What would happen if...
JShr...@gmail.com
How much would this cost?
Then someone paid to have the Lisp libraries modernized to CPAN(etc)
standards?
How much would this cost?
Then someone paid to do something like Rails in the foregoing new
modernized OS Lisp?
How muh would this cost?
Pascal Bourguignon
> Someone paid to have one of the commercial Lisps open sourced?
> How much would this cost?
Assume the company earns 200k/year for the share holders with its
CL implementation. It would be worth at least 4M.
If the share holders are also programmers working on CL
implementation, you'd have to add their salaries, so it would be worth
at least 9M.
> Then someone paid to have the Lisp libraries modernized to CPAN(etc)
> standards?
> How much would this cost?
Depends on the size of your team. At least 50k/year/person.
Let's say 25 man.year: 1.25M.
> Then someone paid to do something like Rails in the foregoing new
> modernized OS Lisp?
> How muh would this cost?
Depends on the size of your team. At least 50k/year/person. Let's
say 10 man.year, 0.5M.
So total you're speaking of something like 10M at least.
(monetary unit being either USD or EUR or similar).
--
__Pascal Bourguignon__ http://www.informatimago.com/
NOTE: The most fundamental particles in this product are held
together by a "gluing" force about which little is currently known
and whose adhesive power can therefore not be permanently
guaranteed.
Tim X
Well, how long is a piece of string....
I don't quite understand the motivation/thoughts underlying this question. What
would be the advantage of paying (probably a lot) to have a commercial lisp
made into open source compared to putting that money into developing/expanding
an existing open source version?
If the objective is just to try and have a rails like app, why not just do that
with one of the existing open source lisps?
Rails has obtained quite a lot of interest and having looked at it, its quite
nice. There are some aspects of it that I don't think quite work, at least not
for the way I like to design and build web apps, but it does feel better to use
than other web development frameworks I've used. In particular, it does show
how easily you can build quite sophisticated web apps without having the
complexities associated with other models, such as Java beans etc, which seem
to have way to much overhead for all but the most complex of applications.
However, I'm not sure there is anything technical stopping someone from
developing an similar (maybe even better) framework using one of the existing
open source or commercial CL implementations. At least I'm not aware of
anything in a commercial implementation which would facilitate this over an
open source implementation.
I think the limitation or reason we don't have such a framework (noting that
some may argue we actually already do given the libraries and projects working
in this area) is more related to not having someone who has both the creative
ability, CL skill and the idea for some web app which fills a perceived need at
just the right time for the market to accept it (i.e. a BaseCamp equivalent).
It may even be a mistake to try and develop a rails competitor at this time.
Anything that tries to do the same sort of role as rails is going to just be
seen as an attempt to jump on the band wagon or ride on the coat tails of ruby
and rails and being done in CL is unlikely to gain any real support or get past
the FUD out there concerning lisp.
If on the other hand, someone came up with a truely original idea which has a
superior implementation that is largely due to being written in CL and it is
supported by some app that fills a real need that isn't being adequatly
addressed already - well then, just maybe.....
--
tcross (at) rapttech dot com dot au
John Thingstad
>
> If on the other hand, someone came up with a truely original idea which
> has a
> superior implementation that is largely due to being written in CL and
> it is
> supported by some app that fills a real need that isn't being adequatly
> addressed already - well then, just maybe.....
>
http://www.franz.com/products/allegrograph/
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
JShr...@gmail.com
> would be the advantage of paying (probably a lot) to have a commercial lisp
> made into open source compared to putting that money into developing/expanding
> an existing open source version?
I believe that the cost of brining any of the current open lisps up to
the level of any of the commerical ones would be about the same as
just open sourcing a commerical one, and the latter plan would be a
great deal faster and more reliable.
> If the objective is just to try and have a rails like app, why not just do that
> with one of the existing open source lisps?
That is not the goal. The goal is to make Lisp a viable competitor in
the current programming language market, dominated by C++, Java, and
Python/Ruby. The Rails part of my question is merely a part of the
overall picture. In order to play in this domain there needs to be a
single, free, excellent Lisp with a modern library and several killer
apps; Rails seems to be a killer app that would be easy to copy --
that's the only reason I mentioned Rails.
My theory is that is we bought and open-sourced ACL or LispWorks
Lisps, including buying support for their engineers as gatekeepers,
and building out the libraries, we could get the community to rally
around that one, flush the confusing morass of current open source
versions, and focus on building out from that root. I'm fairly certain
that this path is the right one for Lisp; What I'm not sure of is how
to do it in such manner as the commercial vendor doesn't fold in the
process. That vendor would have to transition to being an application
level provider, and which one survives that will depend upon which of
them has the best set of verticals, and which one takes the leap to
open source their lisp first, because then their verticals will
automatically run on the new lisp, and also they get to own the very
important hearts and minds of the market.
Tim Bradshaw
> Depends on the size of your team. At least 50k/year/person.
> Let's say 25 man.year: 1.25M.
That's probably a fairly significant underestimate of the cost of
employing a decent programmer for a year. I don't know what
programmer salaries are like, but I know that the daily rate for a
decent SA come to something significantly more than 50k/year (assuming
dollars: even for GBP a reasonable SA should expect that or more in
salary I think, somewhat location-dependent). That's before all the
overheads etc are factored in which would probably be a factor of
1.5-2.
In the UK, a reasonable contract rate for a decent SA might be GBP 300/
day which is (220 days/year) GBP 66,000, or say $130,000 / year.
That is without overheads, so perhaps the total cost might be $200,000/
year.
Alex Mizrahi
(you :wrote :on '(Sun, 10 Jun 2007 14:13:55 +1000))
(
TX> I don't quite understand the motivation/thoughts underlying this
TX> question. What would be the advantage of paying (probably a lot) to
TX> have a commercial lisp made into open source compared to putting that
TX> money into developing/expanding an existing open source version?
because open sourcing existing has known outcome.
funding development has unknown -- team might not deliver results it will
have too. as you might know, _most_ software projects either run slower than
expected, or completely fail.
as for expanding existing open source implementation, the basis they are
written on might be not appropriate for the desired product. CMUCL and SBCL
have big problems porting them to Windows, and i suspect there are also
glitches with other platforms, because of unusual memory handling. ACL and
Lispworks were ported on Windows long time ago and live fine there.
also, time might be crusial for the idea the Shrager wants to implement --
if project will deliver totally cool implementation in, say, two years, it
will have to compete not with existing languages and tools, but with future
ones -- they might be superior. OTOH open sourcing existing implementation
is much faster, if you have the money :).
by the way, i think open sourcing will not totally kill profits of
commercial vendors -- there still should be people willing to pay for
support.
the prominent north-american Linux vendor, Red Hat, has 4.5 billion dollars
market capitalization -- and all they do is selling support. anyone can
download their source code or fully compatible binaries from CentOS, for
example, and even get updates from CentOS. but some people prefer to pay for
support, and you see -- it's profitable.
certainly, that would be a loss of profit at first, since people who don't
need support won't pay. but in long drive, it can bring even more profit,
because of (much) larger user base.
)
(With-best-regards '(Alex Mizrahi) :aka 'killer_storm)
"I am everything you want and I am everything you need")
Rainer Joswig
"Alex Mizrahi" <udod...@users.sourceforge.net> wrote:
> (message (Hello 'Tim)
> (you :wrote :on '(Sun, 10 Jun 2007 14:13:55 +1000))
> (
>
> TX> I don't quite understand the motivation/thoughts underlying this
> TX> question. What would be the advantage of paying (probably a lot) to
> TX> have a commercial lisp made into open source compared to putting that
> TX> money into developing/expanding an existing open source version?
>
> because open sourcing existing has known outcome.
>
> funding development has unknown -- team might not deliver results it will
> have too. as you might know, _most_ software projects either run slower than
> expected, or completely fail.
>
> as for expanding existing open source implementation, the basis they are
> written on might be not appropriate for the desired product. CMUCL and SBCL
> have big problems porting them to Windows, and i suspect there are also
> glitches with other platforms, because of unusual memory handling. ACL and
> Lispworks were ported on Windows long time ago and live fine there.
Franz bought a Windows-based Lisp implementation and merged it with
its own Unix-based offerings...
>
> also, time might be crusial for the idea the Shrager wants to implement --
> if project will deliver totally cool implementation in, say, two years, it
> will have to compete not with existing languages and tools, but with future
> ones -- they might be superior. OTOH open sourcing existing implementation
> is much faster, if you have the money :).
>
> by the way, i think open sourcing will not totally kill profits of
> commercial vendors -- there still should be people willing to pay for
> support.
> the prominent north-american Linux vendor, Red Hat, has 4.5 billion dollars
> market capitalization -- and all they do is selling support. anyone can
> download their source code or fully compatible binaries from CentOS, for
> example, and even get updates from CentOS. but some people prefer to pay for
> support, and you see -- it's profitable.
>
> certainly, that would be a loss of profit at first, since people who don't
> need support won't pay. but in long drive, it can bring even more profit,
> because of (much) larger user base.
>
> )
> (With-best-regards '(Alex Mizrahi) :aka 'killer_storm)
> "I am everything you want and I am everything you need")
JShr...@gmail.com
engineer is $300k/year. I don't actually think that that is the
significant factor here because that's a fairly small recurring cost.
The big cost here will be to buy the product lock-stock-and-barrel
from the vendor. THAT will be the most expensive aspect because, of
course, they have hundreds of person-years behind their products, but
we don't have to use THAT total cost, but rather their current cash
flow and a multiplier (that is, ~ their p/e ratio). Of course, I don't
know what their cash flow is, but lets just make a guess and say that
one of these gets, say, $5K/license and that they sell 2000 licenses/
year. So that's 2000 * $5000 = $10,000,000, times a multipler of, say,
5 ~= $50M; that's more like the calculation I'm looking for.
Is that a reasonable way to calculate this?
Chris Russell
> That is not the goal. The goal is to make Lisp a viable competitor in
> the current programming language market, dominated by C++, Java, and
> Python/Ruby. The Rails part of my question is merely a part of the
> overall picture. In order to play in this domain there needs to be a
> single, free, excellent Lisp with a modern library and several killer
> apps; Rails seems to be a killer app that would be easy to copy --
> that's the only reason I mentioned Rails.
So what features do you think open source lisp is missing that
prevents it from competing efficiently, but which are present in ACL/
lispworks?
Tim Bradshaw
> Let's just in order to get past this detail, agree that a fully loaded
> engineer is $300k/year. I don't actually think that that is the
> significant factor here because that's a fairly small recurring cost.
That recurring cost *is* what it costs to run the company.
To answer your other question, what you are proposing is essentially a
hostile takeover of a privately-held company. There are standard ways
of doing that kind of valuation.
(I think this idea, which has been proposed before, is just as
implausible as it was last time BTW...)
--tim
JShr...@gmail.com
> > engineer is $300k/year. I don't actually think that that is the
> > significant factor here because that's a fairly small recurring cost.
>
> That recurring cost *is* what it costs to run the company.
No, it's only a part of what it costs to run the company. But
regardless, it is small relative to the investment required to buy out
the product, and moreover may be recoupable in, for example, support
contracts (as was mentioned elsewhere in this thread).
> To answer your other question, what you are proposing is essentially a
> hostile takeover of a privately-held company. There are standard ways
> of doing that kind of valuation.
No, it is not. You cannot do a hostile take over of a privately-help
company. All you can do it to try to offer them what they think their
product is worth (or perhaps what the whole company is worth) to sell
the product (or company) to you. And although I believe that there may
be "standard" ways of reading tea leaves for valuation, you apparently
don't intend to share any of your insights on this with us.
> (I think this idea, which has been proposed before, is just as
> implausible as it was last time BTW...)
It would be nice to have a pointer with that claim. But regardless,
the market is in rapid flux, esp. since the exponential rise of Ruby,
and so I believe that the theory is becoming rapidly more plausible.
It is also likely that the previous proposers were simply trying to
talk the commercial vnedors into giving it away, whereas I'm talking
about buying it from them. Since you haven't provided a pointer for
your claim that this was previously discussed, I could be wrong. But
if I'm not, then this should be a very different sort of conversation.
JShr...@gmail.com
> prevents it from competing efficiently, but which are present in ACL/
> lispworks?
A team of brilliant and dedicated engineers who have maintained and
built out the software full time for many years, and who, given the
opportunity, will continue to do so.
