SimH 3.8-0
Unknown
This includes a number of additional PDP11 peripherals, to support the
Unix First Edition reconstruction project; new options in the HP2100
series; the GRI-99 CPU; and the usual bug fixes and whatnot.
Please note the new instructions around the makefile, for non-Linux
systems.
/Bob Supnik
Simon Clubley
> V3.8-0 of SimH has been released.
>
Hello,
I have been trying to contact you, but the simh address on your website
bounces with a suggestion to use simh-owner, which I have done.
Do you have a better address for contact ?
--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world
Stan Barr
>V3.8-0 of SimH has been released.
>
>This includes a number of additional PDP11 peripherals, to support the
>Unix First Edition reconstruction project;
Tell me more...I run 5th Edition on my PDP-11 emulator box, earlier
versions would be most interesting!
--
Cheers,
Stan Barr t-bone .at. dial .dot. pipex .dot. com
The future was never like this!
Christian Corti
> In article <mqft54t11ro0v7g9e...@4ax.com>, Bob Supnik <> writes:
> I have been trying to contact you, but the simh address on your website
> bounces with a suggestion to use simh-owner, which I have done.
> Do you have a better address for contact ?
It seems that Bob doesn't want to receive any email,
or why does he violate the RFCs by specifing an empty address (reserved
for mail bounces) like in "Bob Supnik <>"? I wonder how the news server
could accept that...
Why doesn't he answer any mail sent to simh-owner (I've had the same
problem with the address on his web site)? Why does he put a "wrong"
address on his official web site? I can't see the point in doing that
(Spam is not an excuse).
Christian
jmfbahciv
I'm guessing...he doesn't want to be pestered. Why don't you ask
your question here?
/BAH
Chuck Harris
> It seems that Bob doesn't want to receive any email,
> or why does he violate the RFCs by specifing an empty address (reserved
> for mail bounces) like in "Bob Supnik <>"? I wonder how the news server
> could accept that...
First, RFC's are simply Requests For Comment. They are part of an evolving
internet. Compliance is suggested, but is voluntary. One of the greatest
strengths, and weaknesses, of the internet is you can go your own way.
Second, news servers can do pretty much whatever they like. They are one
of the easiest things on the internet to subvert and spoof.
> Why doesn't he answer any mail sent to simh-owner (I've had the same
> problem with the address on his web site)?
Maybe he is busy? Maybe he has a life besides SimH (Though judging by all
of the amazing work he has done on SimH, that seems unlikely ;-)
Schools in the US are out, and it is summertime, maybe he is on vacation?
Or, to tickle a possible tendency towards paranoia: Maybe he doesn't like you?
Why does he put a "wrong"
> address on his official web site? I can't see the point in doing that
> (Spam is not an excuse).
Well, it is his website, he can do whatever he wants to do. Since he
is the only arbiter of what excuses he may use, any or no excuse is a
valid one.
But it might be simply an oversight.
I maintain a number of websites that don't have my email address on them.
They stand as what they are, serving my purposes only. Comments aren't
welcome. If someone has a legal requirement for me to take something down,
they will find me.
-Chuck
Mark Crispin
> First, RFC's are simply Requests For Comment. They are part of an evolving
> internet. Compliance is suggested, but is voluntary. One of the greatest
> strengths, and weaknesses, of the internet is you can go your own way.
With the possible exception of the second sentence, the above statement is
utter nonsense.
It is like saying that DEC never built any analog circuitry, and never
created any software, because it was the Digital Equipment Corporation.
It is a fallacy to presume understanding the meaning of an acronym by
considering only the words in its expansion.
-- Mark --
http://panda.com/mrc
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.
Pat Farrell
> First, RFC's are simply Requests For Comment. They are part of an evolving
> internet. Compliance is suggested, but is voluntary. One of the greatest
> strengths, and weaknesses, of the internet is you can go your own way.
Amazingly ill informed.
Steve Crocker, who wrote RFC 1, told me that in the collaborative early
days, they didn't want to make it appear that some higher guru was
dictating things, so he/they picked the term RFC to be as gentle as
possible. But they are standards. Its how the internet is defined.
They are not even close to voluntary.
You can, if you wish, propose an new one that does something different.
And if your idea has merit, it might be accepted.
It takes a really great idea, a long time, and a lot of political skill
to turn your idea into an RFC. There is nothing voluntary about it
except that you volunteer to write a draft.
Even the late great Jon Postel didn't invent (most of) the RFCs, he
shepherd them to acceptance.
--
Pat Farrell
http://www.pfarrell.com/
Chuck Harris
> On Thu, 26 Jun 2008, Chuck Harris posted:
>> First, RFC's are simply Requests For Comment. They are part of an
>> evolving
>> internet. Compliance is suggested, but is voluntary. One of the
>> greatest
>> strengths, and weaknesses, of the internet is you can go your own way.
>
> With the possible exception of the second sentence, the above statement
> is utter nonsense.
Whatever. I feel the same about your generalization.
> It is like saying that DEC never built any analog circuitry, and never
> created any software, because it was the Digital Equipment Corporation.
Sometimes a clue can be had by knowing what an acronym stands for. Knowing
that DEC was Digital Equipment Corporation, would give you a very good clue
about what their primary business was (HINT: It wasn't analog.).
>
> It is a fallacy to presume understanding the meaning of an acronym by
> considering only the words in its expansion.
And when you make an assumption, well, you know.
I started programming the year after Steve Crocker submitted RFC #1,
the seminal RFC which started the process of documenting the ARPA network.
I have written numerous drivers based on one RFC or another. I know what they
are, and what strength they carry. It is a shame that you don't seem to.
