returns.
|- Let's not confuse century-old knowledge with some clever new
|-insight.

Actually, it sounds like Marshall anticipated chaos theory. There
seem to have been many thinkers in different disciplines who anticipated
chaos theory but until it was formalized, people kept making the same
mistakes, like trying to figure out long term weather prediction which they
would have known they couldn't do if they'd understood chaos theory. But
then, it turned out that chaos theory was inter-disciplinary, so maybe
there's something going on that's inter-disciplinary between economics
and evolution. You imply something like that happened before when
Darwin read Malthus, why not again?

|-
|-The point of the "path dependence" argument in economics is *not* that
|-some slight historical accident can have major consequences later,
|-"locking in" the behavior of millions of people in regard to, say,
|-QWERTY keyboards. That's obvious.
|-
|-The point of the economic path dependence argument is that the "lock
|-in" force is so strong that it somehow prevents economic actors from
|-improving their welfare through normal market activity -- that they
|-are somehow "locked in" to a *poorer* state of being and *cannot
|-escape*. In other words, market failure.
Yes, to me it seems that Microsoft achieved dominance somehow,
and the force is so strong that it "prevents economic actors from improving
their welfare through normal market activity" (finding something better
than Microsoft) that people are 'locked in', and "market failure" sounds
descriptive of the situation.
|-
|-For anyone who's seriously interested in this subject, here's a good
|-professional article explaining it: "Path Dependence, Lock-In, and
|-History" from the Journal of Law, Economics and Organization,
|-http://wwwpub.utdallas.edu/~liebowit/paths.html (It contains a nice
|-analysis of the VHS/Beta story.)
|- You'll notice the lack of references to Gould or other
|-biological-science sorts of fellows. You'll also note that while it
I wasn't asking who *doesn't* mention a possible connection but
who *does*, and if nobody does, then so be it.
...<snip>...
|-
|- The naive believers in path dependence (QWERTY-myth believers)
|-make the mistake of assuming that because the trivial type of "lock
|-in" is easily seen, it is economically significant. And they overlook
|-the obvious fact that if history should "lock in" the economy to a
|-really costly standard or practice, then really *strong* incentives
|-are created to break it. The more costly the "error" of history, the
|-*greater* is the reward for fixing it.

I agree this is the 'normal' condition. But now I claim there is
a non-normal condition, and that's what needs explaining, the exception.

|- That's why old standards and practices are abandoned all the
|-time, and why it's so hard to find much *measurable* economic cost of
|-path dependence -- in spite of the "new" insights of biology students.

I'm not a biology student, and I'm not claiming that an insight
from biology is what brought about the path dependence concept in economics.
I merely was struck by a similarity and speculated that there might be a
connection. I did not claim there had to be a connection. But I still think
it's worth thinking about some more before being dismissed.
You seem to be giving the whole "path dependence" theory short-shrift.
And you don't think there is economic significance. I noticed, in my lurking,
a lot of posts about a guy named Julian Simon who I never heard of but who
seems to have been a staunch advocate of the value of people resources. If
Microsoft is as bad as I think it is, with all those people struggling to use
it's products, all those engineers working on them, all those
dollars spent developing Windows NT, that's erosion of people as resources,
just like plowing straight furrows up a hill causes erosion of soil.
I know from experience and from talking to other engineers that
sometimes their best creative work disappears down a rat hole, and that's not
good for the spirit, and the better the engineers, the more they are
likely to suffer from it when it happens to them. So if I'm right (and
you probably think I'm wrong, judging by the tone of your post) that's
a big economic expense. Just as QWERTY versus Dvorak is a big expense if
the Dvorak party is right. I seem to have killed a thread in sci.economics,
"Myths of Economics" when I posted not a study, but some old news postings
from 1990 by people who had actually switched to dvorak and were sharing
their experiences. (If any whippersnappers want to see what newsheaders
were like in 1990, with all those !ucbvax!mit-edu!.... type paths, maybe you
can find it on deja-news or something)
<the "|->" lines below are from Andrew Pimlott, who responded to my
original post>
|-
|->The people who led you to believe that path dependance is a failed economic
|->theory have tricked you into taking a very narrow view of the evidence. By
|->a broader view (which allows for common-sense interpretation, by the way),
|->path dependance is much more plausible.
|-
<and Grinch replies...>

|-Do please explain the broader "common sense" interpretation.
|-
This is a purely rhetorical question isn't it? I think you
explained the common sense interpretation yourself; you just don't buy it.

<the "|->>" lines below are me in the original post>
|->>the QWERTY phenomenon, which I'd seen mentioned many times and
|->>believed, is an urban legend. Dvorak and other style keyboards had a
|->>fair trial against QWERTY and were not shown to make typists much more
|->>productive.
<I wrote that after reading Liebowitz and Margolis but before digging
into my archives for postings I vaguely remembered about Dvorak, then
Andrew response to my disillusion:>
|->
|->This argument has been advanced by those who need it for their pet theory.
|->If you really believe that the Dvorak keyboard is anything but clearly
|->superior, read http://www.ccsi.com/~mbrooks/dvorak/dissent.html.
|-
|-Oh, sheesh. This link leads to one howler after another. I'll point
|-out a few at the end of this post.
Yeah, I've noticed that one can find web page after web page
on each side of the argument, very much like reading comp.os.linux.advocacy.
<Here's me in the original post again:>
|->> If somebody came to me 5 years ago and said, "I'm gonna bet
|->>the farm on this new company I'm starting, what kind of computer
|->>system should I use?" I would probably have said, go with Microsoft
|->>and intel style PCs, anything else is too risky if you're betting the
|->>farm, much as I would have hated mouthing those words. But isn't that
|->>a manifestation of this discredited path dependence eonomic theory?
<Grinch replies:>
|-No.
|-Wouldn't the computer system you choose depend on your needs? For
Well, one 'need' is for products from a company that will be in
business for awhile, even if that company makes a lousy product. So why are
they going to be in business for awhile even though they make a lousy product?
That's what I'm trying to find out. Maybe it's "path dependency". Of course,
you can say it's not a lousy product and we can just shout back and forth
at each other, "is so lousy" "is not lousy". And this is maybe why we'll
always be at odds.

...<snip>...
|- What's path dependence got to do with that? Unless you say that
|-*every* company that has *any* kind of competitive advantage anywhere
|-is benefiting from path dependence, which would be trivializing the
|-whole idea out of existence.
And maybe the phenomenon is not path dependence but something else.
I was thinking about this sort of cognitive dissonance we have. Maybe
Microsoft succeeds because, like food for dogs and babies, the people who
byy it (managers) aren't the people who have to use it. People have it
in their homes because it comes already installed on their computers, but
is it already on their computers because of a market failure?
Maybe it's just a skew, and the market is recovering now and in a surprising
and novel way, with free products. At first the free unix like systems
could not compete because they were harder to get started in. But now
they are getting easier to set up (market forces for a 'free' product,
should be an interesting study for an economist), and they seem to be
catching on more and more. But I still fear that Microsoft will do something
to crush the movement using it's whatever-I-was-naming-path-dependency-
advantage to stamp it out. I suppose John D. Rockefeller was exploiting his
whatever-I-was_misnaming-path-dependency advantage to crush the competition
to Standard Oil until the Government stopped him. Of course, he had a
monopoly and Microsoft isn't a monopoly so it can't be quite the same, and
it isn't. Microsoft has to use it's muscle in more sophisticated ways
than Rockefeller did.

...<snip>...
|-
|-Again missing the fundamental point about QWERTY and standards. Some
|-arbitrary keyboard layout must naturally exist as a standard. It might
|-as well be QWERTY as any other, such as IXGLOT or PORKME, so long as
|-no other is demonstrated to be *measurably better*.
|- The fact that QWERTY itself may be no better than many other
|-possible layouts does *not* support the argument for path dependence
|-in economics, because as long as no other standard is measurably
|-better than it, the QWERTY standard carries no economic cost.
Well, just as I can say "Microsoft is lousy" and you reply "No it
isn't", I can say "QWERTY is lousy" and you can say "It's as good as Dvorak".

<Grinch argues against the web page that argues against Liebowitz
and Margolis and then lists other studies besides the Navy study>

|-
|- []A 1953 Australian Post Office study failed to show any benefit
|-versus QWERTY.
|- []A 1978 study at Oregon State University indicated that after 100
|-hours of training, Dvorak typists were up to only 97.6% of their old
|-QWERTY speed.
|- []A 1973 study at Western Electric found that after 104 hours of
|-training on Dvorak, typists were all of 2.6% faster than they had been
|-on QWERTY. (Of course, after 100+ hours of training you'd expect
|-QWERTY typists to be a lot better at QWERTY.)
|- []A. Miller and J. Thomas, of the IBM Research Laboratory, writing
|-in the International Journal of Man-Machine Studies, concluded that
|-"no alternative has shown a realistically significant advantage over
|-the QWERTY for general purpose typing."
|-
|- The Urban Legend of QWERTY/Dvorak no doubt will live on in the
|-wider world. But as far as this newsgroup is concerned, let's bury
|-once and for all the myth that Dvorak is a superior keyboard that has
|-been "suppressed" by QWERTY through "path dependence".
|-
|- BTW, if you like the Dvorak keyboard, go out and buy one. Nothing's
|-stopping you. ;-)
|-
Actually, I am thinking very seriously about doing that very thing.
It may be the only way to get at the truth. You cite studies, but I have
seen testimonials from people who switched. Which is more convincing? In
general, an objective study is probably more convincing. The people who
tried dvorak and didn't get an improvement probably didn't bother to post
an anti-testimonial. With the studies it depends on one's faith in one's
own ability to evaluate their merit or one's trust in someone else to evaluate
for them.
If I try dvorak and really like it then it will lower the credibility
of economists for me but then my faith in economists (or lack of it) is not
something that any economist is going to lose sleep over.

|-
|-Regards
|-
Regards
-----Remove "UhUh" and "Spam" to get my real email address----

[James Youngman]

>>>>> "William" == William T Wilson <flu...@benatar.dunadan.com> writes:

William> Then why are all the world's typing records held by Dvorak
William> typists? Why are the differences in speed between Qwerty
William> and Dvorak statistically significant? You might say that
William> it's simply that the best typists prefer Dvorak, rather
William> than Dvorak making them better typists; but why would they
William> choose it if it weren't better? I doubt from simple
William> elitism.

Are there any "ergonimically shaped" Dvorak keyboards?

[SUSUPPLY]

The thread title is a paraphrase from "Julius Caesar" (that's for those of you
who are Harvard students). The next line is borrowed from a Nobel laureate in
economics, George Stigler (who was responding to a congressman) and is directed
at the recent post by cpwUhUh:

"It is impossible to say anything polite about such analysis."

The gentleman (or woman?) wrote:

> If I did somehow convince Grinch that Microsoft's stuff
>was lousy, then Grinch might consider them an example of market failure.
>If he convinced me that Microsoft wasn't lousy, then I wouldn't be scratching
>my head over why they are so successful.

You have missed the point of the "grinch's" argument completely. He didn't
argue that Microsoft wasn't lousy, and he probably doesn't care, because it is
irrelevant to "path dependence". In the 1960's the Instamatic camera was a
huge commercial success, even though there were many superior cameras, such as
Nikon. This is not an example of market failure. It is the opposite, just as
the fact that Toyota outsells Mercedes, or that people buy more hamburger than
filet mignon are more examples of markets working, not failing.

>I mean that I think that if people had had more information about what was
>going on, they wouldn't have flocked to Microsoft. Which would mean
>the market failed, and I'm trying to figure out what went wrong.

Nothing went wrong. Information is not free, and the laws of supply and demand
apply to information just as they do to anything else. There is an excellent
book that deals with this concept: "Knowledge and Decisions" by Thomas Sowell.
I am sure you will not avail yourself of the insights in it.

> <and Grinch replies...>
>
>|-Do please explain the broader "common sense" interpretation.
>|-
> This is a purely rhetorical question isn't it?

No, it is a straightforward calling of Mr. Pimlott's intellectual bluff by the
"grinch", just as I did when I told him (Pimlott) to consider himself "pressed"
for an explanation.

> And maybe the phenomenon is not path dependence but something else.

You may be on to something here!

>At first the free unix like systems
>could not compete because they were harder to get started in. But now
>they are getting easier to set up (market forces for a 'free' product,
>should be an interesting study for an economist), and they seem to be
>catching on more and more.

Do you really think so? You need to do more reading in economics and less
talking. Economists have studied such things. By the way, "path dependence"
theories implicitly deny that such "free products" strategies can work. What
you have done here is to change horses in midstream.

> But I still fear that Microsoft will do something
>to crush the movement using it's whatever-I-was-naming-path-dependency-
>advantage to stamp it out.

You mean like Microsoft was able to stamp out "Quicken" or AOL?