Kent M Pitman
Just for the record, what opportunity are you offering them?
As you note, they've had years of revenue from selling it. How exactly
would you present the case to them that they should trade an ongoing
source of revenue for a fixed cost and assume they are doing something
financially secure? It sounds a little like asking a farmer how much he'd
want to sell his farm. Unless you were going to finance his retirement,
I don't see him selling. But you used the phrase "will continue to do so",
as if the farmer thinks he'll still be a farmer.
Although it's well-known that I have reservations about open/free
software as a general-purpose market paradigm (although it's also
widely misquoted what those reservations are), I'm not trying to argue
any particular case here so please don't misinterpret my remarks. I'm
not challenging you, at least not at this time. I'm just trying to
understand what is being suggested, before I bother to take a position
on it one way or another.
JShr...@gmail.com
> software as a general-purpose market paradigm ...
As do I. In fact, I believe that is is confused and incoherent.
Unfortuntely, it is an unstoppable force in some areas of software
engineering, and so if one wishes to survive, one must figure out how
to use that force to one's advantage.
> > A team of brilliant and dedicated engineers who have maintained and
> > built out the software full time for many years, and who, given the
> > opportunity, will continue to do so.
>
> Just for the record, what opportunity are you offering them?
Just for the record, I haven't offered anyone anything. However, were
I rich enough to make this happen, what I would offer them is to
become a part of a larger, much more dynamic and profitable set of
vertical operations built on top of their (thence open source) Lisp,
with a larger, much more dynamic and profitable set of vertical
visions, where they could share in the profitability of that
operation, and at the same time have many more resources with which to
develop their (thence open source) lisp. Sort of like Google does with
Python.
> As you note, they've had years of revenue from selling it.
Although this seems to be one a downward spiral.
> How exactly
> would you present the case to them that they should trade an ongoing
> source of revenue for a fixed cost and assume they are doing something
> financially secure?
That wasn't quite what I had in mind. As above, I would buy them out,
and then continue to pay them to work on the product as a part of a
larger operation, where the Lisp is open source (but the larger
operation is NOT!)
> It sounds a little like asking a farmer how much he'd
> want to sell his farm. Unless you were going to finance his retirement,
> I don't see him selling.
> But you used the phrase "will continue to do so",
> as if the farmer thinks he'll still be a farmer.
Exactly so! The goal is to buy the farm and continue to pay the farmer
to farm it, and at the same time to share in the potential profits of
a larger, potentially more profitable operation that makes, for
example, egg nog from his eggs (and he can give away the eggs too! --
this isn't quite analogous because unlike eggs, you can both use your
open source software, and give it away too! :-)
Tim Bradshaw
> > That recurring cost *is* what it costs to run the company.
>
> No, it's only a part of what it costs to run the company.
Ah yes, I forgot all the heavy plant and so on that software companies
need. Silly of me.
>
> No, it is not. You cannot do a hostile take over of a privately-help
> company.
That was my point.
> It would be nice to have a pointer with that claim. But regardless,
> the market is in rapid flux, esp. since the exponential rise of Ruby,
> and so I believe that the theory is becoming rapidly more plausible.
Let's look at it from the point of view of someone who as money and
might want to invest in such a scheme (me, say). What, actually,
you're saying is that commercially developed lisps are far more
valuable than open-source ones. So you propose to buy a commercial
Lisp house and *make their product open source*. OK, well, thanks,
I'll go and put my money somewhere else, I think.
Tim Bradshaw
>
> Exactly so! The goal is to buy the farm and continue to pay the farmer
> to farm it, and at the same time to share in the potential profits of
> a larger, potentially more profitable operation that makes, for
> example, egg nog from his eggs (and he can give away the eggs too! --
> this isn't quite analogous because unlike eggs, you can both use your
> open source software, and give it away too! :-)
Has it occurred to you that the Lisp companies might have considered
this idea? What is their cost to open source their product? Why have
they not done so? I mean, it's not as if this idea is some weird
thing any more: look at Sun, say.
Edi Weitz
>> As you note, they've had years of revenue from selling it.
>
> Although this seems to be one a downward spiral.
Really? Do you have any evidence for this?
--
Lisp is not dead, it just smells funny.
Real email: (replace (subseq "spam...@agharta.de" 5) "edi")
Tim X
Hmmm. I don't think its that simple. While I agree that some find the
number/variety of open source CL implementations confusing and I agree that a
good standard library consisting of high level support for common tasks
required for typical applications (networking, web, databases, portable GUIs
etc) would certainly be welcomed, I don't see that making a commercial lisp
open source will address the market share issue or dispel the FUD associated
with lisp. In fact, there is always a risk that we would lose users,
particularly corporate ones. I suspect some companies that are using one of the
commercial lisps do so because they feel there is more security in using a
product that is commercial and which probably has a well defined road
map/direction and guarantee of timely support. While I don't necessarily agree
with such arguments, they are the ones often used by companies when selecting a
product to use.
It isn't the lack of a single open source implementation with high level
features that is preventing CL from obtaining a greater market share. I think
its more to do with the predominate paradigm of procedural languages, the lack
of good intriduction to CL in teaching, the perception that CL is old and
therefore can't be as good as newer languages and possibly even a drop in skill
levels amongst programmers (I'm meeting more and more programmers who have no
real interest in their 'art' - its a job and they do it while they have to in
the same way a clerical assistant does their job. There is no great drive to
improve your skill, breadth of knowledge regarding techniques, algorithms or
data structures etc. This also seems to be encouraged by the management view of
programming being 'grunt work', which you will outsource to somewhere cheap if
you can. There is a strong perception that programming is what you do when you
first start, but if your any good, you will move up into higher positions that
are further and further removed from writing code. It is rare to find a career
path in a company that will allow you to continue programming - in fact, I ve
often seen the attitude that anyone who is a programmer in their late 30s
onwards is someone with no ambition or career aspirations. Obviusly, there are
exceptions, particularly in some specialist domains, but on the whole...)
Putting all that aside, I also don't see any real business case for the
proposal. Where is all this money going to come from to convince any company to
open source their product and replace that lost revenue. While this may work in
for some companies who can offset this through the provision of other services,
I can't see this working in the area of CL, mainly because there isn't a large
enough user base. Keeping the existing engineers on-board in order to continue
development would be expensive - wehre would that revenue come from? Other
companies that have gone down this road have been able to get new revenue from
providing things like consulting services, high level support etc, but you need
a sizable user base to do this and I don't think there is one at present.
Personally, I don't think we will see CL up there with C/C++, Java, Python,
Ruby in the foreseable future, if ever. I think there are enough people using
it and liking it to see it continue for a long time, but suspect it will remain
the domain of specialist applications or small developers/consultancies. I
think its very unlikely we will see it as a mainstream programming language.
This may change if there are more CL based success stories, particularly ones
that succeed because they are done in CL and are difficult to reproduce using
other languages or ones that somehow change the application langscape in such a
way that CLs advantages become even more self evident and which create
expectations that are difficult to meet using other languages.
Tim
Tamas Papp
>> I don't quite understand the motivation/thoughts underlying this question. What
>> would be the advantage of paying (probably a lot) to have a commercial lisp
>> made into open source compared to putting that money into developing/expanding
>> an existing open source version?
>
> I believe that the cost of brining any of the current open lisps up to
> the level of any of the commerical ones would be about the same as
> just open sourcing a commerical one, and the latter plan would be a
> great deal faster and more reliable.
>
>> If the objective is just to try and have a rails like app, why not just do that
>> with one of the existing open source lisps?
>
> That is not the goal. The goal is to make Lisp a viable competitor in
> the current programming language market, dominated by C++, Java, and
> Python/Ruby. The Rails part of my question is merely a part of the
> overall picture. In order to play in this domain there needs to be a
> single, free, excellent Lisp with a modern library and several killer
> apps; Rails seems to be a killer app that would be easy to copy --
> that's the only reason I mentioned Rails.
You are under the impression that the cost of commercial lisps is the
sole barrier that prevents Lisp from being a "viable competitor". Why
do you think this? When you have some application that is profitable,
a few thousand dollars is peanuts. Or even in a research environment,
software licenses are a relatively small part of your expenditures,
compared to payroll and other costs.
There are nice FOSS Lisps out there, so people can try Lisp before
paying money for a commercial license.
> My theory is that is we bought and open-sourced ACL or LispWorks
> Lisps, including buying support for their engineers as gatekeepers,
> and building out the libraries, we could get the community to rally
> around that one, flush the confusing morass of current open source
> versions, and focus on building out from that root. I'm fairly certain
> that this path is the right one for Lisp; What I'm not sure of is how
> to do it in such manner as the commercial vendor doesn't fold in the
Why would anyone contribute money into this scheme, even if your
theory is correct? (which I doubt). Let people who need a commercial
Lisp pay for it. This way, everyone considers the costs and benefits
of buying a commercial Lisp, and will only do so when they see a
benefit that outweighs the costs. What you are proposing doesn't make
economic sense.
Tamas
MarkHa...@gmail.com
Oddly enough, I was actually thinking about this after reading an
interview with Shuttleworth. Can Common Lisp get a sugar daddy? But
I'm pretty sure it would be a whole lot cheaper for said sugar daddy
to plonk down $10 million (over say 3,4 years) to bring SBCL up to
speed (including a top-notch IDE), then to buy out Allegro or
LispWorks. Nathan Myrvold has Intentional Software. Can anybody
convince Paul Allen to spare $10 million in pocket change?
Tim X
>
> by the way, i think open sourcing will not totally kill profits of
> commercial vendors -- there still should be people willing to pay for
> support.
> the prominent north-american Linux vendor, Red Hat, has 4.5 billion dollars
> market capitalization -- and all they do is selling support. anyone can
> download their source code or fully compatible binaries from CentOS, for
> example, and even get updates from CentOS. but some people prefer to pay for
> support, and you see -- it's profitable.
I thought the jury was still out on Red Hat's profitability? I've not checked
for some time, but last time I looked (maybe two years ago), they still hadn't
made a profit.
>
> certainly, that would be a loss of profit at first, since people who don't
> need support won't pay. but in long drive, it can bring even more profit,
> because of (much) larger user base.
>
Actually, you hit exactly my point in another post of why I don't htink this
will work for CL - you need a large user base and I don't believe CL has one
that is large enough given the size of the initial investment being discussed.
Note also that a big part of Red Hat's success has been their acceptance by
other vendors. For example, products like Oracle are only certified and
supported by Oracle as long as you run them on a specific Linux distro, Red Hat
being one of them (Oracle is possibly an interesting example as they are now
talking about doing their own Linux distro and it will be interesting to see
what they will support). Where I work, we run a lot of Red Hat, not because we
think its the best Linux distro out there, but because its the only one that is
supported by all the different vendor software we run. While they all support
different combinations of Linux, Red Hat Enterprise is the one consistent Linux
distro all the vendors support. Note also that nearly all of these vendors
require the Enterprise version - Fedora Core is not good enough. If your not
running on the enterprise version, your not supported.
I think this is a big pat of Red hat's success so far - they were able to get
support from other vendors. Interestingly, running Red Hat is not that cheaper
than running our old enterprise OS, Tru64. The licensing and support costs for
Red Hat enterprise are a bit cheaper than the discounted price we use to get
from HP (nee Compaq nee Digital), but we have a lot more servers, so that lower
cost is offset by requiring more licenses and Red hat's discount for
educational institutions is not as cheap as the one we got for Tru64 (at least
that was the case when I was managing the data centre a couple of years ago,
things may have changed).
I can't see this working with CL. As stated, I don't believe there is a big
enough user base to offset the costs over a long enough period to possibly
build a larger user base and I can't see the 'lock in' possibilities that Red
Hat was able to get through working with other software vendors. The real
problem is that large initial investment cost. If you were able to get the
software at a very low price, then just maybe it would be possible (but highly
risky IMHO).
Tim
Chris Russell
Well that would be nice, but do ruby or python have that? How far
behind is a cross platform open source lisp like clisp from matching
the core functionality of the ruby/python interpreters, anyway?
Charlton Wilbur
>> So what features do you think open source lisp is missing that
>> prevents it from competing efficiently, but which are present
>> in ACL/ lispworks?
JS> A team of brilliant and dedicated engineers who have
JS> maintained and built out the software full time for many
JS> years, and who, given the opportunity, will continue to do so.