-Chuck
If you want a clue as to how well RFC's are adhered to, check out
HTTP://www.RFC-ignorant.org. You will find tens of thousands of
instances where fairly important RFC conventions are ignored, and
have been for years.
Chuck Harris
> Chuck Harris wrote:
>> First, RFC's are simply Requests For Comment. They are part of an evolving
>> internet. Compliance is suggested, but is voluntary. One of the greatest
>> strengths, and weaknesses, of the internet is you can go your own way.
>
> Amazingly ill informed.
Hi Pat,
You see it your way, I see it mine. Yes, RFC's define the internet, and
yes, they are *supposed* to be followed to the letter, but are they in
truth?
When some programmer subverts the rules to try out a new technique, who sounds
the siren and arrests him and puts him in internet jail?
In some cases, RFC's are very rigorously held (because breaking them
breaks your use of the internet), but in others, such as the OP's complaint,
not at all.
At any given point in time, the internet is composed of software that covers a
wide range of RFC compliance. About as many as there are versions of
software out on the net.
> Steve Crocker, who wrote RFC 1, told me that in the collaborative early
> days, they didn't want to make it appear that some higher guru was
> dictating things, so he/they picked the term RFC to be as gentle as
> possible. But they are standards. Its how the internet is defined.
> They are not even close to voluntary.
OK, put on your siren hat, and enforce away. For a list of perps, a good
starting point is: http://www.rfc-ignorant.org
If you can break the rules, and the only power anyone has to stop you is to look
at you crossly and go, "Tut! Tut!", the rules are voluntary.
-Chuck
Allodoxaphobia
> Pat Farrell wrote:
>> Chuck Harris wrote:
>>> First, RFC's are simply Requests For Comment. They are part of an evolving
>>> internet. Compliance is suggested, but is voluntary. One of the greatest
>>> strengths, and weaknesses, of the internet is you can go your own way.
>>
>> Amazingly ill informed.
>
> Hi Pat,
>
> You see it your way, I see it mine. Yes, RFC's define the internet, and
> yes, they are *supposed* to be followed to the letter, but are they in
> truth?
If there isn't an RFC about hijacking a thread for some OT discussion
without changing the Subject:, there should be.
Pat Farrell
> You see it your way, I see it mine. Yes, RFC's define the internet, and
> yes, they are *supposed* to be followed to the letter, but are they in
> truth?
The wags say what's great about standards is that there are so many to
chose from.
> When some programmer subverts the rules to try out a new technique, who
> sounds the siren and arrests him and puts him in internet jail?
That's a bit silly. When a car maker designs the bolts to be all SAE
(american) and a supplier provides metric threads, are there sirens? No,
but the parts don't interoperate.
> OK, put on your siren hat, and enforce away. For a list of perps, a good
> starting point is: http://www.rfc-ignorant.org
I don't have to look, just see what's from Microsoft and their partners.
> If you can break the rules, and the only power anyone has to stop you is
> to look at you crossly and go, "Tut! Tut!", the rules are voluntary.
In a closed world, yes, you can hack your own. With M$ did many times.
But if you expect something to interoperate, say a PDP-8 with a PDP-11
and a PDP-10, then you have better pay attention.
I'm not sure what your larger point is. Thou shalt not kill is
mandatory, but is broken every day as well.
Mark Crispin
The difference is that my opinions are based upon fact and 30 years of
participation in the Internet standards process. Yes, from the time
before there was an Internet.
Al Gore did not invent the Internet. I did not either; but I worked with
the people who did. I made a few contributions of my own that some people
still find useful today.
I don't know who the hell you are, but you don't seem to have have much of
a presence in the Internet standards community.
> Sometimes a clue can be had by knowing what an acronym stands for. Knowing
> that DEC was Digital Equipment Corporation, would give you a very good clue
> about what their primary business was (HINT: It wasn't analog.).
BAH's claims to the contrary notwithstanding, DEC's primary business was
software and service. They practically gave away what they actually
manufactured (as opposed to rebranded OEM equipment). Perhaps the fact
that DEC didn't correctly identify its primary business helped lead to its
demise.
As for what they actually manufacturered, many important DEC CPUs used
analog logic. Take a look at the prints for the KA10 adder if you don't
believe me.
>> It is a fallacy to presume understanding the meaning of an acronym by
>> considering only the words in its expansion.
> And when you make an assumption, well, you know.
I don't make assumptions when it comes to RFCs.
> I started programming the year after Steve Crocker submitted RFC #1,
> the seminal RFC which started the process of documenting the ARPA network.
That doesn't impress me. I started programming the same year.
However, unlike you, I write RFCs. My count is over two dozen, many of
which are standards-track, and some dating back over 30 years.
Your name does not appear in any RFC author fields. Not a very good c.v.
for someone who makes presumptious statements about RFCs.
> I have written numerous drivers based on one RFC or another. I know
> what they are, and what strength they carry. It is a shame that you
> don't seem to.
Too bad that your limited experience as a driver author has not given you
much of a clue.
> If you want a clue as to how well RFC's are adhered to, check out
> HTTP://www.RFC-ignorant.org. You will find tens of thousands of
> instances where fairly important RFC conventions are ignored, and
> have been for years.
So? I have a long list of how various compilers violate the associated
standard; not to mention how many purported "POSIX" systems violate POSIX.
That doesn't in any way make these standards any less standard.
No guy with a badge and gun is going to to walk in and enforce standards
compliance.
I do not recommend using the rfc-ignorant.org web site as gospel. It
expresses the very selective interpretations by certain individuals of
some (but not all) RFCs. Consequently, some of their claims are valid,
but others are not.