> You cite studies, but I have
>seen testimonials from people who switched. Which is more convincing? In
>general, an objective study is probably more convincing.

I'm sure the scientific method appreciates your backhanded compliment.

> but then my faith in economists (or lack of it) is not
>something that any economist is going to lose sleep over.

Bingo.

Sincerely,
Patrick

[SUSUPPLY]

Gary Forbis wrote:

>I remember a story told several years ago about how one city's bus system
>was primarily electric trollies on rail but was bought out by GM and Shell
>whereupon they replaced the trollies with busses and ripped out the rails.

Politicians' edicts are not examples of market failure, you might want to
acquaint yourself with some of the writings of "public choice" economists.

What you have done with your example is to bring up a rather glaring problem
with "path dependence" theories: If decentralized decision making (by the
many) can't get consumers what they want, why would centralized decision making
(by the few) do so.

>isn't it possible for a concerted effort
>on the part of parties with deep pockets to establish a less than optimal
>standard which then behaves as if a second degree path dependency?
>If such a situation is found to exist would it not fit the author's
>definition of
>a third degree class dependency?

Have you read anything(!) either the "grinch" or I have written on these
threads?

Sincerely,
Patrick

[jim blair]

Grinch wrote:
>Grinch:

>A talk "What Economists Can Learn From Evolutionary Theorists", given
>by a noted economist to the European Association for Evolutionary
>Political Economy, is at http://web.mit.edu/krugman/www/evolute.html

>Teaser: It says some nasty things about Stephen J. Gould.

>Regards

Hi,
Thanks for the interesting link. I think that I expressed some
of the same ideas in the Bio-QWERTY post, now on my web page
QWERTY file. While K thinks of climbing hills and seeking
the highest peak, my physics background led me to think of
seeking the lowest potential energy well. Same idea (if you
just turn everything up-side-down ;-)

Then I read this comment (about maximization and equilibrium):

"Now some people would say that this kind of creation of useful
fictions is a thing of the past, because now we can study complex
dynamics using computer simulations. But anyone who has tried
that sort of thing - and I have, at great length - eventually comes
to realize just what a wonderful tool paper-and-pencil analysis
based on maximization and equilibrium really is. By all means let
us use simulation to push out the boundaries of our understanding;
but just running a lot of simulations and seeing what happens is
a frustrating and finally unproductive exercise unless you can
somehow create a "model of the model" that lets you understand
what is going on."

This reminded me that many years ago, back in the late 1960's
or early 70's, analog computers were as popular as digital ones.
I think the analogs have since disappeared since they were much
less useful for most purposes. But they were very useful for
a situation of analyzing coupled differential equations: the
sort of thing you get from a model of many interacting things,
where the behavior of each is influenced by the others. Like
for example an economy of many firms, or an ecology of many
species.

While digital computers can simulate the situation for any set
of initial conditions and specified parameters, it become very
time consuming to run simulations for all possible sets of
parameters, to decide for example which of the various possible
solutions is "best" or most stable--the local "minimum" (or
as Krugman would say "maximum") problem.

But this was much easier and faster using an analog computer:
you could sweep through an infinite number of parameter values
by simply rotating a variac from 0 to 100%, and watch for any
spikes in output. Then set that variac on the various peaks
while sweeping the variacs that represented other parameters.
Still took some time for complex systems, but I think it was
much faster than trying to use digital computer models.

Anyone out there remember analog computers? Could it be time to
revive them for these kind of problems?

Also note that my Bio-QWERTY post web page version is slightly
different from the posted version.

"....a marsupial can't decide to convert to a mammal"
has been replaced with "...a marsupial cannot decide to
convert to a rodent or primate"

I wrote the original without a reference book and when I checked
later I had forgotten that marsupial is a sub order of mammal (like
rodent
and primate) and not a separate order in the Chordate phylum.

Is this reverting back to my earlier days as an absent-minded college
professor, or is it a forerunner of future Alzheimer's??

--
,,,,,,,
_______________ooo___(_O O_)___ooo_______________
(_)
jim blair (jeb...@facstaff.wisc.edu) For a good time call
http://www.geocities.com/capitolhill/4834

[Robert Vienneau]

Some economists who have considered connections between economics
and evolution:

o Thorstein Veblen ("Why is Economics Not an Evolutionary Science")

o Alfred Marshall

o F. Hayek

o G. Hodgson (who, I believe, is active in the European Association for
Evolutionary Political Economy <http://eaepe.tuwien.ac.at/Welcome.html>)

o Some contributors to a book edited by Philip Mirowski called _Natural
Images in Economic Thought: Nature Read in Tooth and Claw_

There's a lot of interesting work on how history matters in economics that
is not necessarily explicitly connected to biological metaphors, for example,

o Paul Davidson's interpretation of Keynes' GT as being at home in a
non-ergodic environment, unlike the classical special case assumption
of ergodicity

o Richard Goodwin's investigation of chaotic and non-linear dynamical
systems

o Those building on Schumpeter to understand innovations and firms.

Path-dependence relates to more than individual products. Think of how
city layouts are altered by decisions about the use of cars or light
rail and subways. Or how whole related industries of gas stations, materials,
and part manufactors, etc. are affected by a late 19th century/early 20th
century initial dominance of gasoline-powered cars over electric. These
issues have been discussed in the theory of economic development for a
long time, e.g. by Hirschman.

You might also be interested in work coming out of the Sante Fe
Institute. For example, Stuart Kaufmann's _At Home in the Universe_
is a popularization of theoretical biology (which reminds me of
the theory of computing automata).

I don't claim to understand any of this stuff in any depth. Economics
is just a hobby for me.

--
Robert Vienneau Try my Mac econ simulation
r game, Bukharin, at
v
i ftp://csf.colorado.edu/econ/authors/Vienneau.Robert/Bukharin.sea
e m
n o Whether strength of body or of mind, or wisdom,
@ c or virtue, are always found...in proportion to
d . the power or wealth of a man [is] a question
r e fit perhaps to be discussed by slaves in the
e p hearing of their masters, but highly unbecoming
a a to reasonable and free men in search of the
m c truth.
s -- Rousseau

[Wundergeist]

SUSUPPLY wrote:

>If decentralized decision making (by the
> many) can't get consumers what they want, why would
> centralized decision making
> (by the few) do so.

Hi.

Take the prisoner's dillema. Centralized decision making, in the sense
that they commit to a joint strategy, as opposed to decentralized decision,
in which they submit their strategies independently, is better. This does
not mean that *any* centralized decision would be better (which is the usual
non-sequitur made by proponents of state intervention). This means that
*in certain cases, notably when network externalities or free-riding problems
are present* THERE EXISTS a centralized solution that is better than the
decentralized solutions.

IMNSHO, most implementations of centralized decision fare WORSE than the
decentralized decision. Even in the cases when there exists one Pareto-
dominant centralized decision. Existence proofs are not guides to policy.

WG

--
http://web.mit.edu/camoes/public/

[cpw...@spamrahul.net]

susu...@aol.com (SUSUPPLY) wrote:

|-The thread title is a paraphrase from "Julius Caesar" (that's for those of you
|-who are Harvard students). The next line is borrowed from a Nobel laureate in
|-economics, George Stigler (who was responding to a congressman) and is directed
|-at the recent post by cpwUhUh:
That's me.
|-"It is impossible to say anything polite about such analysis."
I'm glad you recognize some of your own limitations, not being able to
be polite that is. You like to borrow other peoples witty rejoinders
I see, the way Beatrice heard that Benedict claimed she got her wit
out of the Hundred Merry Tales. (That's from "Much Ado About Nothing"
for you Harvard Students), appropriate for this little sub-thread.
... <snip> ...
Words from my earlier post are on the "|->" lines.
|-
|-> but then my faith in economists (or lack of it) is not
|->something that any economist is going to lose sleep over.
|-
|-Bingo.
Oooh, Oooh, turning my own words against me. I am so stung!
Are you a Professor?
I really wonder why you bothered to post at all. Of course, I'm
responding, so I can't climb on my high horse about that too much.
|-
|-Sincerely,
|-Patrick

Regards, and no hard feelings,
Carl (I'm a 'he').

------Remove "UhUh" and "Spam" to get my real address---

[Ray Van Tassle]

In article <real.email-ya024080...@news.mit.edu>,
real....@my.webpage (Wundergeist) wrote:

One major problem with centralized decision-making is the way the
incentives run. In general, the central decision maker is not effected by
the consequences of his decision, so he has little incentive to make a
good one. Witness the 1930's famine in Russia and 1950's famine in
China. Stalin and Mao made the decisions, but they were not the ones who
starved to death.

>
> WG
>
> --
> http://web.mit.edu/camoes/public/
--
To email me, remove the antispamfilter from my address.

[SUSUPPLY]

Wundergeist wrote:

>> Take the prisoner's dillema. Centralized decision making, in the sense
>> that they commit to a joint strategy, as opposed to decentralized decision,
>> in which they submit their strategies independently, is better.

Sorry, you are confused. That is not centralized decision making, that is
voluntary cooperation--see Axelrod's "The Evolution of Cooperation" for
example--which is the essence of markets (i.e. decentralized decision making).

Ray Van Tassle is correct to point out that the problem is one of incentives.
That was the point of my question.

Patrick

[SUSUPPLY]

Robt. Vienneau wrote:

>Path-dependence relates to more than individual products. Think of how
>city layouts are altered by decisions about the use of cars or light
>rail and subways.

This may be an interesting topic, but it is not what we are discussing. City
layouts are not determined by decentralized decision makers, quite the
contrary, they are usually the products of politicians.

> Or how whole related industries of gas stations, materials,
>and part manufactors, etc. are affected by a late 19th century/early 20th
>century initial dominance of gasoline-powered cars over electric.

You are on target here, and that is bad news for the" path dependencists" (such
as the fellas at the Santa Fe Institute). Why would anyone build a gas station
if there are no cars. Why would anyone buy a car if there are no gas stations
or replacement parts suppliers or repair shops? If "path dependencists" were
correct we'd still be riding horses.

Patrick

[Gary Forbis]

SUSUPPLY wrote:

Well, this argument may fool creationists but not rational people.

Flammable liquids were in use prior to the invention of automobiles. There were
some very early electric cars but the energy density of oil distillates over
batteries and
their greater availability helped make the choice.

Are you conceiving of "path dependency" in a way no rational person would accept?
As I read the link grinch gave the development of the auto and related industries
is an example of second class path dependeny. Only if electricity was known as
the optimal choice and it was not chosen would it be a third class path dependency
(and this is the kind denied by the link's author.)

[SUSUPPLY]

Gary Forbis wrote:

>Well, this argument may fool creationists but not rational people.
>
>Flammable liquids were in use prior to the invention of automobiles. There
were some very early electric cars but the energy density of oil distillates
over batteries and
>their greater availability helped make the choice.
>
>Are you conceiving of "path dependency" in a way no rational person would
accept?

>As I read the link grinch gave the development of the auto and related
industries is an example of second class path dependeny. Only if electricity
was known as the optimal choice and it was not chosen would it be a third class
path dependency (and this is the kind denied by the link's author.)
>

Gary,

Could you try rewriting the above, I'm trying to understand your point, but I'm
not sure I do.

Patrick

[Chris Beggy]

This Wednesday's Wall Street Journal had a front page article disputing
the positive evaluation of the Dvorak keyboard. The author went on to
examine the effect this has on the claims of the theory of
path-dependence
for adoption of technology.

Chris

[Wundergeist]

SUSUPPLY wrote:

> Ray Van Tassle is correct to point out that the problem is one of incentives.
> That was the point of my question.

That is also the reason why I stated that, while in some cases there exists
one centralized solution that Pareto-dominates independent maximization, I
believe that [probably all or at least] most centralized decisions in practice
(i.e. made by one agent, be it the CEO or the government, rather than by a
omniscient social planner) are WORSE than the independent decisions (or
market decisions).

In fact there are two problems. The incentive problem, of course, is
immediate. The second problem is that of information. A centralized
decision should optimize joint utility. Unfortunately there is no
device or trick that can reliably read people's utility directly.
The market solves that by allowing each agent to make her own tradeoffs.

WG.