I'd concur with this, but I'd wonder where you're going to get the
money to pay the team of brilliant and dedicated engineers what
they're worth.
If you take a look at the languages you're comparing Lisp to here --
Python and Ruby -- you'll see that both of those languages began as
all-volunteer efforts. If you take a look at Perl, another language
taking a similar development path, you'll see that most Perl
development work is done as the result of very small ($500-$1000)
grants or as the side effect of the developer's full-time job. The
developers of PHP aren't paid to work on PHP, but to sell proprietary
addons to PHP that make it faster.
So if you want to successfully buy and open-source a commercial Lisp,
and actively retain the engineers, you need to figure out how to pay
them the same salary to do the same thing without any revenues from
sales of licenses. Otherwise, you lose the principal advantage of
commercial Lisps, which is that engineering and support team. And
you'll lose the experienced engineers to the other Lisp vendor, who is
now benefitting from an influx Lisp users who are not too keen on open
source and who want support contracts with actual engineers behind
them and can now afford to hire more engineers and pay them in
currency rather than good feelings.
Charlton
--
Charlton Wilbur
cwi...@chromatico.net
Rainer Joswig
Charlton Wilbur <cwi...@chromatico.net> wrote:
> >>>>> "JS" == JShrager@gmail com <JShr...@gmail.com> writes:
>
> >> So what features do you think open source lisp is missing that
> >> prevents it from competing efficiently, but which are present
> >> in ACL/ lispworks?
>
> JS> A team of brilliant and dedicated engineers who have
> JS> maintained and built out the software full time for many
> JS> years, and who, given the opportunity, will continue to do so.
>
> I'd concur with this, but I'd wonder where you're going to get the
> money to pay the team of brilliant and dedicated engineers what
> they're worth.
>
> If you take a look at the languages you're comparing Lisp to here --
> Python and Ruby -- you'll see that both of those languages began as
> all-volunteer efforts. If you take a look at Perl, another language
> taking a similar development path, you'll see that most Perl
> development work is done as the result of very small ($500-$1000)
> grants or as the side effect of the developer's full-time job. The
> developers of PHP aren't paid to work on PHP, but to sell proprietary
> addons to PHP that make it faster.
How about MySQL?
They GPL one version and sell also a version with a
different license.
>
> So if you want to successfully buy and open-source a commercial Lisp,
> and actively retain the engineers, you need to figure out how to pay
> them the same salary to do the same thing without any revenues from
> sales of licenses. Otherwise, you lose the principal advantage of
> commercial Lisps, which is that engineering and support team. And
> you'll lose the experienced engineers to the other Lisp vendor, who is
> now benefitting from an influx Lisp users who are not too keen on open
> source and who want support contracts with actual engineers behind
> them and can now afford to hire more engineers and pay them in
> currency rather than good feelings.
>
> Charlton
Tyler Smith
As someone who is just learning lisp, I don't know what this means. I
can understand the advantage of having a full-time, professional
development team. But what has this team produced that is absent from
the free implementations? Do the proprietary implementations produce
faster or more stable programs, or do they provide more/better
libraries?
Tyler
ddd
beeing sold by its vendor. It was a cross platform CL to C compiler
and you could even license the source.
Maybe this ex-comercial? CL can be bought quite cheaply. However I
don't know much about the implementation, especially compared to ECL.
JShr...@gmail.com
> this idea?
Yes, this has occurred to me.
> What is their cost to open source their product?
I don't know, they haven't replied to my post yet.
> Why have they not done so?
I don't know, but I doubt that it's because they are making so much on
Lisp licenses that they are all out on yatchs.
> I mean, it's not as if this idea is some weird
> thing any more: look at Sun, say.
Yes, look at Sun, say. It sounds like you're supporting my position.
Sun makes money (if they do) by selling other products -- specifically
computers and consulting services -- not on Java.
JShr...@gmail.com
> might want to invest in such a scheme (me, say). What, actually,
> you're saying is that commercially developed lisps are far more
> valuable than open-source ones. So you propose to buy a commercial
> Lisp house and *make their product open source*.
No. What I propose is to build valuable verticals on an excellent open
source platform that has guaranteed support (because we are supporting
it ourselves). It may be that at some point in the future, the
platform support subdivision can be once more spun out as a self-
sustaining company, but that would be far into the future. To be more
precise, I claim that Lisp is uniquely positioned for the semantic web
market, and that a semantic web company built on an excellent Lisp
platform can be highly profitable in semantic web verticals, and that
the possible much smaller profit in Lisp engines is negligible and so
the engine (and some support) can be essentially given away. Moreover,
by giving it away, you garner additional support for the engine
through the open source community. In the long run, if your model is
not to make money by selling, but rather by using, Lisp engines, then
I claim that the optimal strategy is to give away the engine in order
to garner an open source community for it, and make your money where
you can.
JShr...@gmail.com
> >> As you note, they've had years of revenue from selling it.
>
> > Although this seems to be one a downward spiral.
>
> Really? Do you have any evidence for this?
No. I have the fact that Lisps are far more expensive than pretty much
any other programming languages, and are getting more, not less
expensive. This suggests, although I suppose does not prove my theory.
JShr...@gmail.com
> > built out the software full time for many years, and who, given the
> > opportunity, will continue to do so.
>
> As someone who is just learning lisp, I don't know what this means. I
> can understand the advantage of having a full-time, professional
> development team. But what has this team produced that is absent from
> the free implementations? Do the proprietary implementations produce
> faster or more stable programs, or do they provide more/better
> libraries?
Uh, let's see: Yes, and Yes, and excellent development environments,
and run on every platform, and ... did I mention a team of brilliant
and dedicated engineers who have maintained and built out the software
full time for many years, and who, given the opportunity, will
continue to do so? Having used both free and commercial Lisps myself
rather extensively, I can tell you that it's like being in another
world entirely!
Daniel Barlow
> Just for the record, I haven't offered anyone anything. However, were
> I rich enough to make this happen, what I would offer them is to
> become a part of a larger, much more dynamic and profitable set of
> vertical operations built on top of their (thence open source) Lisp,
> with a larger, much more dynamic and profitable set of vertical
> visions, where they could share in the profitability of that
> operation, and at the same time have many more resources with which to
> develop their (thence open source) lisp. Sort of like Google does with
> Python.
In other words, they would go from being a profit centre to an overhead,
and likely first for the chop should the "set of vertical operations"
turn out to be less profitable than expected.
Not a move I'd /necessarily/ welcome with open arms, were I in their
position.
-dan
Rainer Joswig
"JShr...@gmail.com" <JShr...@gmail.com> wrote:
Actually this is totally wrong.
Nobody escapes the costs. Especially not software companies.
The company has some income:
* licenses
* trainings
* consulting
* applications
* libraries
It has expenses:
* marketing & sales
* staff
* building & equipment, infrastructure
* tax, insurance, ...
* external costs (writers, web designers, ...)
This is absolutely independent if you are selling Lisp or any
other programming languages. The income could be divided
differently. The licenses could be more or less in number and
costs.
If you have a software company with, say, 20 people
it is is independent if you do Lisp or something else.
If the market has a structure and size to a Lisp business
you'll get automatically similar costs for the license.
Don't believe?
Just check the prices for comparable systems (commercial
language implementations with companies behind them).
* Ada compilers
* Java application servers (BEA Weblogic, IBM Websphere, ...)
* Prolog
* Smalltalk (Visualworks, ...)
* Java (Borland, ...)
* Rule systems (Jrules, ...)
You will find that these are as expensive as commercial Lisp and more.
Kent M Pitman
> > I mean, it's not as if this idea is some weird
> > thing any more: look at Sun, say.
>
> Yes, look at Sun, say. It sounds like you're supporting my position.
> Sun makes money (if they do) by selling other products -- specifically
> computers and consulting services -- not on Java.
Only companies make money on other things. Individuals can't
multitask and break even. So the scheme you describe trades
individual autonomy away from one person for market share in a
different market for others. It's possible to argue there's more
money in it your way, though I think it's not as much of a slam dunk
to do so as some allege, but that's not my point. The money made
other ways goes to other people.
Although they both involve software, some of us don't consider
"design" the same job as "consulting". One is about personal freedom
and expression, the other about doing someone else's bidding and
making enough money to pay the rent. Once in a while, one consults
for someone who wants something freely expressed. But I wouldn't say
that's the norm.
When talking about Lisp companies, I'll wager you're talking the dream
job of the people working there. Doing consulting from those
companies probably pays the bills, but is probably not what they are
there for. And so offering to let them give up the design part, the
one little piece of autonomy they have, in exchange for more ability
to "just pay the bills" sounds like a bizarre thing to think they'd
leap at.
MarkHa...@gmail.com
> In article <87ejkiu38v....@mithril.chromatico.net>,
> Charlton Wilbur <cwil...@chromatico.net> wrote:
Another model is the one employed by Cincom Smalltalk. It's not open
source. You can download the entire thing and you only have to pay
when you start using it commercially. I'm not advocating this for
Franz or LispWorks. I'm sure they've weighed the pros and cons of
various models.
Tim Bradshaw
> > What is their cost to open source their product?
>
> I don't know, they haven't replied to my post yet.
That was a rhetorical question. If they own all the rights to their
product (which is likely to be true to first order) then their cost is
close to zero. In particular it is enormously less than the cost of
buying the company in order to give away its product.
> > Why have they not done so?
>
> I don't know,
Well, that's a question that you need to answer.
>
> > I mean, it's not as if this idea is some weird
> > thing any more: look at Sun, say.
>
> Yes, look at Sun, say. It sounds like you're supporting my position.
Sigh. No, I'm trying to point out that the Lisp companies are not
living inside a chamber sealed off from all outside contact in 1990
and are therefore probably aware that open sourcing their product is
an option. (I wasn't thinking of Java but Solaris, but Java will do.)
So it behooves you to try and explain why they have not taken this
option ("they are all stupid" is not a very good answer).
I'm kind of worried that you seem to have magnificently missed my
points in both my previous articles: perhaps it's because I'm
English. In any case, I'll leave it for others to continue: as I said
this has been gone over many times before in cll and I doubt there is
anything new to say.
--tim
Ron Garret
"JShr...@gmail.com" <JShr...@gmail.com> wrote:
> Someone paid to have one of the commercial Lisps open sourced?
This experiment has already been done. OpenMCL is the result.
rg
Ron Garret
"JShr...@gmail.com" <JShr...@gmail.com> wrote:
> > On Sun, 10 Jun 2007 23:29:12 -0700, "JShra...@gmail.com"
> > <JShra...@gmail.com> wrote:
> > >> As you note, they've had years of revenue from selling it.
> >
> > > Although this seems to be one a downward spiral.
> >
> > Really? Do you have any evidence for this?
>
> No.
But that doesn't necessarily mean it isn't true. Generally companies
whose revenues are in a downward spiral do everything they can to
prevent that fact from being known because it's much harder to raise
capital when people know that you're desperate.
rg
JShr...@gmail.com
Interesting. Not being a PPC user I hadn't considered this, but
apparently they've ported it to x86 Linux now as well (although it's
only in pre-released), so it seems that its market share may well
expand. I'll definitely look more closely into this. Thanks! (I wasn't
aware that the OpenMCL folks did a set of verticals; how do the full
time developers support themselves?)
JShr...@gmail.com
Not to worry.
Tim Bradshaw
Wow. At everything above the level of syntax you don't understand
what I say. How odd.
Ron Garret
"JShr...@gmail.com" <JShr...@gmail.com> wrote:
> > This experiment has already been done. OpenMCL is the result.
>
> Interesting. Not being a PPC user I hadn't considered this, but
> apparently they've ported it to x86 Linux now as well
And OS X as well. I run it on my Macbook Pro. 32-bit processors are so
last-millennium.
> (although it's only in pre-released)
You wanted open-source, this is what you get.
> how do the full time developers support themselves?)
They have day jobs, just all other open-source developers.
rg
JShr...@gmail.com
Wow.
Ken Tilton
JShr...@gmail.com wrote:
> That is not the goal. The goal is to make Lisp a viable competitor in
> the current programming language market, dominated by C++, Java, and
> Python/Ruby.
Wise man, choosing a goal already reached.
> In order to play in this domain there needs to be a
> single, free, excellent Lisp with a modern library and several killer
> apps; Rails seems to be a killer app that would be easy to copy --
> that's the only reason I mentioned Rails.