Chuck Harris
> On Thu, 26 Jun 2008, Chuck Harris posted:
>> Whatever. I feel the same about your generalization.
>
> The difference is that my opinions are based upon fact and 30 years of
> participation in the Internet standards process. Yes, from the time
> before there was an Internet.
>
> Al Gore did not invent the Internet.
I don't much care for the man, but looking over the record, I can't find
any quote where he claimed to have invented the Internet. He did claim
to have understood the importance, and worked to create the funding that
made DARPA possible.
...
> I don't know who the hell you are, but you don't seem to have have much
> of a presence in the Internet standards community.
I have never claimed, or implied, that I had any presence in the internet
standards community. Why is that so important to you?
>> Sometimes a clue can be had by knowing what an acronym stands for.
>> Knowing
>> that DEC was Digital Equipment Corporation, would give you a very good
>> clue
>> about what their primary business was (HINT: It wasn't analog.).
...
> As for what they actually manufacturered, many important DEC CPUs used
> analog logic. Take a look at the prints for the KA10 adder if you don't
> believe me.
I think you had best stick to software, hardware doesn't seem to be your
forte. Just because a digital circuit uses transistors doesn't make it
analog logic. All digital logic uses analog circuitry when you take it
down to the transistor level.
>>> It is a fallacy to presume understanding the meaning of an acronym by
>>> considering only the words in its expansion.
>> And when you make an assumption, well, you know.
>
> I don't make assumptions when it comes to RFCs.
>
>> I started programming the year after Steve Crocker submitted RFC #1,
>> the seminal RFC which started the process of documenting the ARPA
>> network.
>
> That doesn't impress me. I started programming the same year.
>
> However, unlike you, I write RFCs. My count is over two dozen, many of
> which are standards-track, and some dating back over 30 years.
Good for you! I hope your career path has brought you fulfillment. I didn't
go down the standards-tract, I found the work to be stultifying. I work as
an engineer designing and inventing for a living. You worked on more than
two dozen RFC's over 30 years, and I lost count of the number of circuits
and programs I have worked on two decades ago.
> Your name does not appear in any RFC author fields. Not a very good
> c.v. for someone who makes presumptious statements about RFCs.
I never claimed it did. I have had plenty of time to observe the process,
however.
>> I have written numerous drivers based on one RFC or another. I know
>> what they are, and what strength they carry. It is a shame that you
>> don't seem to.
>
> Too bad that your limited experience as a driver author has not given
> you much of a clue.
How so? I followed what was written, and the drivers worked accordingly.
But more importantly, my customers got paid, and I got paid.
>> If you want a clue as to how well RFC's are adhered to, check out
>> HTTP://www.RFC-ignorant.org. You will find tens of thousands of
>> instances where fairly important RFC conventions are ignored, and
>> have been for years.
>
> So? I have a long list of how various compilers violate the associated
> standard; not to mention how many purported "POSIX" systems violate
> POSIX. That doesn't in any way make these standards any less standard.
I have never commented on whether the standard was more or less standard,
just that complying with an RFC up to the discretion of the implementor.
> No guy with a badge and gun is going to to walk in and enforce standards
> compliance.
That makes the compliance optional, doesn't it? Look at how well Microsoft
has complied with the various RFC's they have implemented.
> I do not recommend using the rfc-ignorant.org web site as gospel. It
> expresses the very selective interpretations by certain individuals of
> some (but not all) RFCs. Consequently, some of their claims are valid,
> but others are not.
At this point, RFC's cover such a wide area of expertise that they cannot
really be lumped into a single category.
I think this hijacked thread has gone far enough.
-Chuck
Al Kossow
>> Al Gore did not invent the Internet.
>
> I don't much care for the man, but looking over the record, I can't find
> any quote where he claimed to have invented the Internet. He did claim
> to have understood the importance, and worked to create the funding that
> made DARPA possible.
http://findarticles.com/p/articles/mi_m1568/is_1_31/ai_54956376
"Gore enters the picture a bit later - in 1987, when he supported a drive
by universities to expand funding for NSFNet. That drive became law in
the High Performance Computing Act of 1991, which gave about $1 billion
to high-performance networks and computers; about $150 million of the
funding was new money, with the rest consolidated from other programs."
Gordon Bell was at NSF at the time, and was one of the drivers of the
effort there.
Mark Crispin
>> As for what they actually manufacturered, many important DEC CPUs used
>> analog logic. Take a look at the prints for the KA10 adder if you don't
>> believe me.
> I think you had best stick to software, hardware doesn't seem to be your
> forte. Just because a digital circuit uses transistors doesn't make it
> analog logic. All digital logic uses analog circuitry when you take it
> down to the transistor level.
Apparently you aren't really such a great hardware engineer at all. If
you were, and had seen the prints for the KA10 adder (specifically the
KA10 adder), you might have a clue as to what I am talking about.
Sorry that your bluff didn't work.
> I work as
> an engineer designing and inventing for a living.
A support engineer for Time-Warner Cable in Washington DC, perhaps?
>> Too bad that your limited experience as a driver author has not given you
>> much of a clue.
> How so? I followed what was written, and the drivers worked accordingly.
> But more importantly, my customers got paid, and I got paid.
Lots of people get paid for writing garbage drivers. That doesn't mean
that it isn't garbage, especially when written by a cretin that thinks
that RFCs are "voluntary".
> I have never commented on whether the standard was more or less standard,
> just that complying with an RFC up to the discretion of the implementor.
As Pat Farrell said, that is a silly argument.