--
http://web.mit.edu/camoes/public/

[cpw...@spamrahul.net]

First of all, oldn...@mindspring.com (Grinch) wrote:
>A talk "What Economists Can Learn From Evolutionary Theorists", given
>by a noted economist to the European Association for Evolutionary
>Political Economy, is at http://web.mit.edu/krugman/www/evolute.html

I think he (or she?) wrote it quite awhile back in this thread
but I overlooked it somehow. Now I read it and thanks Grinch, if
you're still following this thread, it was interesting and
stimulating. I might discuss it further in a different thread. (I
started THIS thread with a enough baggage already.)
This whole thread has been really interesting even if
unconclusive. There's something going on with the Microsoft
phenomenon and I've decided that economists just can't account for it.
They really should go lurk in c.o.l.a. (comp.os.linux.advocacy) some.
It might open their eyes a little. This pheonomen is just too complex
and subtle for economists to understand and model and theorize about
now. Krugman in the paper at that website Grinch mentioned commented
about how economists may be too dependent on certain convenient
simplifications for instance and maybe "path-dependence" as it is
understood is inadequate or incorrect. The Free software movement,
GNU, linux, Free/Net/OpenBSD is also an exciting part of this mix, and
perhaps an indicator of how nanotechnology will be addressed in the
next generation, when the power to manipulate matter as well as
information will accessible to the spiritual descendents of the people
writing that software now, and what their relationship will be with
the Microsofts of that era.
It can't be as simple as Microsoft is winning because of some
classic model of providing the best service on some level vs linux.
And truly, I don't believe QWERTY is dominating over dvorak just
because the two are nearly equal in merit either. But I've tried
linux (and had enough exposure to Microsoft to be privately convinced
of it's inferiority) but I haven't tried Dvorak so that's something to
do. Grinch had listed some very impressive sounding pro-QWERTY
studies at the end of one post. I replied that I had testimonials but
agreed more or less that his list of studies seemed more convincing.
Patrick (susu...@aol.com SUSUPPLY) seemed miffed that I wasn't
completely cowed into submission by the list but I admit I was daunted,
and still am daunted, and intrigued by all these mixed signals about
their merit which do seem eerily like the mixed signals about Microsoft
versus the free unix like systems.
One thought I had about the keyboard thing is that maybe the
studies are concerned about how to produce a 'professional' typist. I
do about 25 wpm touch typing and that's fast enough for me but I am
aware of a feeling of awkwardness when I type, and fingers getting
ahead of each other and out of order. I speculated that maybe when
the training is really disciplined as for a pro 80 wpm typist, this
effect is over-ridden or something so that qwerty or dvorak or
whatever don't make much difference but that the people who give
testimony that dvorak helped them are, like me, mostly casual typists
looking for comfort.
Then William T Wilson <flu...@benatar.dunadan.com> made a post
indicating the world's typing records are held by Dvorak typists which
would indicate it isn't just the casual typists who benefit. (though
if dvorak only gave a 1% advantage at the high end, that might be
enough to cinch who the record holders would be, while not looking
impressive enough in some general study.) Guess I have to find out
more about those records. I'll try dvorak. There are some websites
which say it's easy to switch the keyboard and there are online
tutorials, I can take a stab at it and if it works well for me maybe
I'll go track down some of those studies to see what they actually
say.

-----Remove "UhUh" and "Spam" to get my real email address---
You know, this little advisory about my email is not automatic,
not from a .signature file or anything like that. I have
remember to type it in every time. (Those of you who like
to take potshots any chance you get, fire at will.)

[Gary Forbis]

SUSUPPLY wrote:

> Could you try rewriting the above, I'm trying to understand your point, but I'm
> not sure I do.

I'll try.

Grinch supplied the link http://wwwpub.utdallas.edu/~liebowit/paths.html and I
am using it as a starting place for the discussion. Early on it says:

In this paper we identify three distinct forms of path dependence.
Two of these forms -- which we define as first- and second-degree
path dependence -- are commonplace, and they offer little in the way
of an objection to the neoclassical paradigm. Only the third and strongest
form of path dependence significantly challenges the neoclassical paradigm,
and as this paper shows, the theoretical arguments for this form require
important restrictions on prices, institutions, or foresight.

The author defines the three degrees thusly:

First degree path dependency: " instances in which sensitivity to starting
points
exists, but with no implied inefficiency."
Second degree path dependency: "sensitive dependence on initial conditions
leads to outcomes that are regrettable and costly to change" based upon
imperfect information at the time of the decision.
Third degree path dependency: "path dependence, sensitive dependence on initial
conditions leads to an outcome that is inefficient -- but in this case the
outcome
is also 'remediable.'"

The author follows the definition of third degree path dependency with: "That is,
there exists or existed some feasible arrangement for recognizing and achieving a
preferred outcome, but that outcome is not obtained." It is the "or existed..."
area
I am exploring since the author has already accepted the first two degrees

QWERTY is at least a first degree path dependency though it may be a second
degree path dependency.

Many factories are second degree path dependencies as are most home computers
and home computer architectures/instruction sets.

The arguments against third degree path dependencies seems to rely upon the
notion that there is some aggregate optimal or preferred outcome. I don't see
why this need be true and blieve it is not.

Years ago an eighteen year old went to Sears with an improved socket wrench
latching mechanism. Sears encorporated the mechanism without compensating
the inventor. After many years of litigation Sears took the improvement off
the market rather than pay for it. The patent should expire in the near future.
We will see if the improvement is reintroduced on a more global scale.

[SUSUPPLY]

To Gary Forbis:

Gary,

Thank you for taking the time to try to clarify your remarks. However, I’m
going to return to your earlier post. You asked me if I was "conceiving of
‘path dependency’ in a way no rational person would accept?".

The answer is, no. I’m conceiving of it in exactly the same way that Liebowitz
and Margolis do (they are the authors of the article you are relying on as a
"starting place for the discussion"). Their article, "Path Dependence,
Lock-In, and History" is a more rigorous (i.e. scholarly) treatment of the
issue we have been discussing. While their style is appropriate for a
professional audience, I think it is neither necessary nor desirable for those
of us on this newsgroup to adopt it. They did contribute a less rigorously
formulated piece to Reason Magazine that can be accessed online
(www.reasonmag.com). See the section "Is Microsoft a Menace", then click on
the article that is described as something like "2 economists discuss the
QWERTY typewriter".

If you go back to their article ("Path Dependence, Lock-In, and History") you
will see that they clearly state that (what they call) 1st and 2nd degree path
dependency has no relevance to what is claimed by the likes of Brian Arthur and
Paul David for a "revolutionary reformulation of the neoclassical paradigm."
In fact, in their fifth paragraph they write:

"When things that are different are grouped together and treated as things that
are similar, error is assured." Several paragraphs later there is this:

"Third-degree path dependence is the only form of path dependence that
conflicts with the neoclassical model of relentlessly rational behavior leading
to efficient…outcomes."

In other words (what they call) 1st and 2nd degree path dependency is
irrelevant for the issue under consideration by us. 1st degree would be
something as commonplace as the short-run concept of Fixed vs. Variable costs
familiar to any first year accounting student. Path dependence as we have been
discussing on this newsgroup is an issue of (possible) market failure. Their
concepts of 1st and 2nd degree path dependence are actually examples of market
successes. But don’t take my word for it, deep into their article they write:

"Arthur’s version of path dependence is in the third-degree form."

As to your example of Sears and the improved socket wrench, I am not sure what
to say. My memory of the incident is that the inventor won a lawsuit and was
awarded a significant sum of money. You state that Sears, "after many years of
litigation...took the improvement off the market rather than pay for it." That
suggests that the "improvement" may not have been much of an improvement (i.e.,
not worth what they would have had to pay for it). Also, if they could take it
away after several years of selling it, then there must not have been any
"network effects" to begin with, and therefore has nothing to do with the
discussion we have been having.

Sincerely,
Patrick

[Michael Turton]

>The arguments against third degree path dependencies seems to rely upon
the
>notion that there is some aggregate optimal or preferred outcome. I don't
see
>why this need be true and blieve it is not.

People who study the history, sociology and philosophy of technology don't
believe so either. "Efficiency" is a value. The view outlined above has no
overt definition of technology (is it artifacts? artifacts +
organizations? artifficats + organizations + cultural and social
behaviors?).

Let's throw out some examples of "third-degree path dependency." There are
millions, especially when cultures clash. Take Western and Eastern ships.
Chinese junks were superior to Western ships in about every aspect of ship
technology you could name during the age of sail. They were more
seaworthy, could sail closer to the wind, were more stable, etc. Many
western innovations were copied from the Chinese (copper sheathing,
compartmentalization) but other Chinese innovations were ignored. For
example, Chinese ships were built with the thickest part aft of the
centerline, which is the faster and more stable configuration, but western
vessels did not adopt this. Or take wheelbarrows. Many different ones
were in use in China, but only one design got adopted into the west. Why?

How about the way rice growing has reshaped the landscape of Japan, which
has now been tightly adapted to that crop. Yet other crops would be more
productive of calories, or less labor intensive, etc. But Japan hasn't
switched to "more efficient" major crops.

One could go on and on. Take city design in Taipei where I live.
Developers have been allowed to run riot, and have predicatably produced an
urban nightmare. Since buildings are made of concrete and brick,
uninsulated, implementing energy efficiency has been effectively ruled out
should the nation have to save energy. All the buildings would have to be
torn down.

Mike

[Matt Kennel]

On Thu, 26 Feb 1998 15:58:59 -0500, Wundergeist <real....@my.webpage> wrote:
:This means that

:*in certain cases, notably when network externalities or free-riding problems
:are present* THERE EXISTS a centralized solution that is better than the
:decentralized solutions.

:IMNSHO, most implementations of centralized decision fare WORSE than the
:decentralized decision. Even in the cases when there exists one Pareto-
:dominant centralized decision. Existence proofs are not guides to policy.

Just as empirical examples of totalitarian Communist failure should not
be a guide to adopt fundamentalist laissez-faire policy, for the same reason.

--
* Matthew B. Kennel/Institute for Nonlinear Science, UCSD -
* "People who send spam to Emperor Cartagia... vanish! _They say_ that
* there's a room where he has their heads, lined up in a row on a desk...
* _They say_ that late at night, he goes there, and talks to them... _they
*- say_ he asks them, 'Now tell me again, how _do_ you make money fast?'"

[SUSUPPLY]

Mike Turton has just produced a disjointed post on this thread that has
absolutely nothing to do with the theory of "path dependence":

>People who study the history, sociology and philosophy of technology don't
>believe so either. "Efficiency" is a value. The view outlined above has no
>overt definition of technology (is it artifacts? artifacts +
>organizations? artifficats + organizations + cultural and social
>behaviors?).

>Let's throw out some examples of "third-degree path dependency." There are
>millions, especially when cultures clash. Take Western and Eastern ships.

>Chinese junks were superior to Western ships in about every aspect of ship....

and on and on.

Mike, the theory is one of economics. Not "the history, sociology and
philosophy of technology". The theory posits that increasing returns due to
network effects will lead to market failure. There are not millions of
examples, there are zero.

By the way, "efficiency" means something specific to an economist: An outcome
is efficient if no "improvement" will produce net benefits. Take note of the
"net", in the previous sentence.

Patrick

Sincerly,
Patrick

[SUSUPPLY]

I started this discussion on the thread "Myths of Economics". I wrote that
Orrin Hatch seemed to be the most recent dupe of the theory of "path
dependence". I was correct. On page A18 of today's (3-3-98) Wall Street
Journal, the Senator has an opinion piece ("Beware High-Tech Monopolies"). The
second paragraph is an almost perfect specimen of someone falling, hook, line,
and sinker, for the myth of "path dependence".

A little knowledge is a dangerous thing, especially in the hands of powerful
politicians.

Sincerely,
Patrick

[Gary Forbis]

SUSUPPLY wrote of "path dependencies":

> The theory posits that increasing returns due to
> network effects will lead to market failure. There are not millions of
> examples, there are zero.
>
> By the way, "efficiency" means something specific to an economist: An outcome
> is efficient if no "improvement" will produce net benefits. Take note of the
> "net", in the previous sentence.

I appreciate your effort but still fail to fully understand the claim.

What would count as a disproof of the theory? Sometimes I think the outcome is
taken axiomaticly as disproof of path dependency, that is it seems as if the
outcome
is assumed to be the most efficient.

Are you aware of just how much the QWERTY keyboard has changed? Just
as there are copyright laws governing type faces there are (were) copyright laws
governing keyboard layout. Just in the last twenty years the keyboard has
changed significantly. What ever happened to the cent sign? Why are there
twleve function keys rather than ten. Why do most new PC keyboards have
eight keys between the alphabetic and numeric keyboards and why are these
functions duplicated on the numeric keyboard? Even the placement of the tab
key and shape of the enter(return?) key has changed. The Alt and Ctrl keys
have come into existence, doubled in number, and even switched places with
each other.

What would count as a counter example to the proposed theory of path dependency?

[Roy.D...@exeter.ac.uk]

I am a bit surprised that the phenomenon of path dependence
should cause any controversy. The QWERTY keyboard may or may
not be a good example of it but it is not too difficult to
find examples involving a topic of much greater significance
in economics than typewriters - namely money.

We tend to assume that use of coins would precede banking
whereas the opposite is true. Banking was invented at least
a thousand years before coins. Even after the development
of coinage internal financial transactions in the Egypt of
the Ptolemies were carried on to a considerable extent by
credit transfers involving grain - the state granaries
functioned as banks - while coins were used mainly for
military purposes, e.g. payment of soldiers including
foreign mercenaries.