<sigh> You were doing so well. What Lisp has to do is not stop being
Lisp. Man, that sounds easy.
>
> My theory is that is we bought and open-sourced ACL or LispWorks
> Lisps, including buying support for their engineers as gatekeepers,
> and building out the libraries, we could get the community to rally
> around that one,....
PWUAUAHAHAHAHAHAHAHAHAHAHAHAHAAHOAOAHAHAOAHHHOOOOOHWHHEHEEEHHEH...
The idea of this Lisp community "rallying" is about as conceivable as a
hootenany down at the cemetery.
hth,kxo
John Thingstad
wrote:
>
> PWUAUAHAHAHAHAHAHAHAHAHAHAHAHAAHOAOAHAHAOAHHHOOOOOHWHHEHEEEHHEH...
>
> The idea of this Lisp community "rallying" is about as conceivable as a
> hootenany down at the cemetery.
>
> hth,kxo
You have a vierd and twisted mind.. :)
--
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
Josip Gracin
> On Jun 11, 7:29 am, "JShra...@gmail.com" <JShra...@gmail.com> wrote:
>
>> Exactly so! The goal is to buy the farm and continue to pay the farmer
>> to farm it, and at the same time to share in the potential profits of
>> a larger, potentially more profitable operation that makes, for
>> example, egg nog from his eggs (and he can give away the eggs too! --
>> this isn't quite analogous because unlike eggs, you can both use your
>> open source software, and give it away too! :-)
>
> Has it occurred to you that the Lisp companies might have considered
> they not done so? I mean, it's not as if this idea is some weird
> thing any more: look at Sun, say.
My $0.02 on the subject. It seems obvious to me that we are witnessing
the rise in popularity of dynamic languages (due to various reasons).
Stimulated by the hype, more and more people/businesses are looking for
a substitute for C++/Java/C#/etc. Once they find their new pet
language/platform, they will most likely stick with it for the next 5 or
more years. In my opinion, these businesses are mostly small businesses
and I believe current commercial CL prices represent a crucial factor in
discarding CL altogether and going with Python/Ruby/whatever.
My point is: maybe commercial CL vendors should reconsider their
strategy and try to use the hype to gain volume. Once the hype is over
it might be very hard to make living by selling CL compilers. Of
course, vendors would most likely need to broaden their portfolio and
sell other stuff besides development environments.
I think Sun got this right. OTOH, they have enough money to afford
longer term investments.
Alex Mizrahi
(you :wrote :on '(Mon, 11 Jun 2007 14:05:49 -0700))
(
??>> This experiment has already been done. OpenMCL is the result.
J> Interesting. Not being a PPC user I hadn't considered this, but
J> apparently they've ported it to x86 Linux now as well
only 64-bit, which is not yet mainstream.
few days ago i've installed some AMD64 Ubuntu systems -- some apps don't
work there, mostly proprietary with 32-bit binaries.
)
(With-best-regards '(Alex Mizrahi) :aka 'killer_storm)
"I am everything you want and I am everything you need")
Pascal Costanza
> Tim Bradshaw wrote:
>> On Jun 11, 7:29 am, "JShra...@gmail.com" <JShra...@gmail.com> wrote:
>>
>>> Exactly so! The goal is to buy the farm and continue to pay the farmer
>>> to farm it, and at the same time to share in the potential profits of
>>> a larger, potentially more profitable operation that makes, for
>>> example, egg nog from his eggs (and he can give away the eggs too! --
>>> this isn't quite analogous because unlike eggs, you can both use your
>>> open source software, and give it away too! :-)
>>
>> Has it occurred to you that the Lisp companies might have considered
>> this idea? What is their cost to open source their product? Why have
>> they not done so? I mean, it's not as if this idea is some weird
>> thing any more: look at Sun, say.
>
> My $0.02 on the subject. It seems obvious to me that we are witnessing
> the rise in popularity of dynamic languages (due to various reasons).
> Stimulated by the hype, more and more people/businesses are looking for
> a substitute for C++/Java/C#/etc. Once they find their new pet
> language/platform, they will most likely stick with it for the next 5 or
> more years. In my opinion, these businesses are mostly small businesses
> and I believe current commercial CL prices represent a crucial factor in
> discarding CL altogether and going with Python/Ruby/whatever.
Whether a language is dynamic or not is just one of many possible
features in the laundry list for a programming language. Whether a
language makes you more productive or not depends more on how well the
various features are integrated, not whether and which features are
available in the first place.
Whether a language is popular and whether it is available for free or
not are two of the most superficial characteristics of a programming
language. If a company bases its decisions on such superficialities,
it's probably not in a very good shape. Companies have to spend so much
money an assets anyway, so the comparatively small costs for software
cannot possibly be a serious stumbling block.
> I think Sun got this right. OTOH, they have enough money to afford
> longer term investments.
No, Sun just got lucky at one particular point in time. Java became such
a success because of its bundling with Netscape Navigator and the
implied promise of a new deployment model for applications (which never
took off).
Pascal
--
My website: http://p-cos.net
Common Lisp Document Repository: http://cdr.eurolisp.org
Closer to MOP & ContextL: http://common-lisp.net/project/closer/
Josip Gracin
> No, Sun just got lucky at one particular point in time. Java became such
> a success because of its bundling with Netscape Navigator and the
> implied promise of a new deployment model for applications (which never
> took off).
I was actually talking about their current strategy to open source
everything... the way they are attracting people to Solaris in order to
JShr...@gmail.com
> not are two of the most superficial characteristics of a programming
> language. If a company bases its decisions on such superficialities,
> it's probably not in a very good shape. Companies have to spend so much
> money an assets anyway, so the comparatively small costs for software
> cannot possibly be a serious stumbling block.
Unfortunately, this is not the reality that we live in. The open
source movement, and various other factors, has/have convinced nearly
everyone that software should be free -- and unfortunately, they mean
free like beer, not free like freedom, which is exactly the opposite
of its initial intent...but let's let that pass; I really don't want
to start the same fight about open source again!
I have worked at many places, including the most famous and wealthy
university on earth, the most famous (and previously wealthy) computer
science research lab on earth, and several very wealthy places in
between, and I can tell you for absolute fact that every time I have
asked for the money to buy CL licences, or put these into my grant
proposals, the response has been a line item veto, or at least a
serious battle. And these are people who will drop $50,000 on a Mass
Spec that they use rarely, or drop the entire license fee to send a
team on cross-country travel for a week to DC to chat up a potential
client with near zero expected ROI. If you really believe that you can
survive by charging anything with 4 digits before the decimal point
for what looks to them like Ruby or Python, or even Java, only worse
in most *apparent* ways, then you're living in another reality. I hate
the fact that this is how things are, but it's how things are.
JShr...@gmail.com
>
> > Interesting. Not being a PPC user I hadn't considered this, but
> > apparently they've ported it to x86 Linux now as well
>
> And OS X as well. I run it on my Macbook Pro. 32-bit processors are so
> last-millennium.
>
> > (although it's only in pre-released)
>
> You wanted open-source, this is what you get.
Hmmm. Well, I didn't actually want open source in the sense of not
professionally supported. What I wanted (although you'd have to study
the thread to get this) is a company that supports an excellent CL on-
the-side and makes their money through verticals, like Google supports
Guido to do python, and Sun/Java. So, MCL probably doesn't actually
fit the bill afterall.
> > how do the full time developers support themselves?)
>
> They have day jobs, just all other open-source developers.
Right, as above, this isn't what I'm after, so although I will go out
and try it, on your advice (although it looks like I'm going to have
to buy myself a 64 bit machine first!) I'm not sanguine that it's what
I'm looking for.
Pascal Costanza
>>>> This experiment has already been done. OpenMCL is the result.
>>> Interesting. Not being a PPC user I hadn't considered this, but
>>> apparently they've ported it to x86 Linux now as well
>> And OS X as well. I run it on my Macbook Pro. 32-bit processors are so
>> last-millennium.
>>
>>> (although it's only in pre-released)
>> You wanted open-source, this is what you get.
>
> Hmmm. Well, I didn't actually want open source in the sense of not
> professionally supported. What I wanted (although you'd have to study
> the thread to get this) is a company that supports an excellent CL on-
> the-side and makes their money through verticals, like Google supports
> Guido to do python, and Sun/Java. So, MCL probably doesn't actually
> fit the bill afterall.
They exactly fit the bill, AFAICT.
Tamas Papp
> Unfortunately, this is not the reality that we live in. The open
> source movement, and various other factors, has/have convinced nearly
> everyone that software should be free -- and unfortunately, they mean
Really? Nearly everyone? That's quite a strong statement.
> I have worked at many places, including the most famous and wealthy
> university on earth, the most famous (and previously wealthy) computer
> science research lab on earth, and several very wealthy places in
> between, and I can tell you for absolute fact that every time I have
> asked for the money to buy CL licences, or put these into my grant
> proposals, the response has been a line item veto, or at least a
> serious battle. And these are people who will drop $50,000 on a Mass
So this "most famous and wealthy university on earth" didn't have any
commercial software? How did they manage without Mathematica, Gauss,
Matlab, Stata and other similar programs? You must share the secret,
I am sure that less famous and wealthy universities would like to save
money on licenses too.
> Spec that they use rarely, or drop the entire license fee to send a
> team on cross-country travel for a week to DC to chat up a potential
> client with near zero expected ROI. If you really believe that you can
So are you trying to imply that these people are stupid? (that's what
I gather from your description -- spending time and money on a near
zero ROI project qualifies as stupid). Hope you have better luck next
time and find a company with sane people.
Best,
Tamas
evins...@gmail.com
> > > > This experiment has already been done. OpenMCL is the result.
>
> > > Interesting. Not being a PPC user I hadn't considered this, but
> > > apparently they've ported it to x86 Linux now as well
>
> > And OS X as well. I run it on my Macbook Pro. 32-bit processors are so
> > last-millennium.
>
> > > (although it's only in pre-released)
No, there is a 1.0 release available on several platforms. It is true,
however, that most people seem to use the pre-release versions,
presumably since those have more interesting feature sets, and usually
aren't terribly buggy.
> > You wanted open-source, this is what you get.
>
> Hmmm. Well, I didn't actually want open source in the sense of not
> professionally supported.
OpenMCL is professionally supported (assuming you have the money to
pay Clozure to support it).
> What I wanted (although you'd have to study
> the thread to get this) is a company that supports an excellent CL on-
> the-side and makes their money through verticals, like Google supports
> Guido to do python, and Sun/Java. So, MCL probably doesn't actually
> fit the bill afterall.
Well, Clozure develops and supports OpenMCL, and makes its money doing
consulting gigs that sometimes use OpenMCL. Not exactly "verticals",
but other than that, the model seems fairly close to what you
described.
> > > how do the full time developers support themselves?)
>
> > They have day jobs, just all other open-source developers.
>
> Right, as above, this isn't what I'm after, so although I will go out
> and try it, on your advice (although it looks like I'm going to have
> to buy myself a 64 bit machine first!) I'm not sanguine that it's what
> I'm looking for.
Personally, I think it would be unwise for a company like, say,
Clozure, to toss all its eggs in one "vertical" basket, when it can
instead maintain a relatively diverse portfolio of clients.
Conversely, it seems to me that it would be unwise for a company
building a business around some set of "vertical" applications to
start out with the assumption that they are going to commit to some
particular technology platform to do it (winding up with some
particular platform because that's what turned out to work well for
their business is an entirely different matter).
Dimiter "malkia" Stanev
http://www.franz.com/downloads/student.lhtml
From the page:
Requirements and Restrictions
1 Available to undergraduate students enrolled full time at a degree
granting university.
2 Individuals interested in purchasing will be required to validate
their student and enrollment status prior to purchase.
3 Full version of Allegro CL 8.0 with no feature restrictions. Upon
purchase of the Student Version the product will function for a period
of 2 years.
4 Not for use in a commercial setting or in university sanctioned
research projects and applications.
I guess what you've described, is case 4, so it might've not worked for
you back then...
Lispworks also has a Academic Version, which restricts you from
delivering applications/DLL's and distributing those applications free
of charge.
http://www.lispworks.com/products/features.html
Prices for the Academic are higher than Allegro's Student, but still
under 4 figures:
http://www.lispworks.com/buy/prices-1a.html
jimb...@gmail.com
the issues with open source CL implementations is mostly packaging.