>> No guy with a badge and gun is going to to walk in and enforce standards
>> compliance.
> That makes the compliance optional, doesn't it? Look at how well Microsoft
> has complied with the various RFC's they have implemented.
Speaking as someone who has extensive experience with Microsoft and
standards compliance issues, I can asset that -- regardless of the
frustration caused -- most of Microsoft's compliance issues are bugs and
not some deliberate decision to violate a specification.
It is silly and ill-informed of you to claim that because Microsoft has
bugs, it is "optional" to comply with RFCs and that RFCs are somehow
different in force from other specifications because the expansion of the
RFC acronym is "Request For Comments".
What's more, the fact that bugs happen is far more widespread than just
Microsoft, and affects far more standards than just RFCs.
RFCs are as much standards as anything put out by ISO.
>> I do not recommend using the rfc-ignorant.org web site as gospel. It
>> expresses the very selective interpretations by certain individuals of some
>> (but not all) RFCs. Consequently, some of their claims are valid, but
>> others are not.
> At this point, RFC's cover such a wide area of expertise that they cannot
> really be lumped into a single category.
It doesn't work to weasel out when far more knowledgable people call you
out for uttering silly nonsense.
Mark Crispin
> I don't much care for the man, but looking over the record, I can't find
> any quote where he claimed to have invented the Internet. He did claim
> to have understood the importance, and worked to create the funding that
> made DARPA possible.
DARPA existed long before Al Gore was legally eligible to serve in
Congress.
I worked on ARPAnet -- the predecessor to Internet -- years before Gore
entered Congress in 1977.
The part that Gore played in expanding funding for NSFnet in 1991 had the
effect of removing DARPA from the picture. But the origins of NSFnet is
earlier than that; it was well underway by 1987. Gore simply jumped on a
bandwagon that had started moving years earlier.
glen herrmannsfeldt
(snip)
> The part that Gore played in expanding funding for NSFnet in 1991 had
> the effect of removing DARPA from the picture. But the origins of
> NSFnet is earlier than that; it was well underway by 1987. Gore simply
> jumped on a bandwagon that had started moving years earlier.
But would "the Internet as we know it" have developed as fast as it
did, and in the way it did, without him?
I remember the days of rules like connections between commercial
(.com) sites were not allowed to route on non-commercial (educational
and government) networks. Somewhere along the way that rule
changed, but I don't remember exactly how. Putting federal money
into a faster backbone allowed for the expansion of bandwidth
heavy applications (WWW). I also remember telnet connections with
two second response on good days (around 1987).
-- glen
jmfbahciv
<snip>
>> Sometimes a clue can be had by knowing what an acronym stands for.
>> Knowing
>> that DEC was Digital Equipment Corporation, would give you a very good
>> clue
>> about what their primary business was (HINT: It wasn't analog.).
>
> BAH's claims to the contrary notwithstanding, DEC's primary business was
> software and service.
You still don't get it, Mark. This is not _my_ claim. It was the
corporate business statement. Tradeoffs were decided based on
hardware business, not software and services. All software and services
development and maintenance schedules were driven by the _hardware_
schedule..not the other way around.
Assuming that the reverse happened is why you get things wrong.
<snip>
/BAH
Mark Crispin
> But would "the Internet as we know it" have developed as fast as it
> did, and in the way it did, without him?
Yes. Gore was not the only Congresscritter who recognized the value of
the nascent Internet. What Gore did was valuable; but if he had not done
it, someone else would have.
Congress was being actively lobbied in this direction, and Gore was merely
one who took the bait.
Mark Crispin
> You still don't get it, Mark. This is not _my_ claim. It was the
> corporate business statement. Tradeoffs were decided based on
> hardware business, not software and services. All software and services
> development and maintenance schedules were driven by the _hardware_
> schedule..not the other way around.
And that, in turn, was why DEC failed; corporate management did not
correctly identify the primary business.
Chuck Harris
> On Fri, 27 Jun 2008, Chuck Harris posted:
>>> As for what they actually manufacturered, many important DEC CPUs
>>> used analog logic. Take a look at the prints for the KA10 adder if
>>> you don't believe me.
>> I think you had best stick to software, hardware doesn't seem to be your
>> forte. Just because a digital circuit uses transistors doesn't make it
>> analog logic. All digital logic uses analog circuitry when you take it
>> down to the transistor level.
>
> Apparently you aren't really such a great hardware engineer at all. If
> you were, and had seen the prints for the KA10 adder (specifically the
> KA10 adder), you might have a clue as to what I am talking about.
>
> Sorry that your bluff didn't work.
I pulled the prints on the KA10, and sorry Mark, but it is binary
digital. The B138 adder module used to make the KA10 adder is composed of
transistors, diodes, resistors and capacitors, but it is a discrete version
of what is commonly known as DTL, or diode transistor logic. Discrete voltage,
and current levels are used to represent the binary logic levels.
All digital logic uses analog circuitry when you take it down to the
transistor level. The only exception I can think of is the futuristic
quantum logic gates, but they, and quantum computers, are still deep
into the R&D phase.
The KA10 is <<asynchronous>>, but even though both <<analog>> and
<<asynchronous>> start with the letter A, that doesn't make them the
same thing.
Stick to software, hardware doesn't seem to be your forte.
As for the rest of your invective, and ad hominem attacks, they are beneath
you.
-Chuck
Mark Crispin
> digital. The B138 adder module used to make the KA10 adder is composed of
> transistors, diodes, resistors and capacitors, but it is a discrete version
> of what is commonly known as DTL, or diode transistor logic.
Look again. You're missing something really important.