In contrast, although the Romans had banks they used
coins for many transactions that had been carried out
by banks in earlier civilisations because by the time
their banking system developed the use of coins was well
established.

Cutting notches in sticks sounds like a rather primitive
method of keeping records but for centuries this practice
was central the state finance in Britain. The size of
the notch indicated the amount of money involved. The
tally stick would then be split down the middle and one
half would be kept by the creditor and the other half
by the debtor. The tallies functioned as instruments of
credit and this system was so well established by the
end of the 12th century that it continued long after the
development of more sophisticated banking methods should
have made it obsolete.

Not until 1826 were the tallies abolished and a few years
later, in 1834 these redundant pieces of wood were gathered
up and used as fuel in the stoves of the House of Commons.
Unfortunately they proved much better for this purpose than
anticipated and burned so fiercely that Parliament was
razed to the ground!

A third example of path dependence concerns the development
of postal giro systems. The first modern giro was created in
Austria in 1883. Britain did not establish its postal giro
until 85 years later - in 1968. However all the conditions
for a giro existed in Britain long before the Austrian one
was created. Britain was the first country to develop a modern
postal service and the first to have a nationwide railway
network. However the banking system in Britain was also well
developed. It grew considerably during the industrial revolution
(which of course started in Britain) and the banks opposed
attempts by the Post Office to move into what they regarded
as their territory.

For more information on these and related topics see the book
_A History of Money from Ancient Times to the Present Day_
by by Glyn Davies, rev. ed. Cardiff: University of Wales Press,
1996. 716p. ISBN 0 7083 1351 5.

Also see my web site with the same title (the URL is given
below) and especially the chronology.

(In any email messages remove NoSpam and the number from my
address below)
_____________________________________________________________________
Roy Davies | e-mail Roy.D...@NoSpam2193.exeter.ac.uk |
University Library | History of Money URL |
University of Exeter | http://www.ex.ac.uk/~RDavies/arian/llyfr.html |
Stocker Road | |
Exeter EX4 4PT | Financial Thrillers URL |
UK | http://www.ex.ac.uk/~RDavies/arian/linda.html |
----------------------------------------------------------------------

--------------------------------------------------------------------
Posted using Reference.COM http://WWW.Reference.COM
FREE Usenet and Mailing list archive, directory and clipping service
--------------------------------------------------------------------

[SUSUPPLY]

Gary Forbis wrote:

>What would count as a disproof of the theory [of "path dependence"]?

Well, quite a lot of what I have written. But you are putting the cart before
the horse. The "path dependency" theorists are proclaiming that they are on
the cutting edge of a "new economics". It is up to them to provide some
empirical evidence to support their theory that can stand up to scrutiny. A
theory is not assumed to be true, just because someone hasn't disproved it.
Especially a radical theory like "path dependence".

>Are you aware of just how much the QWERTY keyboard has changed?....

> Just in the last twenty years the keyboard has
>changed significantly. What ever happened to the cent sign? Why are there

>twleve function keys rather than ten.....

Uh Gary, did you really mean to submit this as evidence of "path dependency"?
The theory posits "lock-in", remember.

>What would count as a counter example to the proposed theory of path
>dependency?

Over 200 years of scholarship in economics.

Patrick

[SUSUPPLY]

Roy Davies wrote:

>I am a bit surprised that the phenomenon of path dependence
>should cause any controversy. The QWERTY keyboard may or may not be a good

example....

The evidence is overwhelmingly against it.

>We tend to assume that use of coins would precede banking
>whereas the opposite is true. Banking was invented at least
>a thousand years before coins.

What does this have to do with increasing returns due to network effects
causing market failure?

>Cutting notches in sticks sounds like a rather primitive
>method of keeping records but for centuries this practice
>was central the state finance in Britain.

So does painting on large rocks (see M. Friedman's "The Island of Stone
Money"), but I repeat my earlier question.

> The first modern giro was created in
>Austria in 1883. Britain did not establish its postal giro
>until 85 years later - in 1968.

Okay, I'm curious. What is a postal giro? Although, since postal systems are
created through governments and are not the result of market forces, I repeat
for a third time: "What relevance does any of this have for the topic under
discussion.

Sincerely,
Patrick

[Michael Turton]

In article <19980303153...@ladder03.news.aol.com>,
susu...@aol.com (SUSUPPLY) wrote:

>
>and on and on.
>
>Mike, the theory is one of economics. Not "the history, sociology and

>philosophy of technology". The theory posits that increasing returns due
>to network effects will lead to market failure. There are not millions of
>examples, there are zero.

>By the way, "efficiency" means something specific to an economist: An
>outcome is efficient if no "improvement" will produce net benefits. Take
note of the "net", in the previous sentence.
>

>Patrick
>
>Sincerly,
>Patrick
>

How interesting. What is a "benefit?" And what is a technology? How are
"net" benefits calculated? Who determines all these things? And so on.

Mike

[Michael Turton]

In article <19980303205...@ladder02.news.aol.com>,
susu...@aol.com (SUSUPPLY) wrote:

>Okay, I'm curious. What is a postal giro? Although, since postal
>systems are created through governments and are not the result of market

>forces, I repeatfor a third time: "What relevance does any of this have

>for the topic under discussion.
>
>Sincerely,
>Patrick

I dunno, Pat, are organizations a form of technology? If so, is the
corporation a technology, and why doesn't it count as "path dependence."
Are problem-solving strategies technologies?

Your definition is tautological, Pat. If something has net benefit, it
will overwhelm competitors in the marketplace, but if it does, it must have
had net benefit, whatever that may be.

Pat, what is a technology, and how is net benefit determined?

Mike

Mike

[jim blair]

James Youngman wrote:
>
> Are there any "ergonimically shaped" Dvorak keyboards?.

Hi,

Look at

http://www.teleprint.com/

There is also a reference on my web page QWERTY file in the book review
section.

[Grinch]

We now are well into the "nonsense" mentioned in the title of the
thread.
__

mtu...@ms.showtower.com.tw (Michael Turton) wrote:

>I dunno, Pat, are organizations a form of technology?

Are fish a form of bird?

>If so, is the
>corporation a technology, and why doesn't it count as "path dependence."

A corporation is a technology ?
Do you have a dictionary of economics? Do words have meanings?

>Are problem-solving strategies technologies?
>
>Your definition is tautological, Pat.

No, his definitions are clear and in the literature. Yet nobody but
Pat in this thread seems to be aware that words have defined meanings.

The relevant URLs have been posted, perhaps six times. And yet
somehow everybody seems to believe that any accident of history they
see anywhere qualifies as economic "path dependence".

>If something has net benefit, it
>will overwhelm competitors in the marketplace, but if it does, it must have
>had net benefit, whatever that may be.

Exactly backwards.

"Path dependence" says that if something *does* have a measurable net
benefit it *won't* succeed in the marketplace.

In the case of QWERTY, the alleged benefit was a 2,000% return on
investment from retraining typists in Dvorak for companies with
secretarial typing pools -- namely, all big companies in the 1950s.
And yet these profit-seeking companies did not seek this
*measurable* benefit that would have added to their bottom line. This
was allegedly due to the pressure of "path dependence" -- which thus
became a newly recognized, somewhat mysterious force to be explored.

So you see, there's nothing tautological about it, and nothing unclear
about *net benefit* -- for a business the net benefit if profit on the
bottom line.

Now, the whole concept of "path dependence" was put forth in a 1985
paper by economist Paul David, entitled _Clio and the Economics of
QWERTY_.
But last week, in Wednesday's Wall Street Journal, Prof. David
admitted that he now views the QWERTY example to be false because
QWERTY provides no measurable benefit over Dvorak. Thus, there is no
"path dependence" to be explained here.

More to the point, Prof. David said that absent QWERTY there is *no*
example of meaningful economic path dependence *anywhere* in the real
world that he is aware of.

It's fair to assume, I think, that Prof. David is aware of Microsoft,
Sears tools, chinese junks, the history of money, and so on. So if
you people consider these to be examples of "path dependence" then
you're working with very different definitions than he his, and it's
his theory after all.

If any of you can think of a good example where where profit-seeking
businesses today are turning their backs on *real measurable profits*
for no good economic reason, just because they prefer to do things the
old-fashioned, less-profitable way (like sticking with QWERTY
keyboards when switching to Dvorak really would produce 2,000% ROI)
then I suggest you e-mail Prof. David at Stanford with your idea. I'm
sure he'll be happy to hear from you. You may get your name in the
textbooks for reviving path dependence.

But citing one example after another of historical happenstance isn't
going to do it:
*Historical contingency is NOT economic path dependence*.

If all that people are going to do in this thread from here on in is
cite random historical events and torture the meanings of words beyond
all recognition to assert the 'self-evidentness' of a theory when even
its *creator* says it's kaput, then have mercy -- let this thread die.

Path dependence has gone the way of cold fusion.

Let it go.

[Grinch]

Roy.D...@exeter.ac.uk wrote:

>Also see my web site with the same title (the URL is given

>below) and especially the chronology.

This Web site is a great resource on the really interesting subject of
the history of money. Highly recommended. But it has nothing at all to
do with economic path dependence.

Historical contingency is not economic path dependence, although
everybody certainly seems eager to confuse the two.

[Michael Turton]

In article <34fce32...@news.mindspring.com>,
oldn...@mindspring.com (Grinch) wrote:

>mtu...@ms.showtower.com.tw (Michael Turton) wrote:
>
>>I dunno, Pat, are organizations a form of technology?
>
>Are fish a form of bird?

Alas, people who study technology in all its myriad forms wonder about
organizations of things and people. Do technologies have to include
artifacts? I'm sorry if your unfamiliarity with the extensive literature
on the definitions of words like "technology" (a very slippery term, I
might add) has lead you think that you know everything. Since technology
is so clear and simple, give me your definition.

Mike

[Markku Stenborg ®]

On 4 Mar 1998 12:49:09 GMT, mtu...@ms.showtower.com.tw (Michael
Turton) wrote:

[snip]

> Alas, people who study technology in all its myriad forms wonder about
> organizations of things and people. Do technologies have to include
> artifacts? I'm sorry if your unfamiliarity with the extensive literature
> on the definitions of words like "technology" (a very slippery term, I
> might add) has lead you think that you know everything. Since technology
> is so clear and simple, give me your definition.

This is sci.econ NG, and the thread seems to be one of the few that
isn't X-posted all over the Usenet. This should give you a hint on
what types of issues are relevant here. Eg, one would expect that
words are used here as they are in the Science of Economics, not as in
SciFi, as for one instance. Also, one would expect that the nuances or
definitions of words and concepts not really crucial for their Econ
aspect can safely be ignored.

On the top of my head: Technology for Econ purposes is anything that
allows one to transform one bunch of goods to an other by applying
labor and capital.

--
Š Markku Stenborg
OFC & Turku Biz School
ROT13ed for the hell of it:
zne...@hgh.sv <- out-of-order for the time being
znexxh....@svabsp.sv

[Aaron R Kulkis]

James Youngman wrote:
>
> >>>>> "William" == William T Wilson <flu...@benatar.dunadan.com> writes:
>
> William> Then why are all the world's typing records held by Dvorak
> William> typists? Why are the differences in speed between Qwerty
> William> and Dvorak statistically significant? You might say that
> William> it's simply that the best typists prefer Dvorak, rather
> William> than Dvorak making them better typists; but why would they
> William> choose it if it weren't better? I doubt from simple
> William> elitism.

>
> Are there any "ergonimically shaped" Dvorak keyboards?

For what it's worth....

The first ergonomic keyboard I ever saw (1970's) was Dvorak....
--
Aaron R. Kulkis
Unix Systems Administrator

---------------------------------------------------------------------------
I speak for me, not my employer
---------------------------------------------------------------------------
"What's wrong with 3rd party tools? Especially if they are free?
What the hell do you thin unix is anyway? It's a big honkin' party
of 3rd party free tools." --Bob Cassidy (rmca...@uci.edu)

[Michael R. Ward]

Was the senetor from Wordperfect really duped, or is he serving his
constiuents? All the knowledge he needs is that his constituents would
be helped if Microsoft were harmed.

In a similar vein, I remember Senetor Paul Simon from Illinois defending
"free markets" for derrivatives, when they were blamed for the 1987
crash on Wall Street. His other stances are hardly consistent with a
deep faith in free markets. The fact that these derivatives were traded
in the pits in Chicago might better explain his stance.

Mike Ward
----------------------------------------------------------------
Michael R. Ward (217) 244-5667
Dept. of Ag. and Consumer Econ. wa...@uiuc.edu
University of Illinois http://www.uiuc.edu/ph/www/ward1

[Michael R. Ward]

[Gary Forbis]

Grinch wrote:

> The relevant URLs have been posted, perhaps six times. And yet
> somehow everybody seems to believe that any accident of history they
> see anywhere qualifies as economic "path dependence".