There are a lot of open source libraries out there that do most of the
things as the libraries advertised by the commercial lisps. For me,
Emacs (Aquamacs, actually) + SLIME is an outstanding development
environment. Once you understand it, ASDF and ASDF Install is a
pretty good package manager/installer.
However, you need to get your open source lisp, Emacs, and SLIME
separately and figure out how to make them work together. You need to
find the open source libraries you need and investigate which ones are
any good, then confirm they work well with your Lisp. You need to
learn Emacs is you don't already know it. And there's no one to call
if things don't go the way you suspect.
So I think integration and support are the two real services
commercial lisps offer over free ones. Granted, this is based on
playing around with just the free LispWorks and Allegro versions.
There may be more benefits to the paid versions I have not
experienced. There are also things like AllegroCache that look
totally awesome, but I kinda think of those as a separate product
bundled with the high end versions. If the world went to SBCL or some
other open source Lisp, I think Franz could get by just selling
AllegroCache and associated products.
I tried the free versions, but quickly blew through the heap
limitations because I was using Lisp for homeworks that required
processing large datasets. So it was on to SBCL and no looking back.
I'm also somewhat surprised that no one has mentioned the ultimate
candidate for Sugar Daddy and how he took himself out of the running
by deciding to write a totally new Lisp altogether. Someone who wrote
extensive books on Common Lisp, started and then sold a multi-million
dollar software company on Lisp, and denigrates lesser languages by
calling them "Blub".
-jimbo
Pascal Costanza
> I hate the fact that this is how things are, but it's how things are.
I don't think it's a good idea to make things even more how they appear
to be, especially when you hate their very appearance.
Chris Russell
> As a relative newbie to Common Lisp, and specifically SBCL, I think
> the issues with open source CL implementations is mostly packaging.
> There are a lot of open source libraries out there that do most of the
> things as the libraries advertised by the commercial lisps. For me,
> Emacs (Aquamacs, actually) + SLIME is an outstanding development
> environment. Once you understand it, ASDF and ASDF Install is a
> pretty good package manager/installer.
Well there are things out there that integrate all of this by default.
http://www.cliki.net/Lisp%20in%20a%20Box
But maybe more could be done to make them more obvious.
Tim Bradshaw
>
> > I think Sun got this right. OTOH, they have enough money to afford
> > longer term investments.
>
> No, Sun just got lucky at one particular point in time. Java became such
> a success because of its bundling with Netscape Navigator and the
> implied promise of a new deployment model for applications (which never
> took off).
It's fairly important to remember that if any of the Lisp companies
had been through what Sun have they would probably be long, long
dead. Sun are doing better now, but they've been through a very long
run of losing money and their share price is a somewhat over 5% of
what it was at the peak. Sun really didn't get lucky (other than in
the sense that they had enough money in the bank to avoid some awful
takeover).
However, I think the other important thing is that Sun are a *systems*
company and have always been that. I think that means (among other
things) that their value to customers is not really some bit of
software, but in the whole bundle: from well-engineered, supportable
and scalable hardware (which is very profitable at the high end) to
service to software. Sun are *nothing like* a software company, and
that's why it makes sense to them to open source Solaris say.
--tim
(Disclaimer: Sun pay some of my wages at present).
Tim Bradshaw
> So this "most famous and wealthy university on earth" didn't have any
> commercial software? How did they manage without Mathematica, Gauss,
> Matlab, Stata and other similar programs? You must share the secret,
> I am sure that less famous and wealthy universities would like to save
> money on licenses too.
Not to mention the Oracle system that ran their payroll &c &c.
> So are you trying to imply that these people are stupid? (that's what
> I gather from your description -- spending time and money on a near
> zero ROI project qualifies as stupid). Hope you have better luck next
> time and find a company with sane people.
Always remember the the correct fix for stupid decisions is never
educating people so they can make smart ones. Instead do everything
you can to support the stupidity.
Ron Garret
"JShr...@gmail.com" <JShr...@gmail.com> wrote:
> > > > This experiment has already been done. OpenMCL is the result.
> >
> > > Interesting. Not being a PPC user I hadn't considered this, but
> > > apparently they've ported it to x86 Linux now as well
> >
> > And OS X as well. I run it on my Macbook Pro. 32-bit processors are so
> > last-millennium.
> >
> > > (although it's only in pre-released)
> >
> > You wanted open-source, this is what you get.
>
> Hmmm. Well, I didn't actually want open source in the sense of not
> professionally supported.
Clozure supports OpenMCL to the best of their ability. And they are
some of the most professional software engineers I have ever had the
privilege of working with.
> What I wanted (although you'd have to study
> the thread to get this) is a company that supports an excellent CL on-
> the-side and makes their money through verticals, like Google supports
> Guido to do python, and Sun/Java. So, MCL probably doesn't actually
> fit the bill afterall.
I see. You want someone to make a pile of money and then use some of it
to develop and maintain a product that you can use without paying. Is
that a fair restatement of your position?
rg
jimb...@gmail.com
wrote:> Well there are things out there that integrate all of this by default.http://www.cliki.net/Lisp%20in%20a%20Box
>
> But maybe more could be done to make them more obvious.
Actually, if you combined LispBox with Starter Pack:
into a single, double clickable application, I think you would answer
the vast majority of problems getting started with Common Lisp. It
would be a lot closer to what people get out of the box with Python
and Ruby.
I'd love to help, but unfortunately don't know enough yet to be much
use.
I actually did start with LispBox, and graduated to AquaMacs/SBCL. As
I mentioned previously the free Allegro version I downloaded with
LispBox hit the heap limit early on. AquaMacs has SLIME pre-
installed, which is pretty sweet.
For the Mac, a single bundle that properly installed SBCL, Aquamacs,
and the Starter Pack libraries, would rival most development
environments out there, Lisp or otherwise, IMHO. As long as someone
knows how to get around Emacs, that is.
-jimbo
Alex Mizrahi
(you :wrote :on '(Mon, 11 Jun 2007 18:31:30 +1000))
(
TX> I can't see this working with CL. As stated, I don't believe there is a
TX> big enough user base to offset the costs over a long enough period to
TX> possibly build a larger user base and I can't see the 'lock in'
TX> possibilities that Red Hat was able to get through working with other
TX> software vendors.
there are some products that work only with Allegro CL -- AllegroStore,
AllegroGraph, RACER (iirc), and possibly their web dev products live much
better on the 'real' Allegro.
so there are some killer products -- afaik AllegroGraph and RACER are the
best products for SemanticWeb. and market for Semantic Web products growth
as it gets more widespread.
so it can actually can make sense to make base product more available,
focusing on supporting software for building advanced next-generation web
applications.
but certainly Franz won't like to risk for possible significant expansion,
already having some stable income.
TX> The real problem is that large initial investment cost.
yup, it's obvious that if someone has lotsa money, he can do an interesting
experiment, that could possibly change landscape of programming language
market..
but it's very unlikely to find a source for money, given that it's a highle
risky
Rainer Joswig
"Alex Mizrahi" <udod...@users.sourceforge.net> wrote:
> (message (Hello 'Tim)
> (you :wrote :on '(Mon, 11 Jun 2007 18:31:30 +1000))
> (
>
> TX> I can't see this working with CL. As stated, I don't believe there is a
> TX> big enough user base to offset the costs over a long enough period to
> TX> possibly build a larger user base and I can't see the 'lock in'
> TX> possibilities that Red Hat was able to get through working with other
> TX> software vendors.
>
> there are some products that work only with Allegro CL -- AllegroStore,
> AllegroGraph, RACER (iirc), and possibly their web dev products live much
> better on the 'real' Allegro.
Racer runs in other Lisps, too. Racer is a product
of Racer Systems, Hamburg, Germany.
http://www.racer-systems.de/index.phtml?lang
>
> so there are some killer products -- afaik AllegroGraph and RACER are the
> best products for SemanticWeb. and market for Semantic Web products growth
> as it gets more widespread.
>
> so it can actually can make sense to make base product more available,
> focusing on supporting software for building advanced next-generation web
> applications.
>
> but certainly Franz won't like to risk for possible significant expansion,
> already having some stable income.
>
> TX> The real problem is that large initial investment cost.
>
> yup, it's obvious that if someone has lotsa money, he can do an interesting
> experiment, that could possibly change landscape of programming language
> market..
>
> but it's very unlikely to find a source for money, given that it's a highle
> risky
>
> )
> (With-best-regards '(Alex Mizrahi) :aka 'killer_storm)
> "I am everything you want and I am everything you need")
JShr...@gmail.com
> some of it to develop and maintain a product that you can
> use without paying. Is that a fair restatement of your position?
I'm not sure; it doesn't sounds like what I mean, so I guess perhaps
not. A fair restatement would be this: I want to invest in one of the
commercial Lisp vendors to repurpose them to develop a specific set of
semantic web verticals which I believe Lisp is uniquely positioned
for. That company would change their business model to build and sell
those verticals instead of building and selling Lisps. Since they will
have to have a well supported Lisp (which, of course, they already
have, but can't sell for nearly the margin that the verticals will
sell for), they will continue to support their Lisp for their own
purposes, and will significantly addition enhance that support by
spinning off the lisp as open source freeware. The implementors, now
working for a semantic web verticals company, not a Lisp products
company, although still doing more-or-less exactly what they did
before. (Note that I'm investing additional money to buy additional
engineers to add the Semantic Web products layer!) The result would be
a freeware lisp with a set of highly motivated, full time gatekeepers,
supported initially by my investment, and eventually by the income
from the semweb apps. Of course, there are a number of difficult
questions that immediate come to mind: What are the semweb apps? What
is the business plan basde upon these that could convince an investor
or set of investors to go for this? Unfortuntely, unless you are a
serious interested investor, I can't go into that (and even if you
were, we obviously wouldn't do it in this forum!) What I do not know,
and what the community might be able to help me figure out, is (are)
the question(s) that I started out asking: If one of the commercial
Lisps was made into open source freeware, etc. would the community be
likely to rally around that? I shouldn't have gone down the valuation
path for the commercial vendors; that was a red herring; as has been
pointed out, that's really a matter that has to be figured out by
analysis of the specific companies involved. I did want to get a sense
of what the community thought it might cost, and how much time it
would take, to do the same starting from one of the freeware bases.
The pointer for the new, more broadly useable OpenMCL is a very useful
one, for example, so I consider that I'm getting useful information
from this thread. (Again, given that asking for a valuation of Franz
or Lispworks was a stupid question for this forum!)
JShr...@gmail.com
help and/or time limits and licensing limitations (as you point out,
e.g., case 4!), they typically are not the "enterprise" editions, so
important connectivity and web development tools are missing.
(Although all of this changes variously, so my info on this could be
dated, but not by far; I fairly deeply explored this a couple of years
ago.)
On Jun 12, 8:54 am, "Dimiter \"malkia\" Stanev" <mal...@mac.com>
wrote:
> > the fact that this is how things are, but it's how things are.- Hide quoted text -
>
> - Show quoted text -
JShr...@gmail.com
> have any commercial software? How did they manage without
> Mathematica, Gauss, Matlab, Stata and other similar
> programs? You must share the secret, I am sure that less
> famous and wealthy universities would like to save
> money on licenses too.
First off, I didn't say that they didn't do it; there was just often a
serious battle. Also, on the other hand, you're correct that I was not
specific enough (which mislead others as well). Let me rephrase: It
has been my experience more often than not that decision makers who
normally would not balk at dropping a few thousand dollars for
hardware or travel, will often reject (or fight) paying for slightly
expensive programming language software based, I think, upon, among
other things, a series of misunderstandings of what "open source"
means, which are not improved by the fact that there are nearly
equivalent (in these decision makers' limited knowledge) programming
languages out there that (a) "everyone else it using", and (b) are
demonstrably free. Now, I'll admit that my experience here is limited
to Lisp, so I can't speak for how these decision makers decide to buy
Oracle or Mass Specs or send folks on trips for thousands of dollars
to accomplish thing that could mostly be done by email. Obviously they
do sometimes pay for software, but mostly these are incompariable
cases because MatLab, Mathematica, etc. are (for the most part)
unique, and do not have well respected freeware similar tools. Another
case that I do know of, R, which is basically a freeware SPlus has
essentially deleted the market for SPlus.
Tim X
> JShr...@gmail.com wrote:
>
>> I hate the fact that this is how things are, but it's how things are.