At least you identified the correct module.
kkt
> On Fri, 27 Jun 2008, glen herrmannsfeldt posted:
> > But would "the Internet as we know it" have developed as fast as it
> > did, and in the way it did, without him?
>
> Yes. Gore was not the only Congresscritter who recognized the value of
> the nascent Internet. What Gore did was valuable; but if he had not done
> it, someone else would have.
We don't know that. Lots of good ideas are proposed in Congress for
session after session but if no one really pushes for them they die
for lack of time.
We also could have ended up with a higher-fee, pay-per-packet model,
which would have changed how the internet is used.
-- Patrick
Mark Crispin
>>> But would "the Internet as we know it" have developed as fast as it
>>> did, and in the way it did, without him?
>> Yes. Gore was not the only Congresscritter who recognized the value of
>> the nascent Internet. What Gore did was valuable; but if he had not done
>> it, someone else would have.
> We don't know that. Lots of good ideas are proposed in Congress for
> session after session but if no one really pushes for them they die
> for lack of time.
Gore was not the only player in this game. There were many others.
> We also could have ended up with a higher-fee, pay-per-packet model,
> which would have changed how the internet is used.
That is certainly what the European PTTs wanted. In the late 1980s, we
all were required to pay lip service to GOSIP. The victory of TCP/IP (and
hence Internet) was much more due to the independent guerilla efforts of
sites worldwide who chose a technology that was available now and with
open source implementations.
Pat Farrell
> Gore was not the only player in this game. There were many others.
There is no point in arguing who was important, the facts are that he
never said he invented the internet, and he was a big supporter of the
Information Superhighway
> That is certainly what the European PTTs wanted. In the late 1980s, we
> all were required to pay lip service to GOSIP. The victory of TCP/IP
> (and hence Internet)
But, DEC was one of the major players pushing GOSIP aka the ISO/OSI
model. That was why DECnet was the way it was, as DEC made it comply
with, or at least play nice with OSI/ISO.
It was more than needing to pay lip service, there were GSA procurement
rules that said all systems purchased by the Feds had to talk Gosip
after a certain date, which I think was April 87 -- I may be fuzzy on
the date. If your system, be it Vaxen, Data General, etc., didn't speak
GOSIP, you had to do tons of bureaucratic red tape to justify why.
DEC drove it, perhaps with some others. IBM sure didn't want it, it
nearly killed Wang to try to comply.
Mark Crispin
> But, DEC was one of the major players pushing GOSIP aka the ISO/OSI
> model.
I know.
> It was more than needing to pay lip service, there were GSA procurement
> rules that said all systems purchased by the Feds had to talk Gosip
> after a certain date
I remember that all quite well. It led to the creation of a large body of
software that purportedly was GOSIP capable, but never actually worked
(much less interoperated) in other than a toy environment. It was quite
embarassing that France and Germany could not exchange X.400 email with
each other!
Note that FIPS 146-1, the first US government publication, wasn't until
1990; and the towel was thrown in by FIPS 146-2 in 1995. Parts of OSI
remain today, but the vast bulk of it was dead by 1996.
I recognized that GOSIP was doomed in the late 1980s, and I did my modest
part to undermine its credibility. The writing was plainly on the wall
for X.400 once we developed MIME. Certainly, none of us paid any
attention to X.400 at all (even lip service) after that point.
Mark Crispin
> I recognized that GOSIP was doomed in the late 1980s, and I did my modest
> part to undermine its credibility. The writing was plainly on the wall for
> X.400 once we developed MIME. Certainly, none of us paid any attention to
> X.400 at all (even lip service) after that point.
If you haven't read Carl Malamud's "Exploring the Internet: A Technical
Travelogue", I suggest you do. It explains, better than just about any
other source, why TCP/IP won and OSI did not.
Mark Crispin
> Travelogue", I suggest you do. It explains, better than just about any other
> source, why TCP/IP won and OSI did not.
Speaking of which, I see that it's online:
glen herrmannsfeldt
> Mark Crispin <m...@Panda.COM> writes:
(snip)
>>Yes. Gore was not the only Congresscritter who recognized the value of
>>the nascent Internet. What Gore did was valuable; but if he had not done
>>it, someone else would have.
> We don't know that. Lots of good ideas are proposed in Congress for
> session after session but if no one really pushes for them they die
> for lack of time.
> We also could have ended up with a higher-fee, pay-per-packet model,
> which would have changed how the internet is used.
That is what I was wondering about. As I said, there was a
time when commercial traffic (between any two .com sites) was
not allowed on .edu and .gov nets. It could have gone to a dual
(commercial/non-commercial) internet, with a pay per packet
model for the commercial net. The non-commercial
(government/academic) net government funded. I don't know who
was responsible for that not happening, though.
-- glen
Morten Reistad
I have understood what happened in DEC now.
Even if the DEC brass stated they sold iron only didn't mean
they were right. They were clueless. Normally such an organisation
would sink in weeks; but somehow DEC thrived, in direct violation
of management. That is the thing that is so unique about DEC.
The subculture of fixing things was so ingrained in DEC that they
survived. Engineers fixed stuff, and made it work. But they got
no credit, except from customers.
It couldn't last. I am still searching for how such a culture
arose; so different from e.g. Prime, which was virtually next door.
-- mrr
Morten Reistad
glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
>Mark Crispin wrote:
>(snip)
>
>> The part that Gore played in expanding funding for NSFnet in 1991 had
>> the effect of removing DARPA from the picture. But the origins of
>> NSFnet is earlier than that; it was well underway by 1987. Gore simply
>> jumped on a bandwagon that had started moving years earlier.