I'm trying to understand what is being claimed. I am using the definition
of"path dependency" given in http://wwwpub.utdallas.edu/~liebowit/paths.html
(which you gave earlier.) I guess I'm slow.

> "Path dependence" says that if something *does* have a measurable net
> benefit it *won't* succeed in the marketplace.

> More to the point, Prof. David said that absent QWERTY there is *no*

> example of meaningful economic path dependence *anywhere* in the real
> world that he is aware of.
>
> It's fair to assume, I think, that Prof. David is aware of Microsoft,
> Sears tools, chinese junks, the history of money, and so on. So if
> you people consider these to be examples of "path dependence" then
> you're working with very different definitions than he his, and it's
> his theory after all.

It's very likely normal people will interpret phrases within the context oftheir
experiences. "Path dependency" does not have the feel of technical
phrase only to be interpreted within the confines of a specific theory.

I watched yesterday's senate hearings concerning Microsoft and the internet.
Some interesting comments were made. It is generally acknowledged Microsoft
has achieved monopolistic proportions in the microcomputer OS world. It was
suggested OS prices are rising. Even though the OS is much more than the user
presentation, the user presentation is the layer users think about when they
think
about the OS. Microsoft is marrying their OS and internet user presentations.
Microsoft is using its OS domination to leverage its internet visibility and is
using
its internet visibility to fight innovation by its competition though pricing
mechanisms.
Bookmarks let the individual remember locations they think important and
preloading
these with the browser give certain locations advantages over others. Channels
also
give specific locations advantages. Since Netscape is competing against a
"free"
browser it has to find alternative funding. It has moved to a model where
internet
sites pay for channel access and recoup this cost though their own means.
Microsoft
has a policy to deny "free" channel access (and like its "free" browser, channel

access is being defined as part of the OS) to those who pay for access
elsewhere.

Others at the hearing said no moneys are being made available to compete
directly
against Microsoft because the risks are just too high. Microsoft is being given

a wide berth because few can afford development costs whose recovery depends
upon the sales of products Microsoft may bundle with their operating system and
so technological innovation is being stifled.

This sure looks like a market failure. If you don't want to call a failure due
to
"path dependency" then what do you want to call it?

> But citing one example after another of historical happenstance isn't
> going to do it:
> *Historical contingency is NOT economic path dependence*.

I'm eager to learn new phrases. "Historical contingency" and "path dependency"
seem synonymous at first glance. I guess their difference usage in economics
is an example of historical contingency.

[SUSUPPLY]

Michael R. Ward wrote:

>Was the senetor from Wordperfect really duped, or is he serving his
>constiuents? All the knowledge he needs is that his constituents would
>be helped if Microsoft were harmed.

Yes, I made the same point in my original post, that Novell is in Utah.
Politicians seem to be "path dependent" on the strategy of lying. Perhaps Mike
Turton could enlighten us as to whether or not that is a "technology".

I have recently read that Newt Gingrich is sharing his feelings with the people
of Iowa about how much he loves ethanol subsidies

Patrick

[Roy.D...@exeter.ac.uk]

On 3 Mar 1998 20:54:55 GMT, susu...@aol.com (SUSUPPLY) wrote:
> Roy Davies wrote:
>
> >I am a bit surprised that the phenomenon of path dependence
> >should cause any controversy. The QWERTY keyboard may or may not be a good
> example....
>
> The evidence is overwhelmingly against it.
>
> >We tend to assume that use of coins would precede banking
> >whereas the opposite is true. Banking was invented at least
> >a thousand years before coins.
>

> What does this have to do with increasing returns due to network effects

> causing market failure?
>
> >Cutting notches in sticks sounds like a rather primitive
> >method of keeping records but for centuries this practice
> >was central the state finance in Britain.
>
> So does painting on large rocks (see M. Friedman's "The Island of Stone
> Money"), but I repeat my earlier question.
>
> > The first modern giro was created in
> >Austria in 1883. Britain did not establish its postal giro
> >until 85 years later - in 1968.
>

> Okay, I'm curious. What is a postal giro? Although, since postal systems are
> created through governments and are not the result of market forces, I repeat
> for a third time: "What relevance does any of this have for the topic under
> discussion.
>

I was writing about evolution and path dependence. All three
examples above show that what exists at any given time depends
at least partly on what went before (hardly a controversial
point) and therefore the present state of affairs is not
necessarily optimal. That is the only point I was making.

A postal giro is a simple system of transferring money from one
person's account to another. Postal services reach everybody
(at least in developed countries) and have a much bigger
system of branch offices than any commercial bank. Therefore
the post office is well placed to offer money transmission
services cheaply. Because of the universality of postal
services accounts can be credited and debited at a single
national centre whereas because no single bank is as omnipresent
a single payment by cheque (or check to Americans) will
frequently involve transactions in more than one bank.

Most European countries and Japan have giro services. In some
countries they are no longer run by the post offices, e.g. in
Britain Girobank was sold to Alliance and Leicester bank (a
former building society or mortgage lending institution like
the American thrifts) a few years ago but its new owners have
a contract with the Royal Mail (or British post office) which
continues to provide the same services as it did when it owned
the giro. I believe that in Denmark the giro has merged with
a bank too.

The money transfer services provided by giros are in competition
with those of commercial banks and therefore market forces do
operate. The giros tend to be cheaper. Even for international
transfers, payment from an account in one giro to another,
e.g. from the British one to the Danish one, would normally be
quite a bit cheaper than the charges typically levied by
commercial banks - and the giros still manage to make a profit.

The most detailed book in English on this subject, despite
being nearly 25 years old) is still -

Davies, Glyn _National Giro : modern money transfer_
London : Allen and Unwin, 1973. ISBN 0-04-332054-6.

Regards,
_____________________________________________________________________
Roy Davies | e-mail Roy.D...@exeter.ac.uk |

University Library | History of Money URL |
University of Exeter | http://www.ex.ac.uk/~RDavies/arian/llyfr.html |

[jim blair]

Robert Vienneau wrote:

>Path-dependence relates to more than individual products. Think of how
>city layouts are altered by decisions about the use of cars or light
>rail and subways. Or how whole related industries of gas stations, materials,
>and part manufactors, etc. are affected by a late 19th century/early 20th
>century initial dominance of gasoline-powered cars over electric. These
>issues have been discussed in the theory of economic development for a
>long time, e.g. by Hirschman.

Hi,

Several have objected to this as unrelated to path dependence, but I
think it is.

Consider how often you have seen a situation where a series of decisions
have been made where each decision seemed perfectly logical (at the
time).
But the result of the series is not very good, and almost anyone can
suggest a better way than the final result. In many cases, the best
solution is to "go back to square one and start over".

I think this is related to the "local minimum" problem described in my
web page file "Does Path Dependence apply to Biology?" in the QWERTY
section of Book Reviews.

Sometime the "less than optimum" result is because of a change in the
conditions (as I think is the case with the QWERTY keyboard); but often
it is just the result of starting something before thinking through the
whole picture: the "painting yourself into a corner" syndrome.

And often (to return to Robert Vienneau's point), no single part of the
system can be evaluated except by reference to the total picture.
Or to try to get at this differently: the system cannot be understood
by looking only at the individual parts. Or it is more like ecology
than like physics.

[SUSUPPLY]

Gary Forbis wrote:

"It's very likely normal people will interpret phrases within the context of

their experiences. 'Path dependency' does not have the feel of technical

phrase only to be interpreted within the confines of a specific theory."

Beautiful! This is a discussion group about issues in economics, it is not
about feelings--maybe you're confusing the grinch with Bill Clinton? Would you
feel it proper to write about touchdowns and field goals in a discussion of
baseball?

"I watched yesterday's senate hearings concerning Microsoft and the internet.
Some interesting comments were made. It is generally acknowledged Microsoft
has achieved monopolistic proportions in the microcomputer OS world."

Gary, you need to learn to distinguish between assertions and facts.
Especially when the assertions are made by interested parties (i.e. Microsoft's
competitors).

"Others at the hearing said no moneys are being made available to compete
directly against Microsoft because the risks are just too high. Microsoft is
being given a wide berth because few can afford development costs whose
recovery depends upon the sales of products Microsoft may bundle with their
operating system
and so technological innovation is being stifled."

Gary, how do you suppose the following happened: With a true business giant,
IBM, dominating the computer industry, a couple of college dropouts found the
money to start Apple Computer and carve out a very profitable niche for their
company.

Then, when Apple was selling about $2 billion a year (both the hardware and the
software!) two more college dropouts found the money to start Microsoft, and
become even more successful.

"This sure looks like a market failure. If you don't want to call a failure
due to 'path dependency' then what do you want to call it?"

I would call it an example of "the interested sophistry of merchants and
manufacturers confounding the common sense of mankind"

Sincerely,
Patrick

[SUSUPPLY]

Roy Davies wrote:

>... what exists at any given time depends

>at least partly on what went before (hardly a controversial

>point)....

You're right, that is not controversial. And it is not what we have been
talking about either.

Thank you for taking the time to explain about the "postal giro". However, as
I suspected it is not apposite to this thread. Welcome to the club.

You wrote:

>The money transfer services provided by giros are in competition with those of
commercial banks and therefore market forces do operate.

The theory under discussion posits that market forces will not operate (not
successfully, anyhow). So if you've done anything here, it's to offer a
rebuttal of the theory of "path dependence". Though, as I said, government
agencies actions are irrelevant to this issue.

Sincerely,
Patrick

[SUSUPPLY]

Jim Blair wrote:

>Consider how often you have seen a situation where a series of decisions
>have been made where each decision seemed perfectly logical
>(at the time).
>But the result of the series is not very good, and almost anyone can
>suggest a better way than the final result. In many cases, the best
>solution is to "go back to square one and start over".

Yes I have. Such as my decision to start this discussion in the first place.

Patrick

[SUSUPPLY]

Mike Turton wrote:

>How interesting. What is a "benefit?" And what is a technology? How are
>"net" benefits calculated? Who determines all these things? And so on.

Are you the same Mike Turton who wrote:

"Chinese junks were superior to Western ships in about every aspect of ship
technology you could name during the age of sail. They were more seaworthy,
could sail closer to the wind, were more stable, etc."

Put aside the truth or falsity of the assertions, and ask yourself if the Mike
Turton who wrote the above had a grasp of "benefit", or thought he knew what he
meant by "ship technology"

As to " who determines all these things", well in a market economy (which is
the area under discussion) the consumers do. They are often aided by
professionals such as economists and accountants, in their decision making.

With little tolerance for sophistry,
Patrick

[Aaron R Kulkis]

jim blair wrote:

Hmm... ever see the movie Brazil...

all the typewriters are implemented with 1920's keyboards
coupled into modernistic flouresct+LCD screens...

just to point out a different path that even something as
simple as the typewriter could have gone down....

> Robert Vienneau wrote:
>
> >Path-dependence relates to more than individual products. Think of how
> >city layouts are altered by decisions about the use of cars or light
> >rail and subways. Or how whole related industries of gas stations, materials,
> >and part manufactors, etc. are affected by a late 19th century/early 20th
> >century initial dominance of gasoline-powered cars over electric. These
> >issues have been discussed in the theory of economic development for a
> >long time, e.g. by Hirschman.
>
> Hi,
>
> Several have objected to this as unrelated to path dependence, but I
> think it is.
>

> Consider how often you have seen a situation where a series of decisions
> have been made where each decision seemed perfectly logical (at the
> time).
> But the result of the series is not very good, and almost anyone can
> suggest a better way than the final result. In many cases, the best
> solution is to "go back to square one and start over".
>

> I think this is related to the "local minimum" problem described in my
> web page file "Does Path Dependence apply to Biology?" in the QWERTY
> section of Book Reviews.
>
> Sometime the "less than optimum" result is because of a change in the
> conditions (as I think is the case with the QWERTY keyboard); but often
> it is just the result of starting something before thinking through the
> whole picture: the "painting yourself into a corner" syndrome.
>
> And often (to return to Robert Vienneau's point), no single part of the
> system can be evaluated except by reference to the total picture.
> Or to try to get at this differently: the system cannot be understood
> by looking only at the individual parts. Or it is more like ecology
> than like physics.
>
>
> --
> ,,,,,,,
> _______________ooo___(_O O_)___ooo_______________
> (_)
> jim blair (jeb...@facstaff.wisc.edu) For a good time call
> http://www.geocities.com/capitolhill/4834

--

[Matt Kennel]

On 3 Mar 1998 20:38:28 GMT, SUSUPPLY <susu...@aol.com> wrote:
:Gary Forbis wrote:
:
:>What would count as a disproof of the theory [of "path dependence"]?

:
:Well, quite a lot of what I have written. But you are putting the cart before
:the horse. The "path dependency" theorists are proclaiming that they are on
:the cutting edge of a "new economics". It is up to them to provide some
:empirical evidence to support their theory that can stand up to scrutiny. A
:theory is not assumed to be true, just because someone hasn't disproved it.