>
> I don't think it's a good idea to make things even more how they appear to be,
> especially when you hate their very appearance.
>
Things are more like they are now than they ever have been before!
--
tcross (at) rapttech dot com dot au
John Thingstad
> On Jun 12, 1:25 pm, Chris Russell <christopher.m.russ...@gmail.com>
> wrote:
>> On 12 Jun, 16:59, jimbo...@gmail.com wrote:
>> Well there are things out there that integrate all of this by
>> default.http://www.cliki.net/Lisp%20in%20a%20Box
>>
>> But maybe more could be done to make them more obvious.
>
> Actually, if you combined LispBox with Starter Pack:
>
> http://weitz.de/starter-pack/
>
> into a single, double clickable application, I think you would answer
> the vast majority of problems getting started with Common Lisp. It
> would be a lot closer to what people get out of the box with Python
> and Ruby.
Not sure what you mean by Lispbox.
From gigamonkey there is LispInABox (Peter Seibel.)
This is a EMACS with SLIME (Superior Lisp Interaction Mode) preinstalled.
Works with a number of free and comercial Lisp's.
However the starter pack you mention works with LispWorks.
It provides some SLIME functionality in LispWorks IDE + Web library,
regexp's etc.
Chris Russell
Ok, so a grossly simplistic version of your investment scheme, would
be:
1)Spend lots of money.
2)Give valuable things away to everyone.
3)Sell other things.
4)Profit.
And frankly you've lost me at the second half of step 2, why bother
giving your lisp away to everyone? I can see the value in bundling a
site licence for the lisp in with your semantic web goodness, and in
fact that is what some of the semi web guys offer (I think), but that
doesn't mean you need to open source it.
As investment schemes go, it makes me worry about losing the remaining
lisp companies to a coming web2.0 winter.
JShr...@gmail.com
> scheme, would be:
>
> 1)Spend lots of money.
> 2)Give valuable things away to everyone.
> 3)Sell other things.
> 4)Profit.
Yes, you're right. That's grossly simplistic. But let's run with it
for the moment...
> And frankly you've lost me at the second half of step 2,
> why bother giving your lisp away to everyone? I can see
> the value in bundling a site licence for the lisp in
> with your semantic web goodness, and in fact that is
> what some of the semi web guys offer (I think), but that
> doesn't mean you need to open source it.
>
> As investment schemes go, it makes me worry about losing the
> remaining lisp companies to a coming web2.0 winter.
And Iraq has weapons of mass destruction!
But scare phrase marketing aside, you apparently don't realize that
your grossly simplistic model is the same as the Open Source
(misunderstood as "freeware" -- i.e., free as in beer) model. "Why
would anyone give away valuable software?" Indeed, why would they?
Maybe if you answer that you'll be closer to answering some of your
other questions for yourself.
OTOH, since I don't personally think that the freeware model makes
sense, I won't argue very strongly for it. The only reason that it
might make sense in this case is that we're trying to chop our way out
through a forest of freeware competitors that operate according to
this model. I'm just tyring to find a path out of that forest.
Daniel Barlow
> OTOH, since I don't personally think that the freeware model makes
> sense, I won't argue very strongly for it. The only reason that it
> might make sense in this case is that we're trying to chop our way out
> through a forest of freeware competitors that operate according to
> this model.
No you're not. If the value is in the verticals, the freeware lisps are
not competing with you in the "valuable" space.
-dan
Chris Russell
You are not suggesting the the open source model, you have not talked
about selling support. Your idea as you've presented it is to take a
profitable company, and have them churn out the same goods, at the
same rate, and providing the same level of support for free.
You intend to finance this by selling a marvellous web 2.0 product,
which unfortunately is too secret to describe in public. And
apparently you want to do this because there are too many open-source
lisps out there already.
Jock Cooper
<snip>
> You intend to finance this by selling a marvellous web 2.0 product,
> which unfortunately is too secret to describe in public. And
> apparently you want to do this because there are too many open-source
> lisps out there already.
I don't think he means open-source lisps, I think he means other
open-source and/or free/low-cost dev tools such as python, ruby, perl,
java, visual studio, etc.
Raffael Cavallaro
> If one of the commercial
> Lisps was made into open source freeware, etc. would the community be
> likely to rally around that?
Some part of the lisp using community would, but others who are deeply
comitted to existing open source lisps might very well not. You might
want to look into merging an existing open source list (e.g., sbcl)
into an existing commercial lisp which you would buy and open source -
this would definitely increase mindshare. i.e., you'd have more of a
juggernaut with an open source lispworks-sbcl or allegro-sbcl than with
an open source lispworks or open source allegro alone.
Tim Bradshaw
wrote:
>
> As investment schemes go, it makes me worry about losing the remaining
> lisp companies to a coming web2.0 winter.
Although there will undoubtedly be some kind of web 2.0 winter, I
think it's clear that there is genuine value in a lot of web 2.0
stuff. TiddlyWiki, for instance, has changed the way I work to the
extent that when I recently started a contract where it wasn't going
to be realistic to run it on my desktop (bank, change control, locked
down machines etc) I went out and bought a laptop to run it without
thinking twice: even though I will not be able to connect the machine
to the bank's network it was still a no-brainer.
But the semantic web is kind of a different issue. It's just
astonishing how closely it is repeating the early AI hype: the idea
that once there is a basically sufficient representational system
(sexps then, XML now) everything else will just be easy. How wrong
they were then, how wrong they are now. I guess the hope must be
that, as before, some vast wash of government money will appear. This
is kind of moderately plausible as there's going to be a lot of
surveillance stuff coming out of the invented war on terror. But the
winter is going to be hard when it comes.
jimb...@gmail.com
> Not sure what you mean by Lispbox.
http://www.gigamonkeys.com/lispbox/
"Lispbox is a version of Lisp in a Box, which was originally created
by Matthew Danish and Mikel Evins, customized for use with Practical
Common Lisp."
> However the starter pack you mention works with LispWorks.
Exactly. If it worked with, say, SBCL, and you packaged the Starter
Pack interface and SBCL into a LispBox/in a Box distribution, I think
you would have something almost as good as the low end Lispworks and
Allegro CL products.
-jimbo
jimb...@gmail.com
> (long, complex plan for world domination snipped)
I still don't understand why you're proposing this long, complex plan
for buying out a commercial lisp distribution instead of just sprucing
up SBCL with an IDE and include batteries to get mostly where you want
to be. Can you be specific about what's so awful about SBCL besides
the lack of polish?
It would probably be easier to entice current SBCL developers to work
for you anyway, as they could then do for money what they currently do
only for love (although for some of the developers, you may have been
already beaten to it; I think I recall reading a SBCL developer's blog
who said he was hired explicitly to contine working on SBCL for pay).
So write up your secret semantic web business plan, get funded by
YCombinator (they certainly won't have a problem with you writing
software in Lisp), get an SBCL core developer as cofounder, dogfood
your own World Beating Lisp Development Environment while you get
rich, then release to the world.
Much simpler, right?
Don Geddis
> makes their money through verticals, like Google supports Guido to do
> python, and Sun/Java.
Ron Garret wrote:
>> I see. You want someone to make a pile of money and then use
>> some of it to develop and maintain a product that you can
>> use without paying. Is that a fair restatement of your position?
Ron's summary seems to match JShrager's wish list, to me at least.
"JShr...@gmail.com" <JShr...@gmail.com> wrote on Tue, 12 Jun 2007:
> I want to invest in one of the commercial Lisp vendors to repurpose them to
> develop a specific set of semantic web verticals which I believe Lisp is
> uniquely positioned for. That company would change their business model to
> build and sell those verticals instead of building and selling Lisps.
If you have a great business opportunity in semantic web verticals, why do
you need the Lisp company? Just license some existing Lisp, and go make your
huge profits.
Then, after you're rich, you can get into philanthropy. Perhaps you'd like
to save tigers and pandas. Or maybe you want to support some open source
Common Lisp.
But I really don't see any necessary connection between your philanthropic
desires, and some profitable business opportunity.
> Since they will have to have a well supported Lisp (which, of course, they
> already have, but can't sell for nearly the margin that the verticals will
> sell for), they will continue to support their Lisp for their own purposes,
> and will significantly addition enhance that support by spinning off the
> lisp as open source freeware.
But you agree that there are already commercial Lisp companies providing a
product of sufficiently high quality? Why do they need to be involved at all
in your new business venture?
> Of course, there are a number of difficult questions that immediate come to
> mind: What are the semweb apps? What is the business plan basde upon these
> that could convince an investor or set of investors to go for this?
Uh, yeah. That's always the hard part, isn't it? Creating a successful,
profitable business out of nothing.
> Unfortuntely, unless you are a serious interested investor, I can't go into
> that (and even if you were, we obviously wouldn't do it in this forum!)
Yet, everything else you've talked about is useless if you can't deliver on
this part of the problem. And we know that lots of people already want to
get rich, but few succeed. You really have zero credibility on this issue.
Which makes the rest of your concerns moot.
> What I do not know, and what the community might be able to help me figure
> out, is (are) the question(s) that I started out asking: If one of the
> commercial Lisps was made into open source freeware, etc. would the
> community be likely to rally around that?
> I did want to get a sense of what the community thought it might cost, and
> how much time it would take, to do the same starting from one of the
> freeware bases.
These are all silly hypotheticals without some separate, independent way to
make tons of money. Meanwhile, if you have that separate thing, then these
questions are all irrelevant anyway.
Unless you're already rich, and this is your idea of a charity. Why don't
you come back to us when that's the situation, and we'll solve it for you
then.
-- Don
_______________________________________________________________________________
Don Geddis http://don.geddis.org/ d...@geddis.org
I wish my name was Todd, because then I could say, "Yes, my name's Todd. Todd
Blankenship." Oh, also I wish my last name was Blankenship.
-- Deep Thoughts, by Jack Handey
Ron Garret
"JShr...@gmail.com" <JShr...@gmail.com> wrote:
> > I see. You want someone to make a pile of money and then use
> > some of it to develop and maintain a product that you can
> > use without paying. Is that a fair restatement of your position?
>
> I'm not sure; it doesn't sounds like what I mean, so I guess perhaps
> not.
Doesn't "perhaps not" mean the same as "perhaps so"?
> A fair restatement would be this: I want to invest in one of the
> commercial Lisp vendors to repurpose them to develop a specific set of
> semantic web verticals which I believe Lisp is uniquely positioned
> for.
Do *you* want to invest, or do you want *someone else* to invest? If
the former, why don't you just do it instead of wasting your time (not
to mention everyone else's) chatting about it?
> That company would change their business model to build and sell
> those verticals instead of building and selling Lisps. Since they will
> have to have a well supported Lisp (which, of course, they already
> have, but can't sell for nearly the margin that the verticals will
> sell for), they will continue to support their Lisp for their own
> purposes, and will significantly addition enhance that support by
> spinning off the lisp as open source freeware.
OK, but why does it have to be a commercial Lisp? Why can't you do what
you propose with SBCL or CLisp? There are lots of existence proofs that
commercial software is not a prerequisite to financial success (and
indeed, some evidence that it might actually be a hindrance).
> Of course, there are a number of difficult
> questions that immediate come to mind: What are the semweb apps?
Yeah. It's those damn annoying little details, like "how do we make
money?" that shoot down the brilliant ideas every time. Damned annoying.
> What
> is the business plan basde upon these that could convince an investor
> or set of investors to go for this? Unfortuntely, unless you are a
> serious interested investor, I can't go into that
See http://rondam.blogspot.com/2006/10/top-ten-geek-business-myths.html
myth #3.
> If one of the commercial
> Lisps was made into open source freeware, etc. would the community be
> likely to rally around that?
Since you ask, IMO the answer is pretty clearly no. The Lisp community
hasn't rallied around anything in twenty years, and as I pointed out
elsewhere in this thread, the experiment has already been done at least
once with (so far) negative results.
I think you are making a very fundamental strategic error: you are
conflating the mechanism by which you intend to make money with your
plans for what to do with that money once you've made it. I also think
that if your overarching desire is to unify the Lisp community then
you're tilting at a mighty big windmill.
rg
Kent M Pitman
> Ok, so a grossly simplistic version of your investment scheme, would
> be:
>
> 1)Spend lots of money.
> 2)Give valuable things away to everyone.
> 3)Sell other things.
> 4)Profit.