>
>But would "the Internet as we know it" have developed as fast as it
>did, and in the way it did, without him?
It got a funding boost, and was pushed to resovle the AUP issues.
But I would place Gordon Bell in a far more important role for
that to happen. OK; Gore was the first senior politician after
Boris Yeltsin and Carl Bildt to understand what the Internet was
about. The Queen was also an early adopter.
>I remember the days of rules like connections between commercial
>(.com) sites were not allowed to route on non-commercial (educational
>and government) networks. Somewhere along the way that rule
>changed, but I don't remember exactly how. Putting federal money
The CIX wars, which were won by The Good Guys (US; I was there),
in late january 1992. It got hot sometime in october 1991, and
the embryonic commercial Internet fought it's way to victory.
Free peering. Content neutrality. Back then the ISPs embraced
content neutrality to get carrier immunity under Common Law.
>into a faster backbone allowed for the expansion of bandwidth
>heavy applications (WWW). I also remember telnet connections with
>two second response on good days (around 1987).
As a founder of a large ISP, we fought tooth and nail to
get hold of international bandwidth. We had to use some satellite
capacity just to get the bits through: but it was never popular.
We tuned it so telnet didn't fly, but web, ftp and news did.
-- mrr
Morten Reistad
Mark Crispin <m...@Panda.COM> wrote:
>On Fri, 27 Jun 2008, Pat Farrell posted:
>> But, DEC was one of the major players pushing GOSIP aka the ISO/OSI
>> model.
>
>I know.
>
>> It was more than needing to pay lip service, there were GSA procurement
>> rules that said all systems purchased by the Feds had to talk Gosip
>> after a certain date
The feds were not alone. Europeans were in many respects worse.
>I remember that all quite well. It led to the creation of a large body of
>software that purportedly was GOSIP capable, but never actually worked
>(much less interoperated) in other than a toy environment. It was quite
>embarassing that France and Germany could not exchange X.400 email with
>each other!
In EUnet we ran what was probably the biggest X.400 network in the world
in terms of country coverage, peerings, and bandwidth-miles of connections.
We had to to enter public bids. It was all Marshall Rose's and friends' code.
If worked well enough for a rigged demo.
But it carried almost no traffic. The Internet did. We also ran a CLNS
network for it, and another for IS-IS. And we ran X.25 on top of TCP
a lot. All for silly requirements. But we fulfilled the requirements
better than almost anyone else.
>Note that FIPS 146-1, the first US government publication, wasn't until
>1990; and the towel was thrown in by FIPS 146-2 in 1995. Parts of OSI
>remain today, but the vast bulk of it was dead by 1996.
>
>I recognized that GOSIP was doomed in the late 1980s, and I did my modest
>part to undermine its credibility. The writing was plainly on the wall
>for X.400 once we developed MIME. Certainly, none of us paid any
>attention to X.400 at all (even lip service) after that point.
We dismantled most of the X.400 stuff in early 1997, except for a few
gateway servers to SMTP and back.
-- mrr
Rich Alderson
> It seems that Bob doesn't want to receive any email,
> or why does he violate the RFCs by specifing an empty address (reserved
> for mail bounces) like in "Bob Supnik <>"? I wonder how the news server
> could accept that...
Perhaps because the RFCs in question are talking about the *implementation* of
the Simple Mail Transfer Protocol, and not about J. Random User's use of the
implementation? An implementation of an Internet standard can be in violation
of the relevant RFC(s), but a human being cannot.
> Why doesn't he answer any mail sent to simh-owner (I've had the same
> problem with the address on his web site)? Why does he put a "wrong"
> address on his official web site? I can't see the point in doing that
> (Spam is not an excuse).
In point of fact, it is a very good excuse. In the early days of the World
Wide Web, before spiders and other web-crawling robots started gathering e-mail
addresses non-stop, it was a wonderful idea to include one's address on a web
page. In these latter fallen days, it's a stupid idea.
Perhaps Bob should put *nothing* there, but there is nothing in any Internet
RFC that says he has to put *anything* of *any* particular format, or that he
*must not* put something of some particular format, and he is old enough to
decide for himself.
If you, or the esteemed Simon Clubley, or anyone else wants to get in touch
with Bob, it's bery easy to do so: Subscribe to the SimH mailing list, and
post your query/suggestion/attaboy/complaint/whinge there. Bob reads the list,
as is proven by his responses on the list when they are called for.
But thanks for starting another yardstick war.
--
Rich Alderson "You get what anybody gets. You get a lifetime."
ne...@alderson.users.panix.com --Death, of the Endless
Eric Smith
> Amazingly ill informed.
>
> Steve Crocker, who wrote RFC 1, told me that in the collaborative early
> days, they didn't want to make it appear that some higher guru was
> dictating things, so he/they picked the term RFC to be as gentle as
> possible. But they are standards.
Only a very small number of RFCs are standards. As of January 2008,
there were fewer than 70 standards out of more than 5000 RFCs. A much
larger number are on the "standards track", but that is most certainly
NOT the same thing as being a standard. It's more like a recommended
practice. Many RFCs aren't even on the standards track.
> Its how the internet is defined.
Yes.
> They are not even close to voluntary.
No one is forced to conform to them, so they most certainly are voluntary.
If someone refuses to conform, the worst that usually happens is that they
are ostracized.
By comparison, there are many standards from other national standards
bodies that are mandatory (based on laws and regulations), and for which
nonconformance can result in fines.
Eric
Christian Corti
> Perhaps Bob should put *nothing* there, but there is nothing in any Internet
> RFC that says he has to put *anything* of *any* particular format, or that he
> *must not* put something of some particular format, and he is old enough to
> decide for himself.