So, do you assume that economy is really just like a system in full
equilibrium with a linear, homogeneous, 'utility function' so that travelling
in all degrees of freedom along the local gradient of "utility" actually does
reach the overall global maximum?

That's the theoretical justification for 'laissez faire', and as far as I can
tell, it hasn't been experimentally proven. I feel it is based upon an
entirely supurious analogy to equilibrium classical thermodynamics. In
physics, the thermodynamical laws are derived from underlying microsocopic
theory and apply only in some circumstances---they are not taken as axioms!

Some brief experience in minimizing large-degrees of freedom systems with
nonlinear and time dependent interactions will quickly disabuse one of the
notion that ''gradient descent'' is the pinnacle of achievement.

:Over 200 years of scholarship in economics.

:Patrick

--
* Matthew B. Kennel/Institute for Nonlinear Science, UCSD -
* "People who send spam to Emperor Cartagia... vanish! _They say_ that
* there's a room where he has their heads, lined up in a row on a desk...
* _They say_ that late at night, he goes there, and talks to them... _they
*- say_ he asks them, 'Now tell me again, how _do_ you make money fast?'"

[SUSUPPLY]

Mike Turton wrote:

>I dunno, Pat, are organizations a form of technology? If so, is the

>corporation a technology, and why doesn't it count as "path dependence."

Okay Mike, anything you say.

Now explain to me the how network effects of corporate ownership produce
increasing returns, and how corporate ownership is thereby "locked-in", to the
detriment of the superior form of ownership, which is ...(communism?).

Eagerly awaiting your reply,
Patrick

[Michael Turton]

In article <19980304180...@ladder02.news.aol.com>,
susu...@aol.com (SUSUPPLY) wrote:

>Michael R. Ward wrote:
>
>Yes, I made the same point in my original post, that Novell is in Utah.
>Politicians seem to be "path dependent" on the strategy of lying. Perhaps
Mike
>Turton could enlighten us as to whether or not that is a "technology".
>

No problem Pat, as soon as you can tell me how net benefit can be
calculated without reference to values.

Mike

[Robert Vienneau]

Markku Stenborg writes:

> On the top of my head: Technology for Econ purposes is anything that
> allows one to transform one bunch of goods to an other by applying
> labor and capital.

Markku,

You didn't answer all the questions. "Are organizations a form of
technology?" And why don't you formulate your definition so as to
be consistent with the well-accepted result in the "Science of
Economics" that capital is not an input?

In my opinion, the Neo-Pareto definition of "efficiency" supposedly
used by Liebowitz and Margolis is a norm that should be
of little interest to economists. Those who tend to worry about how
history matters in economics tend to reject that norm. For example,
consider my favorite Scandinavian economist - present company
excepted - Gunnar Myrdal and his use of the concept of "cumulative
causation."

Another defect of Liebowitz and Margolis' paper is their assumption
that economic agents can determine the underlying parameters of a
model generating observed patterns from a single realization.
Technological change is one place where one would expect non-ergodicity
to be important. Liebowitz and Margolis sort of acknowledge this, but
conclude that economists must rely on the special case assumption of
ergodicity on the grounds that non-ergodicity is supposedly
non-empirical. It's always suspect when a dubious principle of dated
philosophy of science is used to decide a substantive point.

So I find the framework in which Grinch thinks conclusive results have
been demonstrated not fully adequate for discussing the issue of path
dependence.

--
Robert Vienneau Try my Mac econ simulation
r game, Bukharin, at
v
i ftp://csf.colorado.edu/econ/authors/Vienneau.Robert/Bukharin.sea
e m
n o Whether strength of body or of mind, or wisdom,
@ c or virtue, are always found...in proportion to
d . the power or wealth of a man [is] a question
r e fit perhaps to be discussed by slaves in the
e p hearing of their masters, but highly unbecoming
a a to reasonable and free men in search of the
m c truth.
s -- Rousseau

[Markku Stenborg ®]

On 5 Mar 1998 11:33:15 GMT, rv...@see.sig.com (Robert Vienneau) wrote:

> Markku Stenborg writes:
>
> > On the top of my head: Technology for Econ purposes is anything that
> > allows one to transform one bunch of goods to an other by applying
> > labor and capital.
>
> Markku,
>
> You didn't answer all the questions. "Are organizations a form of
> technology?" And why don't you formulate your definition so as to
> be consistent with the well-accepted result in the "Science of
> Economics" that capital is not an input?

I thought that I did formulate that to be consistent with the idea
that labor and capital are distinct from inputs? [I didn't think this
taxonomy -- input, capital, labor, output -- really is a result?]

Bunch of Technology Bunch of other
goods or apply Labor get goods or
Inputs Capital Outputs

So, organization *could* be seen as technology, if usefull for some
purposes, as labor buzzing about and capital laying around would not
do this transformation as efficiently as some organized processes
would.

Equally well, one could make a distinction b/w tech and organization:

Technology
Inputs --> L, K --> Outputs
Organization

I have no idea what difference this distinction would make on anything
of any relevance to any econ issue.

Umm, while we're defining terms, What do you mean by "organization"
here, and why does it matter for something within econ that some stuff
is called organization and some other stuff is called something else?

[Haven't payed that much attention to this thread, so I don't recall
what this Liebowitz and Margolis' stuff is about.]

--
© Markku Stenborg

[Robert Vienneau]

In article <350196ad...@news.eunet.fi>, real.a...@bottom.of.msg
(Markku Stenborg Ž) wrote:

> > You didn't answer all the questions. "Are organizations a form of
> > technology?" And why don't you formulate your definition so as to
> > be consistent with the well-accepted result in the "Science of
> > Economics" that capital is not an input?
>
> I thought that I did formulate that to be consistent with the idea
> that labor and capital are distinct from inputs?

Markku,

Yes, I misread you. I tend to get suspicious whenever economists
talk about "capital," and I still think your further explanation
is confused. But I don't think I need to bore our audience with my
ideas on capital here and now.

I was more interested in whether you would say "organizations" are
part of what economists mean or should mean by "technology." I think
we are agreed, the answer is "perhaps."

[..]

> Umm, while we're defining terms, What do you mean by "organization"
> here, and why does it matter for something within econ that some stuff
> is called organization and some other stuff is called something else?

Since my question was quoted from a prior poster, I don't need to answer.
I do think a weakness of mainstream economics is the lack of a
coherent and well-established theory of the firm.

[Robert Vienneau]

A demonstration that the usual mainstream economists use of "efficiency"
is unfounded can be found at <http://www.wau.nl/wub/wep/nr9705/wep05.htm>.
This is an internal critique.

American institutionalists, such as Marc Tool, are an important school
of thought in economics who insist that economics should be explicitly
value-based. They tend to draw on the pragmatic tradition in philosophy,
as represented, for example, by John Dewey. Institutionalists would not
find Grinch's reference convincing.

[George Avery]

Markku Stenborg ® <real.a...@bottom.of.msg> wrote in article
<35015c6c...@news.eunet.fi>...

> On 4 Mar 1998 12:49:09 GMT, mtu...@ms.showtower.com.tw (Michael
> Turton) wrote:
>
> [snip]
>
> > Alas, people who study technology in all its myriad forms wonder about
> > organizations of things and people. Do technologies have to include
> > artifacts? I'm sorry if your unfamiliarity with the extensive
literature
> > on the definitions of words like "technology" (a very slippery term, I
> > might add) has lead you think that you know everything. Since
technology
> > is so clear and simple, give me your definition.
>
> This is sci.econ NG, and the thread seems to be one of the few that
> isn't X-posted all over the Usenet. This should give you a hint on
> what types of issues are relevant here. Eg, one would expect that
> words are used here as they are in the Science of Economics, not as in
> SciFi, as for one instance. Also, one would expect that the nuances or
> definitions of words and concepts not really crucial for their Econ
> aspect can safely be ignored.
>

> On the top of my head: Technology for Econ purposes is anything that
> allows one to transform one bunch of goods to an other by applying
> labor and capital.

Herbert Simon, who is extremely relevant to the topic of this thread, as
his work on organizational structure, and March's "Garbage Can" model are
forebears of the complex systems work, which includes Arthur's models,
specifically defined the structure, rules, taboos, etc. of an
organizational entity as its "technology." Of course, at the time the
management types didn't like cybernetics because it was "economics", and
the economists rejected it for the opposite reason...

[George Avery]

Gary Forbis <for...@accessone.com> wrote in article
<34FD9847...@accessone.com>...

>
>
> Grinch wrote:
>
> > The relevant URLs have been posted, perhaps six times. And yet
> > somehow everybody seems to believe that any accident of history they
> > see anywhere qualifies as economic "path dependence".
>
> I'm trying to understand what is being claimed. I am using the
definition
> of"path dependency" given in
http://wwwpub.utdallas.edu/~liebowit/paths.html
> (which you gave earlier.) I guess I'm slow.
>
>
> > "Path dependence" says that if something *does* have a measurable net
> > benefit it *won't* succeed in the marketplace.

Path dependency does *not* mean that something that has a measurable net
worth won't succeed. What it means is that the mechanism by which an
economic system evolves is irreversible. That is, a dynamic economy does
not obtain an equilibrium, but functions as a steady state.

It would help if you consider the model in terms of classical
thermodynamics. Remember, the Santa Fe guys were strongly influenced by
collaboration with the physics community. Traditionally, economists have
talked in terms of equilibria because the linear mathematics that implies
are solvable. It was always implicitly understood that the real world was
not a static eqilibrium system, bt an equilibrium model allowed a snapshot
of a problem. A dynamic system, like a real economy, is much tougher. Try a
model of an economy that accounts for all the feedback mechanisms and see
how tough it is to find a finite set of solutions that are not sensitively
dependent on the input conditions.

Consider the problem of reversing the evolution of an economy by
eliminating a major individual technology. The economy cannot revert to the
pre-technology state, because the influences of the technology on other
developments, and vice-versa, has changed the whole nature of the system.
The path by which the technology developed is important. Imagine, for
example, trying to eliminate the organizational technology of the
corporation. (See Fukuyama's "Trust", for example).

>
> > More to the point, Prof. David said that absent QWERTY there is *no*
> > example of meaningful economic path dependence *anywhere* in the real
> > world that he is aware of.
> >
> > It's fair to assume, I think, that Prof. David is aware of Microsoft,
> > Sears tools, chinese junks, the history of money, and so on. So if
> > you people consider these to be examples of "path dependence" then
> > you're working with very different definitions than he his, and it's
> > his theory after all.
>

> It's very likely normal people will interpret phrases within the context
oftheir
> experiences. "Path dependency" does not have the feel of technical
> phrase only to be interpreted within the confines of a specific theory.
>

> I watched yesterday's senate hearings concerning Microsoft and the
internet.
> Some interesting comments were made. It is generally acknowledged
Microsoft

> Others at the hearing said no moneys are being made available to compete
> directly
> against Microsoft because the risks are just too high. Microsoft is
being given
>
> a wide berth because few can afford development costs whose recovery
depends
> upon the sales of products Microsoft may bundle with their operating
system and
> so technological innovation is being stifled.
>

> This sure looks like a market failure. If you don't want to call a
failure due
> to
> "path dependency" then what do you want to call it?

Yet is it a failure? Even equilibrium models have to accept, that with a
significant enough advantage in utility, a product can dominate a market.
This is a market working as it is supposed to. The idea that a monopoly is
some sort of failure is rooted in ant-trust politics, not good theory. The
market, in fact, may be exceptionally efficient in these cases in
transmitting information.

A monopoly doesn't have to be permanent, or even strong enough to dictate a
market. When I was a kid, IBM was in the situation where Microsoft is now.
In a monopoly position, it became so rigid that younger, more nimble
start-ups were able to take the market away-first the mini-computer
companies such as Digital, the Apple, Microsoft, Sun, etc.

In a static economy, a monopoly would be able to dictate prices. Of course,
the economy is *not* a static, equilibrium system. The only way a monopoly
could corner a market, dictating prices over a long term, and be absolutely
secure in a dynamic economy is to control a raw material fundamental to an
economy - such as oil or water, or obtain a legal sanction that forbids
competition, such as the electric and phone companies once enjoyed. It is
no coincidence that the great trusts of the 19th century controlled basic
raw materials such as steel, oil, and sugar.

>
> > But citing one example after another of historical happenstance isn't
> > going to do it:
> > *Historical contingency is NOT economic path dependence*.
>
> I'm eager to learn new phrases. "Historical contingency" and "path
dependency"
> seem synonymous at first glance. I guess their difference usage in
economics
> is an example of historical contingency.

Again, Arthur, in "Path Dependence..." argues that they are the same thing.