If this is correct, I tend to believe the claim more when the person
proposing it (a) has money to start with, (b) is the one planning to
give the money away, and (c) plans to profit on net when the endeavor
is done.
That is, incidentally, what the Lisp vendors have done... the ones
we're suggesting should be doing something else.
The thing is that when it's your own money, and you really need to
come out ahead, you're much more inclined to not ignore the details.
Tim Bradshaw
>
> If you have a great business opportunity in semantic web verticals, why do
> you need the Lisp company? Just license some existing Lisp, and go make your
> huge profits.
I expect better of you: this is an absolutely standard trick which is
played by people on themselves all the time. We all have many
versions of it, one of mine is a variant on "if only I had a '59 Les
Paul I could play like Jimmy Page". Well, you know that's bullshit:
what's stopping me from playing like that is not being willing to
practice 5-10 hours a day and fucking myself up with speed to be able
to do so, not being willing to spend the next 5 years in a van, and
finally lack of talent. But I'd rather not face those awkward truths.
Probably almost half of the traffic on cll is people doing this
trick. Just kill them.
Don Geddis
> In article <1181715211.3...@g37g2000prf.googlegroups.com>, "JShr...@gmail.com" <JShr...@gmail.com> wrote:
>> What is the business plan basde upon these that could convince an investor
>> or set of investors to go for this? Unfortuntely, unless you are a serious
>> interested investor, I can't go into that
>
> See http://rondam.blogspot.com/2006/10/top-ten-geek-business-myths.html
> myth #3.
Similar thoughts have been echoed by Paul Graham, for example here:
http://paulgraham.com/ideas.html
[People] overvalue ideas. They think creating a startup is just a
matter of implementing some fabulous initial idea. And since a
successful startup is worth millions of dollars, a good idea is
therefore a million dollar idea.
[...]
Actually, startup ideas are not million dollar ideas, and here's an
experiment you can try to prove it: just try to sell one. Nothing
evolves faster than markets. The fact that there's no market for
startup ideas suggests there's no demand. Which means, in the narrow
sense of the word, that startup ideas are worthless.
And also perhaps
http://paulgraham.com/start.html
An idea for a startup, however, is only a beginning. A lot of
would-be startup founders think the key to the whole process is the
initial idea, and from that point all you have to do is
execute. Venture capitalists know better. If you go to VC firms with
a brilliant idea that you'll tell them about if they sign a
nondisclosure agreement, most will tell you to get lost. That shows
how much a mere idea is worth. The market price is less than the
inconvenience of signing an NDA.
Another sign of how little the initial idea is worth is the number of
startups that change their plan en route. Microsoft's original plan
was to make money selling programming languages, of all things. Their
current business model didn't occur to them until IBM dropped it in
their lap five years later.
Ideas for startups are worth something, certainly, but the trouble
is, they're not transferrable. They're not something you could hand
to someone else to execute. Their value is mainly as starting points:
as questions for the people who had them to continue thinking about.
What matters is not ideas, but the people who have them. Good people
can fix bad ideas, but good ideas can't save bad people.
_______________________________________________________________________________
Don Geddis http://don.geddis.org/ d...@geddis.org
If you ever get whipped by a bullwhip, try to breathe _in_ as the whip is going
back, and _out_ as it hits your back. Or is it the other way around? Anyway,
you'll figure it out. -- Deep Thoughts, by Jack Handey [1999]
Cor Gest
wrote this mindboggling stuff:
(selectively-snipped-or-not-p)
> These "student" editions are not useful for real work; in addition to
> help and/or time limits and licensing limitations (as you point out,
Gosh, is that why they're called student-edition?
is it really true , that they really are meant for study ? ...
Cor
--
(defvar MyComputer '((OS . "GNU/Emacs") (IPL . "GNU/Linux")))
The biggest problem LISP has, is that it does not appeal to dumb people
If that fails to satisfy you, start reading the HyperSpec or woman frig
mailpolicy @ http://www.clsnet.nl/mail.php
Tayssir John Gabbour
> Whether a language is popular and whether it is available for free
> or not are two of the most superficial characteristics of a
> programming language. If a company bases its decisions on such
> superficialities, it's probably not in a very good shape. Companies
> have to spend so much money an assets anyway, so the comparatively
> small costs for software cannot possibly be a serious stumbling
> block.
I just asked someone, and was told (from imprecise memory) that Franz
wanted both:
* over 10k EUR per developer * platform * version
* over 5% revenue of any product using Franz's lisp
Not to single out proprietary Lisps; obviously our societies are
biased in favor of wealth, and sufficiently sane plans are always a
struggle. (Technically speaking, in this discussion we observe the
free rider problem of market economies.)
There are alternatives to private funding. For example, Lisp obviously
did well from government subsidy, during its most innovative periods.
Some countries offer significant subsidies for employers, like I'm
told the Netherlands does. (Or at least parts of it.)
Also, I would pledge significantly to a Lisp organization which was
run reasonably democratically and transparently, if I thought it had a
chance of effectively allocating resources to free software
production.
Tayssir
JShr...@gmail.com
> the Lisp community then you're tilting at a mighty big
> windmill.
Apparently so.
Kent M Pitman
Probably. But then, I don't think you've really got the data needed
to adopt what sounds like a pessimistic tone here.
I don't generally like speaking for others, but you seem to be having
trouble getting a gestalt on this discussion, and while I'd not
presume to speak for others, I'll tell you what I personally hear them
saying, though paraphrasing and aggregating according to my own personal
sense of things. (Being truly objective is hard, in other words.)
It seems to me that, yes, people have made some critical comments.
But I don't perceive that any of the comments have been intended to
say "don't try to find a way to help". I believe the comments have
been focused on "are you sure you're looking in the right place",
"what problem are you really trying to solve", etc.
People like to have help when they perceive it as help. But it's one
thing to say "how do I rearrange the icons on my desktop?" and have
someone show you. It's quite another to have someone walk in and say
"I'm here to help you clean up your desktop since it's obviously a
mess" when you've just finished setting up your icons in a way that
you like but you suddenly find that someone else doesn't.
You probably should have a discussion here on what you perceive the
problem is, if there is one, before you set off to solve it. You just
dove straight into a solution to an unstated problem, and no wonder
you got pushback.
If you skip that step, you'll end up thinking people are rejecting
your help (as I perceive you're saying above) when perhaps they don't
see it as help because they don't perceive the problem as you do.
It's much easier to evaluate the usefulness of a solution if you know
what problem it's trying to solve.
Look at how the X3J13 issues (attached to the Common Lisp HyperSpec)
are written. (I'm pretty sure it was Larry Masinter who forced us to
separate problem descriptions from proposals for solutions, and it was
a HUGE simplification of discussion.) If people can't agree on a
problem description, they have no business talking solutions, since
competing "solutions" to problems that are not agreed upon in the
first place cannot be evaluated usefully.. it just melts down with
people thinking the others are irrational because they can easily see
that the solution doesn't match their description of the problem.
It's very important to have a shared understanding of a problem if you
want to avoid that notion of people feeling others are irrational.
What the process we adopted while designing CL caused us to see was
that "the language lacks a foo construct" is NOT a proper problem
description, since linguistically/socially it only accommodates
solutions like "add a foo construct". It barely admits discussion
on the question of whether the problem description is wrong (after
all, it DOES lack a foo construct, so how COULD it be wrong? ... and
since it's a description of a problem, and the problem is clear, how
could that not be the answer? nothing to think about, right? let's
just do it.) Good problems are not described that way.
Good problem descriptions are written in a way that invite discussion
on the question of whether the problem is characterized correctly, and
whether there is a problem at all. Good problem descriptions allow
others to join the problem description with additional information
without fighting over the problem description. For example, if it's a
problem that person A wants to do tail recursive calls, person B can
also have that problem, or can want some other kind of related thing.
But person B cannot say "person A doesn't have a problem". So
eventually when everyone has spoken, you get to a place where the
problem description correctly says what the problem is.
Tim Bradshaw
<tayssir.j...@googlemail.com> wrote:
> On Jun 12, 11:48 am, Pascal Costanza <p...@p-cos.net> wrote:
>
> > Whether a language is popular and whether it is available for free
> > or not are two of the most superficial characteristics of a
> > programming language. If a company bases its decisions on such
> > superficialities, it's probably not in a very good shape. Companies
> > have to spend so much money an assets anyway, so the comparatively
> > small costs for software cannot possibly be a serious stumbling
> > block.
>
> I just asked someone, and was told (from imprecise memory) that Franz
> wanted both:
> * over 10k EUR per developer * platform * version
> * over 5% revenue of any product using Franz's lisp
1. I bet they negotiate, and imprecise memory is not the best source
of information.
2. There are other Lisp vendors and at least one of them has very
straightforward terms (no runtime fees).
Tayssir John Gabbour
> On Jun 14, 12:31 am, Tayssir John Gabbour
> > wanted both:
> > * over 10k EUR per developer * platform * version
> > * over 5% revenue of any product using Franz's lisp
>
> 1. I bet they negotiate, and imprecise memory is not the best source
> of information.
This was after negotiation. And I conservatively lowered the numbers
-- the numbers I was quoted were HIGHER than 10k and 5%. And I've
heard similar anecdotes about the over-5% revenue elsewhere.
In other forums, Franz representatives absolutely refuse to estimate
their pricing, arguing that you should contact them privately. Despite
repeated public questioning. So Franz is not a source of public
information here.
That said, I also spoke in person with a Franz salesman, and discussed
the issue with them on a general level. (We didn't talk about price
specifics, but rather the broad issue.) I won't go into specifics
about our conversation. But the reasoning, as I understand it, is the
most obvious one: differential pricing.
If anyone has better stats, I welcome them to offer it. I've observed
CEOs talk about this pricing privately, and they just found it weird.
Personally though, I can imagine that paying this cost is justifiable,
depending on your situation. So for example, I expect most people get
pretty low wages. (If they're even on a wage system.) Being in a
situation where you can afford to pay Franz's price isn't such a bad
situation after all, all things considered.
Tayssir
Tim X
Its interesting you make the comparison between the AI hype and what some may
call the semantic web hype. I too think there are a lot of similarities. some
see the concept of the semantic web as something great and new thats just
around the corner. However, people have been working on this and doing research
etc on the semantic web since not long after the whole web stuff kicked off. I
originally came across these ideas back in 95 and ironically, many of these
ideas came from the same people doing research in AI, particularly in the areas
of knowledge representation, ontologies, inductive reasoning etc.
I'm not discounting the potential of a semantic web, but have to put it into
the same box as many of the goals of AI - would be nice, but I'm not convinced
we understand things enough yet to get to that next level or can even agree
enough to see the sort of consistent adoption that would be required to move
the web from its rather primitive pattern matching indexing to something that
would provide more high level representations and potentially more
sophisticated links/relationships that could provide the basis for that next
level of searching and retrieval of data/information.
I do think we will move towards a semantic web and I also still believe that we
will make advances in areas that were traditionally thought of as AI, but I
think the progress will be a lot slower and take a lot longer than many are
arguing. Like AI, I think there are some fundamental issues which need to be
resolved and which many people seem to be glossing over. Until some of these
are resolved, I don't see any real progress being made - we will probably see
some advances and from the outside, many of them will possibly look quite
amazing. However, just like with the AI stuff, a lot of these 'advances' will
likely be little more than already existing solutions that were not practicle
originally, but because of increased processing power, lower storage costs and
faster networks, some of these previously impracticle solutions become
feasible.
Going back to the OPs original idea of getting a high quality commercial lisp,
making it open source and thereby hopefully bringing people back/over to lisp
from other open source languages - I can see where he is coming from and can
understand the reasoning. Possibly too many people have concentrated on the
costs and other aspects of such a proposal and missed the underlying
motivation. I thought the comment regarding trying to find a path through the
forest of all other open source languages was interesting as I think this could
be where the proposal has shaky foundations. I think the error here may be in
trying to find a new path in a forest that is already rich in paths.
Using a different analagy, Java, python, ruby et. al. are the new team
favorites. They are seen as young, fit, powerful and sexy. they are the home
town favorites with some high profile wins. CL is seen as old, a bit fat, slow
and out classed which hasn't had a high profile win in recent memory. giving
them new uniforms is unlikely to make any difference. Trying to compete with
the home town favorites on their own turf under their rules is very unlikely to
succeed.