RFC 1036 chapters 2.1. and 2.2.
In short: you are *required* to put *something* in a *particular* format.
Christian
Guesser
oh noes, it's not compliant, so what're you gonna do?
jmfbahciv
> On Fri, 27 Jun 2008, jmfbahciv posted:
>> You still don't get it, Mark. This is not _my_ claim. It was the
>> corporate business statement. Tradeoffs were decided based on
>> hardware business, not software and services. All software and services
>> development and maintenance schedules were driven by the _hardware_
>> schedule..not the other way around.
>
> And that, in turn, was why DEC failed; corporate management did not
> correctly identify the primary business.
That's what I've been writing all along. And it wasn't corporate
management; it was Gordon Bell's doing. During the last half of
the 1970s, he got rid of any VP that had any software background.
What we did was work around his abject stupidity, cleaning up
after him whenever he opened his mouth in public. The
last mess he made, we couldn't clean up and he did nothing
to fix it.
Now, all software ships were driven by hardware schedules. Period.
No matter how much you protest, that was the reality of the
business. We did the best we could and managed to keep DEC
functioning until the BoD decided to strip all assets out of
the company and sell the piece that Compaq wanted. Unfortunately,
the piece that Compaq wanted required the infrastructure of
soft/hardware development to keep it useful. Dumping that
infrastructure, spectacularly finished off by Carleybaby,
made the help desk a blathering nonentity.
/BAH
/BAH
jmfbahciv
> In article <NZ-dnWlSNs6TRPnV...@rcn.net>,
> jmfbahciv <jmfbahciv@aol> wrote:
>> Mark Crispin wrote:
>>
>> <snip>
>>
>>>> Sometimes a clue can be had by knowing what an acronym stands for.
>>>> Knowing
>>>> that DEC was Digital Equipment Corporation, would give you a very good
>>>> clue
>>>> about what their primary business was (HINT: It wasn't analog.).
>>> BAH's claims to the contrary notwithstanding, DEC's primary business was
>>> software and service.
>> You still don't get it, Mark. This is not _my_ claim. It was the
>> corporate business statement. Tradeoffs were decided based on
>> hardware business, not software and services. All software and services
>> development and maintenance schedules were driven by the _hardware_
>> schedule..not the other way around.
>>
>> Assuming that the reverse happened is why you get things wrong.
>
> I have understood what happened in DEC now.
>
> Even if the DEC brass stated they sold iron only didn't mean
> they were right. They were clueless. Normally such an organisation
> would sink in weeks; but somehow DEC thrived, in direct violation
> of management. That is the thing that is so unique about DEC.
Thank you.
>
> The subculture of fixing things was so ingrained in DEC that they
> survived. Engineers fixed stuff, and made it work. But they got
> no credit, except from customers.
This was where DECUS was vital. Those engineers wouldn't have had
any idea how to fix it "correctly" without knowing what their
customers needed (not wanted, needed). There existed a few
savvy customers who groomed JMF and TW (I'm only talking about
TOPS-10 business, now...I'm sure the same thing happened in the
other product lines). They would keep an ongoing conversation
through phone calls throughout the year.
Whenever Bell, or anybody else, made a mess, these customers
would call inhouse to find out the truth.
>
> It couldn't last. I am still searching for how such a culture
> arose; so different from e.g. Prime, which was virtually next door.
Hint: hippies. ;-) And a corporate attitude that allowed mistakes
and didn't punish people for working on "midnight" hacks.
As long as we did what we were paid to do, we could do anything
else. The successful attempts made it to the product lines.
/BAH
Simon Clubley
>
> If you, or the esteemed Simon Clubley, or anyone else wants to get in touch
> with Bob, it's bery easy to do so: Subscribe to the SimH mailing list, and
> post your query/suggestion/attaboy/complaint/whinge there. Bob reads the list,
> as is proven by his responses on the list when they are called for.
>
Thanks for the suggestion.
Until I dug a little deeper into the website just now, I didn't realise that
it was a public mailing list as there's no such indication on the pages I
looked at; I just thought that Bob Supnik had come up with a unique way of
making sure that humans and not spam-ridden machines contacted him.
I'll join the mailing list.
Simon.
--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world
Pat Farrell
> They were clueless. Normally such an organisation
> would sink in weeks; but somehow DEC thrived, in direct violation
> of management. That is the thing that is so unique about DEC.
"thrived' is way too strong. They had the Vax, it was an amazing cash
cow. A better term would be that they survived far longer than anyone
would expect.
By 85/86, they were dead, but still walking.
jmfbahciv
> Morten Reistad wrote:
>> They were clueless. Normally such an organisation
>> would sink in weeks; but somehow DEC thrived, in direct violation
>> of management. That is the thing that is so unique about DEC.
>
> "thrived' is way too strong.
No, it's not.
> They had the Vax,
We had PDP-11s.
>it was an amazing cash
> cow. A better term would be that they survived far longer than anyone
> would expect.
>
> By 85/86, they were dead, but still walking.
Corporate seemed to be starting to think in terms of systems rather
than hardware with a little software added in sometime during
the Alpha development phase. VMS had evolved to the point of
being useful.
/BAH
>
Morten Reistad
Until about 1981-82 I would call DEC Thriving without blinking.
They were growing, profitable and had a strong, loyal customer
base in segments some salesmen would kill for.
One thing that doesn't figure is why they didn't embrace
the "alternate software" if they were so intent on selling
just hardware. Unix, ITS, Tenex etc. developed, sold and
supported by third parties should fit the DEC policy like
a glove. yet, it didn't.