[SUSUPPLY]

Mike Turton wrote:

>No problem Pat, as soon as you can tell me how net benefit can be
>calculated without reference to values.

Mike you can't measure your shoe size without values. Are you suggesting no
one has values? Or perhaps it's that everyone is afraid to refer to them?

Sophomoric attempts to hide behind semantics don't make a person look clever.
Only foolish.

How are you doing on your homework assignment?

Patrick

[SUSUPPLY]

Matt Kennel wrote:

>So, do you assume that economy is really just like a system in full
>equilibrium with a linear, homogeneous, 'utility function' so that travelling
>in all degrees of freedom along the local gradient of "utility" actually does
>reach the overall global maximum

>That's the theoretical justification for 'laissez faire', and as far as I can

>tell, it hasn't been experimentally proven. I feel it is based upon an
>entirely supurious analogy to equilibrium classical thermodynamics. In
>physics, the thermodynamical laws are derived from underlying microsocopic

>theory and apply only in some circumstances---they are not taken as axioms!

Let me guess. It's spring break for colleges?

Patrick

[SUSUPPLY]

Robt. Vienneau wrote:

>In my opinion, the Neo-Pareto definition of "efficiency" supposedly
>used by Liebowitz and Margolis is a norm that should be
>of little interest to economists. Those who tend to worry about how
>history matters in economics tend to reject that norm. For example,
>consider my favorite Scandinavian economist - present company
>excepted - Gunnar Myrdal and his use of the concept of "cumulative
>causation."
>
>Another defect of Liebowitz and Margolis' paper is their assumption
>that economic agents can determine the underlying parameters of a
>model generating observed patterns from a single realization.
>Technological change is one place where one would expect non-ergodicity
>to be important. Liebowitz and Margolis sort of acknowledge this, but
>conclude that economists must rely on the special case assumption of
>ergodicity on the grounds that non-ergodicity is supposedly
>non-empirical. It's always suspect when a dubious principle of dated
>philosophy of science is used to decide a substantive point.

Well, this should be a lot of use to those who profess not to even understand
the meaning of the word "benefit". Perhaps you should help Mike with his
homework.

Patrick

[SUSUPPLY]

Markku Stenborg wrote:

>I have no idea what difference this distinction would make on anything
>of any relevance to any econ issue.
>

>Umm, while we're defining terms, What do you mean by "organization"
>here, and why does it matter for something within econ that some stuff
>is called organization and some other stuff is called something else?

Amen, brother.

Patrick

[George Avery]

SUSUPPLY <susu...@aol.com> wrote in article
<19980303203...@ladder02.news.aol.com>...

> Gary Forbis wrote:
>
> >What would count as a disproof of the theory [of "path dependence"]?
>
> Well, quite a lot of what I have written. But you are putting the cart
before
> the horse. The "path dependency" theorists are proclaiming that they are
on
> the cutting edge of a "new economics". It is up to them to provide some
> empirical evidence to support their theory that can stand up to scrutiny.
A
> theory is not assumed to be true, just because someone hasn't disproved
it.

> Especially a radical theory like "path dependence".

The arguments against seem to come largely from economists who are so
hidebound in misunderstanding the language and history of their profession
that they are unable to accept innovation.

Path dependence isn't radical. It does, however, upset conventions used in
neo-classical models (But not the classical model of Adam Smith). Smith was
able to avoid these conventions because he did not try to establish a rigid
*mathematical* model.

Neo-classical economics, as the traditionalists define it, depends on the
concept of an equilibrium. An equilibrium system, however, requires a
static system-a class I or class II Wolfram system for those interested in
the technical details. It was always implicit in the neo-classical model,
however, that this was just a convention to bring the mathematics into a
managable form (complex, non-linear systems tend to be chaotic). This
allowed a formalized approach to the problem, a glimpse at the real
workings of a limited set of the real world. The same simplification can be
seen in game theory, with the focus on two-player, zero sum games. In the
real world, these only exist as...games. But they do help isolate features
of decision making and provide a framework for considering more complex,
real world situations.

The corrollary to this is that the real economy is dynamic. It has feedback
loops, the inputs change, the output is siphoned off into other processes.
Equilibria as such cannot be established. Psuedo-equilibria form, but these
are simple what a physical scientist would describe as "steady state"
functions. A good analogy would be to the biochemical processes of a living
organism, which is a sort of economy in its own right (economics, after
all, is merely the study of decisions and resource allocation). As an old
Biochem prof of mine once said, an organism at equilibrum is ....dead. No
"real" economy is static, or relies on fixed transactions (nobody always
trades "widgets" only for "food", and vice versa).

Any dynamic system has to be path dependent....because the existence of a
complex system of pathways makes it too costly to reverse an action. The
amount of information it takes to describe the beginning and end states,
and the pathways between them, is tremendous, effectively trans-finite in a
real system, and hence the cost of reversing it is effectively
prohibitive. Thuse, the process must be irreversible due to the entropy of
the system. This is a necessary corrollary to Hayek's formal demonstration
of the superiority of a market to an individual in processing information,
and general information theory, both of which are accepted by mainstream
economics.

> >Are you aware of just how much the QWERTY keyboard has changed?....
>
> > Just in the last twenty years the keyboard has
> >changed significantly. What ever happened to the cent sign? Why are
there
> >twleve function keys rather than ten.....
>
> Uh Gary, did you really mean to submit this as evidence of "path
dependency"?
> The theory posits "lock-in", remember.

Lock-in of the path, not the final product. Even those keyboard withot the
cent sign still have the letters "SDFGHJKL" following the "A".
>
> >What would count as a counter example to the proposed theory of path
> >dependency?

>
> Over 200 years of scholarship in economics.

Which largely lacked the mathematics and the conceptual framework to
address complexity. This argument is hardly more than sophistry. For over
1000 years, Aristotilean scholars worked firmly in a paradigm of the
physical universe...that was wrong. As Kuhn noted, most scholarship is
derivative, occuring within the boundaries of a generally accepted
paradigm. When that paradigm is no longer adequate, a scientific revolution
will occur, and those who see beyond the boundaries are generally reected
by an old gaurd because it "doesn't fit the model"-even though the model is
wrong.

[Stewart Goldwater]

rv...@see.sig.com (Robert Vienneau) wrote:

> I tend to get suspicious whenever economists talk about "capital,"

> But I don't think I need to bore our audience with my ideas on
> capital here and now.

Oh do, Robert and tell us why you get suspicious.

[Stewart Goldwater]

mtu...@ms.showtower.com.tw (Michael Turton) wrote:

> This is sci.econ NG, and the thread seems to be one of the few that
> isn't X-posted all over the Usenet. This should give you a hint on
> what types of issues are relevant here. Eg, one would expect that
> words are used here as they are in the Science of Economics, not as in
> SciFi, as for one instance. Also, one would expect that the nuances or
> definitions of words and concepts not really crucial for their Econ
> aspect can safely be ignored.

Yes, good, I'm very glad to have found you Michael! You seem to be
the ideal person to ask for a definition of 'capital'.

[Michael Turton]

In article <19980305234...@ladder03.news.aol.com>,

Pat, I said efficiency is a value, you said, no, to economists, it's a
calculation of net benefits. I guess we agree, then, no?

Mike

[cpw...@spamrahul.net]

susu...@aol.com (SUSUPPLY) wrote:

>Gary Forbis wrote:
>
>Others at the hearing said no moneys are being made available to compete
>directly against Microsoft because the risks are just too high. Microsoft is
>being given a wide berth because few can afford development costs whose
>recovery depends upon the sales of products Microsoft may bundle with their
>operating system
>and so technological innovation is being stifled."
>

>Gary, how do you suppose the following happened: With a true business giant,
>IBM, dominating the computer industry, a couple of college dropouts found the
>money to start Apple Computer and carve out a very profitable niche for their
>company.
>

I'm not Gary but I'd like to answer. IBM never lost it's position
in the market it controlled, mainframes. Apple was duking it out in a new
market with other companies. Then IBM decided to get in. They introduced
a PC that would have been laughed out of the market if it hadn't had IBM's
name on it. I would posit that this was an example of what we laymen have
been calling path dependancy and market failure. Furthermore, for reasons
still unclear (there are many conflicting accounts) IBM chose Microsoft
for the OS. Supposedley they bought the original version of DOS from a
company where a guy had written it in 3 weeks.
IBM, presumably paying more attention to there mainframe biz made
what many now consider to have been strategic blunders, like allowing clones,
but those clones still used DOS! That's how a couple of drop-outs did it,
not by being competitive in the normal sense. But our laymen's notion of
path dependence could explain it.

---- remove UhUh and Spam to get my real email address ---

[Markku Stenborg ®]

On 5 Mar 1998 22:44:23 GMT, rv...@see.sig.com (Robert Vienneau) wrote:

[snip]

> I do think a weakness of mainstream economics is the lack of a
> coherent and well-established theory of the firm.

Care to elaborate? No, I'm not thinking of Arrow-Debreu f(k,l,.) black
box, (which is really not a *theory* of firm in the sense that it does
explain pretty much nothing of firm), but, eg, firm as nexus of
incomplete contracts type of theory.

Anyways, in econ, firms appear in many contexts. The aspects of firms
that are relevant in some problems, can be quite irrelevant in others.

[SUSUPPLY]

Mike Turton wrote:

>Pat, I said efficiency is a value, you said, no, to economists, it's a
>calculation of net benefits. I guess we agree, then, no?

No, and how far are you going to press this childishness? I most assuredly did
not say efficiency is a "calculation" (of anything). It is of course a result
of such calculations, whether formal or informal. Just as you demonstrated
with your remarks about Chinese junks--I guess you missed that post of mine.

I repeat, how are you coming with your homework assignment? You remember, the
one where you are going to explain how corporations exhibit increasing
returns due to network effects and thereby "lock-in" the corporate form of
ownership, to the detriment of a superior form of ownership. Don't forget to
be specific about what that superior form of ownership is.

Best,
Patrick

[SUSUPPLY]

Carl wrote:

> I'm not Gary but I'd like to answer. IBM never lost it's position
>in the market it controlled, mainframes. Apple was duking it out in a new
>market with other companies. Then IBM decided to get in. They introduced
>a PC that would have been laughed out of the market if it hadn't had IBM's
>name on it. I would posit that this was an example of what we laymen have
>been calling path dependancy and market failure. Furthermore, for reasons
>still unclear (there are many conflicting accounts) IBM chose Microsoft
>for the OS. Supposedley they bought the original version of DOS from a
>company where a guy had written it in 3 weeks.
> IBM, presumably paying more attention to there mainframe biz made
>what many now consider to have been strategic blunders, like allowing clones,
>but those clones still used DOS! That's how a couple of drop-outs did it,
>not by being competitive in the normal sense. But our laymen's notion of
>path dependence could explain it.
>

Carl, I realize that nothing that can be written on this newsgroup will
penetrate, but you have just done a moderately successful job of rebutting what
Paul Krugman used to call "The Economics of QWERTY" (i.e. "path dependence").

Was it "not by being competitive in the normal sense", when, say, the VW Beetle
made its successful inroads in the American automobile market. After all there
were many existing companies making far better cars (more powerful, more
comfortable, larger capacity, safer). All the VW had was this itsy bitsy,
teeny weeny gimmick: it was cheaper (both to buy and operate).

Patrick

[Robert Vienneau]

In article <3507ceee...@news.eunet.fi>, real.a...@bottom.of.msg
(Markku Stenborg Ž) wrote:

> On 5 Mar 1998 22:44:23 GMT, rv...@see.sig.com (Robert Vienneau) wrote:
>
> [snip]
>
> > I do think a weakness of mainstream economics is the lack of a
> > coherent and well-established theory of the firm.
>
> Care to elaborate? No, I'm not thinking of Arrow-Debreu f(k,l,.) black
> box, (which is really not a *theory* of firm in the sense that it does
> explain pretty much nothing of firm)

Exactly. Firms don't really exist in General Equilibrium.

> but, eg, firm as nexus of
> incomplete contracts type of theory.

But this theory has only received much attention very recently. That's
why I had the qualification "well-established" in my post.

> Anyways, in econ, firms appear in many contexts. The aspects of firms
> that are relevant in some problems, can be quite irrelevant in others.

Would technological change be one of those areas where one would expect
the theory of the firm to be relevant?

[jim blair]

Gary Forbis wrote:

Gary Forbis <for...@accessone.com>:

>QWERTY is at least a first degree path dependency though it may be a second
>degree path dependency......

Hi,

I think that in trying to use the QWERTY/Shoales vs Dvorak keyboard
debate as a metaphor or example of some general economic, biological
or even Cosmic Principle, the reality and history has been lost or
distorted.

I suggest you read:

http://www.teleprint.com/keyboard/intro.html

and my web page QWERTY file (which I think has links to all of the
important references).

Then consider these points: QWERTY was the "better" keyboard when it
was adopted as the "industry standard" (in Ed Lee's terms): it was
"better" BECAUSE it was slower.