What CL needs to do is get some high profile wins based on its strengths
playing by the rules it knows best and using the tactics and advantages it has
to get some high profile wins. this is likely to be in a new area (possibly
even the semantic web) where the tactics haven't already been set by the other
teams and possibly, if we are lucky, establish rules and expectations which
favor CL and put the young fast, but inexperienced and often brute force power
of the others at a disadvantage.
However, having said that, I'm extremely skeptical we will ever see CL with the
sort of uptake/popularity of these other languages. I find it far more likely
that CL will maintain a respectable level of support and will likely outlast
some of these newer languages, but I don't see it ever having the popularity of
these languages again (though I don't discount the possibility of a lisp like
language gaining considerable popularity at some point in the future).
I do feel that arguing that the main reason CL isn't more popular is because
there isn't a high quality open source implementation is too simplistic. The
open source languages aren't popular just because they are open source - they
are popular because they were/are as good/productive as anything that was/is
available commercially. In fact, I suspect Java, python and ruby would still be
more popular than CL even if the whole open sorce movement had never occured
and you had to pay for them (like we use to pay for Borland C, C++ and Prolog,
which I suspect were more popular than lisp back then as well - well, maybe not
prolog). other factors also need to be taken into account, such as what venture
capital is prepared to invest in, what is taught and the early experiences
programmers have with various paradigms, the attitudes of managers to
innovation and change, the types of problems we are trying to solve and the
differences in time to master (I think Cl is possibly the easiest language I've
ever learnt, but one of the hardest I've tried to master - this is likely to
impact on adoption as well). Any attempt to increase the market share of CL
needs to consider all of these factors (and others) and not just focus on one
aspect. A high quality open source CL is certainly a worthy goal, but its not
sufficient to somehow reserect CL and put it up there with today's popular
languages.
Tim
Charlton Wilbur
JS> But scare phrase marketing aside, you apparently don't realize
JS> that your grossly simplistic model is the same as the Open
JS> Source (misunderstood as "freeware" -- i.e., free as in beer)
JS> model. "Why would anyone give away valuable software?" Indeed,
JS> why would they? Maybe if you answer that you'll be closer to
JS> answering some of your other questions for yourself.
In the early days, because the software was developed as research
projects or as ancillary to other concerns, and was given away. This
is where BSD, early gcc, Emacs, early Linux, Perl, Apache, and the
like came from.
In the later days, because companies like IBM and Apple determined
that it was more cost-effective to take a solid open-source foundation
and contribute the specific improvements they needed to the public
good than it was to develop the whole operating system and toolchain
from the ground up. This is where later Linux and gcc come from.
Buying a closed-source in order to give it away and make money on
hypothetical poorly-defined "verticals" matches neither of these.
The root problem is that programmers need to eat. The successes of
open source happen where the programmers get to eat, because they are
getting paid by research grants or student loans, because they are
paid for doing something else and gave away this nifty tool they
created to make doing that thing easier, or because they can create
something they can sell for money for less because some of the work is
already done.
Your "buy a closed source Lisp and make it open source" fantasy does
not explain how the programmers will eat. This is a major failing.
Charlton
--
Charlton Wilbur
cwi...@chromatico.net
Charlton Wilbur
JS> It has been my experience more often than not that decision
JS> makers who normally would not balk at dropping a few thousand
JS> dollars for hardware or travel, will often reject (or fight)
JS> paying for slightly expensive programming language software
I've fought this fight before.
The principal difference is that the decision maker understands the
difference that paying a few thousand dollars for hardware or travel
makes. The decision maker knows that there are very good languages
and language implementations available for free, and so will balk at
spending any money for them. It's your job to articulate *why*
the commercial implementation is worth spending the extra money for.
"Because you spent $2000 for Joe and Bob to go to that conference in
Washington, DC" does not have any bearing on the merits of a
commercial software package. Neither does "Because you spent $5000
for the server that's currently not doing anything."
And, frankly, if you can't articulate to a decision maker why the
software package is worth the money, it's a reasonable conclusion that
you don't really need it.
Tim Bradshaw
wrote:
>
> This was after negotiation. And I conservatively lowered the numbers
> -- the numbers I was quoted were HIGHER than 10k and 5%. And I've
> heard similar anecdotes about the over-5% revenue elsewhere.
Remember, you're a Lisp person. That means you probably have the
negotiating skills of a bat (as have I before you think I'm being
nasty). My guess is you were likely negotiating from a position of
weakness such as being a small company or a university or something.
>
> If anyone has better stats, I welcome them to offer it. I've observed
> CEOs talk about this pricing privately, and they just found it weird.
>
It's not that uncommon a model in the non-toy software market.
--tim
Tayssir John Gabbour
> On Jun 14, 9:29 am, Tayssir John Gabbour <tayssir.j...@googlemail.com>
> wrote:
> > This was after negotiation. And I conservatively lowered the numbers
> > -- the numbers I was quoted were HIGHER than 10k and 5%. And I've
> > heard similar anecdotes about the over-5% revenue elsewhere.
>
> Remember, you're a Lisp person. That means you probably have the
> negotiating skills of a bat (as have I before you think I'm being
> nasty). My guess is you were likely negotiating from a position of
> weakness such as being a small company or a university or something.
I didn't negotiate; the costs were told to me by others. But I think
you're right, other than "getting to yes" by paying attention to
peoples' interests, I don't know that much about negotiation.
I doubt Franz plays hardball though. If you can offer them a deal that
makes sense to them, why should they reject it? Taking a chunk of your
revenue means they'd rather you make more revenue.
> > If anyone has better stats, I welcome them to offer it. I've observed
> > CEOs talk about this pricing privately, and they just found it weird.
>
> It's not that uncommon a model in the non-toy software market.
The last company I worked with had kind of similar pricing for one of
their products, actually. Somewhat ironically, they rejected Franz in
favor of Lispworks, partly because of the revenue thing... but also
because of Lispworks' CAPI.
But I suppose most of the people reading this can only afford "toy
software," and are presumably like children.
Tayssir
Ron Garret
Kent M Pitman <pit...@nhplace.com> wrote:
> Good problem descriptions are written in a way that invite discussion
> on the question of whether the problem is characterized correctly, and
> whether there is a problem at all. Good problem descriptions allow
> others to join the problem description with additional information
> without fighting over the problem description. For example, if it's a
> problem that person A wants to do tail recursive calls, person B can
> also have that problem, or can want some other kind of related thing.
> But person B cannot say "person A doesn't have a problem". So
> eventually when everyone has spoken, you get to a place where the
> problem description correctly says what the problem is.
OK, I'll bite since Jeff doesn't seem to want to pick up the ball...
Here's a problem statement:
The current landscape of Lisp implementations seems to have a big hole
in the middle, in the sense that there are a fairly large number of free
implementations, all of which are deficient in one way or another [1],
and a smaller number of commercial implementations which are less
deficient but all very expensive. There is no Lisp of commercial
quality which is affordable by the new breed of self-funded web 2.0
startups running on a shoestring, the result of which is that most such
startups are using Python, Ruby, or PHP instead. The result is a
vicious cycle where those startups, once they succeed, don't pump any
money back into the Lisp economy, which makes it even harder for vendors
to lower their prices.
And here's a proposed solution:
If I were starting a company I would approach the commercial Lisp vendor
of my choice and propose to them that they 1) give me licenses in
exchange for some equity in my company and 2) make it more widely known
that they were amenable to such arrangements. This is almost a no-lose
situation for the vendor because the incremental cost of the license is
zero, the opportunity cost is also most likely zero (since the
alternative is Python) but the upside potential is substantial. That,
it seems to me, is a much more defensible strategy for getting into
verticals that trying to raise a bunch of capital to open-source a
commercial implementation.
rg
jimb...@gmail.com
> In article <u3b0vf7tx....@nhplace.com>,
JShr...@gmail.com
> up the ball...
Sorry; I wasn't able to read the list again until just now...
> Here's a problem statement:
>
> The current landscape of Lisp implementations seems to
> have a big hole in the middle, in the sense that there are
> a fairly large number of free implementations, all of
> which are deficient in one way or another and a smaller
> number of commercial implementations which are less
> deficient but all very expensive. There is no Lisp of
> commercial quality which is affordable by the new breed
> of self-funded web 2.0 startups running on a shoestring,
> the result of which is that most such startups are using
> Python, Ruby, or PHP instead. The result is a vicious
> cycle where those startups, once they succeed, don't pump
> any money back into the Lisp economy, which makes it even
> harder for vendors to lower their prices.
Yes, that pretty much summarizes it.
> And here's a proposed solution:
>
> If I were starting a company I would approach the commercial
> Lisp vendor of my choice and propose to them that they
> 1) give me licenses in exchange for some equity in my
> company and 2) make it more widely known that they were
> amenable to such arrangements. This is almost a no-lose
> situation for the vendor because the incremental cost of
> the license is zero, the opportunity cost is also most
> likely zero (since the alternative is Python) but the
> upside potential is substantial. That, it seems to me,
> is a much more defensible strategy for getting into
> verticals that trying to raise a bunch of capital to
> open-source a commercial implementation.
This plan has several problems. First, you are not counting the
incremental costs of supporting all those free users. One of the
important things that the commercial vendors provide is support at all
levels. By open sourcing the Lisp, you spread this load across the
community. Deep problems may still have to be resolved by the vendor,
but many, probably most issues, could be resolved w/o having to go
through the vendor by way of the OS community -- or even oneself.
Second, your plan does not give folks who are not forming companies
any reason to use these quasi-open Lisps -- that is, those who are not
actually building companies: students, personal hackers, etc. The
companies would not have any reason to bother giving those away --
they have only downside (via support as above).
jimb...@gmail.com
>
> The current landscape of Lisp implementations seems to have a big hole
> in the middle, in the sense that there are a fairly large number of free
> implementations, all of which are deficient in one way or another [1],
> and a smaller number of commercial implementations which are less
> deficient but all very expensive. There is no Lisp of commercial
> quality which is affordable by the new breed of self-funded web 2.0
> startups running on a shoestring, the result of which is that most such
> startups are using Python, Ruby, or PHP instead. The result is a
> vicious cycle where those startups, once they succeed, don't pump any
> money back into the Lisp economy, which makes it even harder for vendors
> to lower their prices.
I'm not Jeff, but I'll agree that you describe a real problem.
> And here's a proposed solution:
Another solution: SBCL, Lisp-in-a-Box, Edi Weitz Starter Pack (ported
to SBCL), combined into a double click and run, self contained package
on Linux, Windows, Mac OS.
I guess the hard step is finishing the port of SBCL to Windows. Would
switching CLisp for SBCL solve that problem?
I suspect this is not very hard, just beyond my expertise.
Would such a distribution be lacking compared to commercial CLs?
Would it have enough "batteries included" to get a second look from
those tempted by Python and Rails?
I guess my proposal requires more capital and effort than your
solution, but less than Jeff's.
-jimbo
Tim X
My experience is that it is very common in the software market once you move
away from mass produced personal software. The last software package I had to
negotiate a purchase for at work started out at $250,000+ for the software and
$70,000 licensing/support per year. After considerable negotiation, we got what
we wanted for less than $120,000 for the package and $10,000 for licensing and
support per year. I've since found out through talking to others in the sector
who have also purchased this software that while the price I was able to get
wasn't the cheapest in the sector, it was one of the lowest and some other
businesses had payed way more. The cheap price I got wasn't mainly due to my
negotiation skills - it was mainly due to the fact the vendor wanted to get a
foothold in the sector and therefore was willing to drop their price
accordingly.
Tim
Tim X
I think you missed Kent's point. You cannot just say this is the problem and
here is the solution. We need to discuss/debate what the problem is until we
find agreement on the problem definition before putting forward a solution and
the problem needs to be presented in a way that encourages
discussion/contribution.
For example, I find no evidence in your problem definition that indicates the
small startup even considered lisp, little only discounting it because it was
too expensive. That seems to be somewhat of a leap. I actually suspect that few
startups even consider lisp as a possible development platform and never get as
far as looking at the costs of commercial implementations or even evaluate the
open source offerings. This is why I don't beleive a high quality open source
version would make a jot of difference. It is possible that a startup may come
to this conclusion if they did honestly evaluate/consider lisp, but I don't
believe that initial step is actually occuring. Therefore, I feel we need to
start with a much more basic problem definition/question - why is CL not being
considered as a viable development platform for startups?