The Dvorak DSK requires less finger movement to touch type a given
document and uses the alternating right and left hands for the common
alternating vowel/consonant pattern. Just look at the layout at:

http://www.ccsi.com/~mbrooks/dvorak/layout.html

At some point in typewriter development, DSK became "better" (faster,
and less stressful) than QWERTY, but by then the standard was
established and hard to change. Individuals convinced of the
superiority of DSK could not change their typewriter.

With computers, you can select the keyboard pattern, and DSK is
making a comeback. But if the physical keys are re-arranged
from QWERTY, they make the board irregular, and it is hard for
many people to adapt to a keyboard if the label on the key
is wrong (hit D and get E, etc.)

My suggestion: a set of little labeled tape key caps to put over
the keys with their DSK meaning. By the time they wear out, you
won't need them. I may try that if I get time.

(Hey, if that works and improves my typing, I can make and sell
sets of them :-)

>The arguments against third degree path dependencies seems to rely upon the
>notion that there is some aggregate optimal or preferred outcome. I don't see
>why this need be true and believe it is not.

Of a set of many possible outcomes, there will be some that are
generally accepted to be "better" than others. But I don't see this
in terms of being "locked out" of the "best" outcome by early arbitrary
choices. At least not in the field of economics (but perhaps in
evolutionary biology). See my web page "local minimum comments".

I think more in terms of the transition time to a lower minimum after
changes in technology make the lower minimum possible, or someone
discovers that it exists.

And note in the Ed Lee article on Standards and Innovation that the first
technology is not "locked in": he claims that many variations are tried
before an industry standard is established. Only then is moving to
a new industry standard made more difficult (but not impossible).

MARKET FAILURE?

This whole process of falling into a "local minimum" has been called
"market failure", but "planning failure" is a better term. It MAY
be the "market" but is often government and planning agencies that
have failed.

I will cite two examples where I think government decisions can/have
directed towards locking in a "bad" result.

LAKE WEEDS

When a lake has a problem with weeds there are 2 common solutions. Spray
chemicals to kill the weeds, or harvest and remove them. City governments
may like the chemicals because they get quick results with minimum cost.
The weeds most sensitive to the chemicals die and decay. The more
resistant weeds survive and the strategy fails in the long run as the
weeds become ever more resistant to the chemicals.

The harvesting combined with control of nutrients is more expensive
at first and shows limited results for a few years. But it deals
with the cause of the problem which is the fertilizer level in the
water. Typically phosphorous is the limiting factor: cutting and
removing the weeds removes phosphorous from the lake and combined
with limiting the input, will eventually limit the weeds.

But since it costs more up front for weed harvesting machines, and
the improvements don't show until a new Mayor and city council are
likely to be in office, harvesting often is passed over for
chemicals.

CARS AND HIGHWAYS

When people had to pay their transportation costs, trolley lines
or subways were in most cities, and trains connected them.

During the 1950's the federal government built the Interstate
Highway System as a military/defense project and set the country
on the path to the car dominated society. I see this a kicking
us OUT of a deep minimum of ever faster and more efficient transit,
and into a dead-end of ever more expensive highways full of stalled
cars, while we import ever more oil. (see my web page economics files
on Gas Tax, Who Pays for Roads, and Trucks)

But these can hardly be called "market failures".

[George Avery]

A number of very good economic theories of the firm exist....starting with
Herbert Simon's (See his book _Administrative Behavior_) and continuing on
with James G. March's refinements to the so-called "Garbage Can" theory
(March and Simon, "Organizations"; March and Olsen, "Ambiguity and Choice
in Organizations"; March and Weissinger-Baylon, "Ambiguity and Command"),
Oliver Williamson's work ("Markets and Hierarchy"), and numerous others.
Check the journals- Administrative Science Quarterly, J. Political Economy,
J of Economic Behavior and Organization, etc.

I didn't have any problems finding theoretical work on the economics of
organization while working on my thesis....the problem was pruning the
rferences to a managable level.

Markku Stenborg ® <real.a...@bottom.of.msg> wrote in article

<3507ceee...@news.eunet.fi>...

> On 5 Mar 1998 22:44:23 GMT, rv...@see.sig.com (Robert Vienneau) wrote:
>
> [snip]
>
> > I do think a weakness of mainstream economics is the lack of a
> > coherent and well-established theory of the firm.
>
> Care to elaborate? No, I'm not thinking of Arrow-Debreu f(k,l,.) black
> box, (which is really not a *theory* of firm in the sense that it does

> explain pretty much nothing of firm), but, eg, firm as nexus of

> incomplete contracts type of theory.
>

> Anyways, in econ, firms appear in many contexts. The aspects of firms
> that are relevant in some problems, can be quite irrelevant in others.
>

[SUSUPPLY]

George Avery has written two posts which are of interest mainly because they
seem to be part of a graceful exit strategy for the "path dependicists" such as
Brian Arthur. George has been corresponding with me by e-mail, so I was
already familiar with slightly different versions of the arguments he just
posted, and he has already read some of what I am now posting.

What I told George via e-mail, was that I thought he (and Brian Arthur) was
attempting to arbitrarily define the issue away. That is, as ever more
embarrassing criticisms are directed at the theory, and its shortcomings are
exposed, it becomes necessary to construct a more restricted version of what
the theory is, and what it predicts. Now, George is writing from Arkansas,
and, as the whole country is discovering, Arkansans can be pretty audacious
with definitions.

George’s posts, and his e-mail to me, remind me of what former Seattle Mariners
broadcaster Ken Levine called Levine’s law. Which is: "A lead-off walk always
comes around to score......Unless it doesn’t". Most of Levine’s jokes are
funnier than that; he is now employed as a scriptwriter in Hollywood, on
"Frasier" (though they aren’t as funny as some of his blunders as a
play-by-play man. See his memoir, titled something like: "Back, Back...It’s
Outahere!..... No, Wait a Minute").

Avery’s Law, I presume, would be something like "Technologies that exhibit
increasing returns will be locked-in........Unless they aren’t".

I had recommended that George read the work of Liebowitz and Margolis, as they
have already dealt quite effectively with his arguments. His recent two
postings show no hint that he has done that, so I will again recommend to him,
and other interested parties, such papers as:

"Chicken Little Comes Home To Roost: A Misplaced and Flawed Economic Theory
Bedevils Microsoft."

"We Don’t Know Why She Swallowed the Fly: Policy and Path Dependence"

"The Fable of the Keys" (or "Typing Errors" in "Reason Magazine")

"Path Dependence, Lock-In, and History"

"Are Network Externalities A New Source of Market Failure?"

All of the above can be found online, as the "grinch" has pointed out.

Now a few specifics about George’s recent comments: George writes that "Path
dependence isn’t radical." He then tries to defend that statement with some
very pedestrian observations that almost no one disputes, such as: "...the

real economy is dynamic. It has feedback loops, the inputs change, the output
is siphoned off into other processes. Equilibria as such cannot be

established." (Really going out on a limb there, George.) There are other
such yawners, I am not going to bother quoting them.

However, George’s heart really isn’t in it. For at the end of one of his posts
he brings up that old chestnut (beloved of every self-described misunderstood
genius), Kuhn’s "The Structure of Scientific Revolutions":

>As Kuhn noted, most scholarship is
>derivative, occuring within the boundaries of a generally accepted
>paradigm. When that paradigm is no longer adequate, a scientific revolution

>will occur, and those who see beyond the boundaries are generally rejected

>by an old gaurd because it "doesn't fit the model"-even though the model is
wrong.

See, George wants to have it both ways. "Path dependence isn’t radical." and,
at the same time, it is "a scientific revolution". If anyone can figure out
how this is possible, they can also explain how Bill Clinton didn’t "have
sexual relations with that woman".

Of course the evidence is that the "path dependicists" originally did not
believe any such things as they (and George Avery) are now professing. Brian
Arthur is quoted in Gary Reback’s "White Paper" (which Federal judge Stanley
Sporkin relied upon) as follows:

"My greatest fear [of allowing the Intuit merger to occur] is that an inferior
technology could be locked in and bring progress to a halt.".

Later in the same paper is this Orwellian claim:

"It is difficult to imagine that in an open society such as this one with
multiple information sources, a single company could seize sufficient control
of information transmission so as to constitute a threat to the underpinnings
of a free society. But such a scenario is a realistic (and perhaps probable)
outcome."

The above are typical sentiments from Brian Arthur and Paul David. The
examples that these two, the most prominent of the "path dependicists", have
given us as proof of the theory are QWERTY and VHS. Now that those examples
have been exposed as fraudulent (see the Wall Street Journal of Feb. 25, 98,
page B1) they are left with NOTHING. And, conspicuous by their absence in
George Avery’s recent posts, are any real-world examples of the supposed
phenomenon.

As Liebowitz and Margolis have said, it is some "new paradigm" that cannot even
marshal one piece of evidence in its support. Which was exactly what the
"grinch" was getting at with his pointed remarks. If you do know of even one
example, George, Brian Arthur and Paul David, I am sure, would be ecstatic to
hear about it. I guess I am "hidebound" to point this out.

As I have pointed out before, this is not an ivory-tower dispute. Both the
U.S. Dept. of Justice, and the Senate Judiciary committee are taking this
drivel seriously. Pace Liebowitz and Margolis, it is some "scientific
revolution" that needs stuffed-shirt politicians and government prosecutors to
impose it on the economy.

Sincerely,
Patrick

[Robert Vienneau]

Mr. Goldwater,

Some of my views on capital I set forth in various essays on my FTP
site. These essays are also available under:

http://csf.Colorado.EDU/pkt/pktauthors/Vienneau.Robert/

I don't think the existence of land alters the problems I find.
Nevertheless, I warn you none of my examples include land.

At one point you asked Michael Turton to define capital. But the
text you were addressing was misattributed. It was Markku Stenborg
you should have been asking.

Mr. Avery,

How would you characterize the economics profession's reception
of Simon's theory?

[jim blair]

Grinch wrote:
>
>Grinch:

>Now, the whole concept of "path dependence" was put forth in a 1985
>paper by economist Paul David, entitled _Clio and the Economics of
>QWERTY_.
> But last week, in Wednesday's Wall Street Journal, Prof. David
>admitted that he now views the QWERTY example to be false because
>QWERTY provides no measurable benefit over Dvorak. Thus, there is no
>"path dependence" to be explained here.

Hi,

Is there a URL or on line copy of that David paper? It looks
something I should have in my QWERTY file.

And I bet you meant to say that the WSJ quotes David as saying
that Dvorak provides no measurable benefit over QWERTY? But I
doubt this. Based on looking at the two patterns, I would
expect Dvorak to be better. But what advantage in speed or
hand stress would that translate into? Would need a study
to say, and that gets back to which study do you believe.

But I can picture this situation: You own/manage a company with
X typewriters and produce a moderate amount of typing. You think
the DSK keyboard is 10% faster, but since no one makes a typewriter
with that keyboard, you use the "off the shelf" QWERTY machines.
Your company does not buy enough typewriters to justify a
special order to Remmington for X of the DSK's, and you can't
teach your typists DSK without them. It would take an order
of, say 100X typewriters to justify the cost of the special
order.

Now there are 200 OTHER companies out there each with the
same situation. Collectively, it would be worth it to change,
but no one can do it alone. And most of them do not even know
about DSK.

I suspect this was the situation sometime AFTER typewriters
improved to the point that mechanical key jams were no longer
a problem for people who could type "too fast". (Before that
time, QWERTY was the "better" keyboard BECAUSE it was slower!)

Now, since computers, each individual can change to DSK.
My guess is that there will be little key labels to mark the
keys with their DSK meaning on sale (for about a dollar), and
that many people will change.

(hey, I may make and sell those key labels myself if no one
else does ;-)

It is not a case of being "locked in" to an older system, but
of a "time of transition" to make the jump to a better system
when it come along.

>More to the point, Prof. David said that absent QWERTY there is *no*
>example of meaningful economic path dependence *anywhere* in the real
>world that he is aware of.

CARS OF THE 50's??

I will suggest one. Back in the 1950's all US cars had the motor
in front but were rear wheel drive. This causes a loss of room
where the rod that transmitted the power from the motor to
the rear went, and a loss of energy (lower gas milage). But
all cars had this design. Why?

The rumor was that since the horse came before the cart, the
motor must belong in front. And it was easier to steer with the
front wheels, so the power went to the rear ones.

This pattern was broken by the German Volkswagen Bug with its
rear mounted (air cooled!) motor. Cheverolet make a similar
design (convair?), and now front wheel drive front end motor
is the "normal" design. That is probably the best design,
but it took a while to become the "industry standard". And
some variation of a turbine or external combustion motor may
someday replace the current internal combution motor. I think
it is in wide use more because it is considered normal than
because it is really the best. Time will tell.