-- mrr
Morten Reistad
Prime, DG and Wang existed in the same environment, took the
MIT graduates from the same classes, hired hippes from the same
concerts, but they turned out very different.
Did you have any contact with the competition? They were
located within walking distance from the core DEC facilities
in central Massachussets too.
-- mrr
Mark Crispin
> One thing that doesn't figure is why they didn't embrace
> the "alternate software" if they were so intent on selling
> just hardware. Unix, ITS, Tenex etc. developed, sold and
> supported by third parties should fit the DEC policy like
> a glove. yet, it didn't.
Prior to some time in the 1970s, it did. However, at that time someone at
DEC realized that the money was in software and services, and those
customers who did not buy such from DEC were not profitable.
Bill Gunshannon
Morten Reistad <fi...@last.name> writes:
> In article <rd2dnZakzJTqrPvV...@rcn.net>,
> jmfbahciv <jmfbahciv@aol> wrote:
>>Morten Reistad wrote:
>>> In article <NZ-dnWlSNs6TRPnV...@rcn.net>,
>>> jmfbahciv <jmfbahciv@aol> wrote:
>>>> Mark Crispin wrote:
>>>>
>
>>>
>>> It couldn't last. I am still searching for how such a culture
>>> arose; so different from e.g. Prime, which was virtually next door.
>>
>>Hint: hippies. ;-) And a corporate attitude that allowed mistakes
>>and didn't punish people for working on "midnight" hacks.
>>
>>As long as we did what we were paid to do, we could do anything
>>else. The successful attempts made it to the product lines.
>
> Prime, DG and Wang existed in the same environment, took the
> MIT graduates from the same classes, hired hippes from the same
> concerts, but they turned out very different.
I can't speak for DG or Wang, but Prime was a bad ewxample for comparison.
Started after DEC, had a much superior product to the VAX at that point
in time. And dependingon what you consider their actual demise, lasted
either 9 or 6 years shorter than DEC. Unless by "different" you actually
meant less successful. :-)
bill
--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bill...@cs.scranton.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>
Pat Farrell
> Prime, DG and Wang existed in the same environment, took the
> MIT graduates
And the best from UMass, and others schools up there. The old 128
circuit was an amazing place.
> Did you have any contact with the competition? They were
> located within walking distance from the core DEC facilities
> in central Massachussets too.
I'm not Barb, but I did, worked with DEC, DG, and Wang folks. All were
very smart, but rarely talked to folks at other firms.
Even Wang, up in Lowell, was not far from the home of LCG
Pat Farrell
> Pat Farrell wrote:
>> They had the Vax,
> We had PDP-11s.
And they were important in 85/86 in what way?
As the original expensive Vaxen were replaced with later versions,
everyone I knew ran running and screaming from TKB and the other curses
of 16 bit code.
The 11's did live a long time in embedded and custom boxes.
I don't remember when the microVax hit the market, I know it was mass
produced by late 80s. There was not much point in a general purpose 11
once the uVax arrived.
Morten Reistad
Bill Gunshannon <bill...@cs.uofs.edu> wrote:
>In article <5p284g....@eden.reistad.name>,
> Morten Reistad <fi...@last.name> writes:
>> In article <rd2dnZakzJTqrPvV...@rcn.net>,
>> jmfbahciv <jmfbahciv@aol> wrote:
>>>Morten Reistad wrote:
>>>> In article <NZ-dnWlSNs6TRPnV...@rcn.net>,
>>>> jmfbahciv <jmfbahciv@aol> wrote:
>>>>> Mark Crispin wrote:
>>>>>
>>
>>>>
>>>> It couldn't last. I am still searching for how such a culture
>>>> arose; so different from e.g. Prime, which was virtually next door.
>>>
>>>Hint: hippies. ;-) And a corporate attitude that allowed mistakes
>>>and didn't punish people for working on "midnight" hacks.
>>>
>>>As long as we did what we were paid to do, we could do anything
>>>else. The successful attempts made it to the product lines.
>>
>> Prime, DG and Wang existed in the same environment, took the
>> MIT graduates from the same classes, hired hippes from the same
>> concerts, but they turned out very different.
>
>I can't speak for DG or Wang, but Prime was a bad ewxample for comparison.
>Started after DEC, had a much superior product to the VAX at that point
>in time. And dependingon what you consider their actual demise, lasted
>either 9 or 6 years shorter than DEC. Unless by "different" you actually
>meant less successful. :-)
Prime was software-based, sales driven and controlled from financial
investors from day 1. Their demise was when their financial games
backfired; but that was the financial markets and the Prime financiers
that interacted; Prime (Or CV) was not directly involved in it's own
demise. They had a sane response to the killer micros; in pushing
software and gracefully unwinding the 50-series.
There was a lot of frustration internally that around 3 out of 5
development projects were canned when the prototype was ready; but
this seems a pretty sound way to operate a company in an environment
such as Prime/DEC was in. You needs innovation, and lots of it, but
you need to be very selective where you put your resources.
So, Prime was the "opposite number" from DEC in many respects.
I just wondered it there was any kind of interaction between DEC
and Prime/DG/Wang, like there always is in Silcon Valley.
-- mrr
Bill Gunshannon
Mentec, who took it over from DEC, only recently has stopped offering
any apparent further support, much to the chagrin of a lot of people.
And, it was only a few years ago they stopped developing the CPU line
itself.
Carl Appellof
"Bill Gunshannon" <bill...@cs.uofs.edu> wrote in message
news:6cpr0oF...@mid.individual.net...
> I can't speak for DG or Wang, but Prime was a bad ewxample for comparison.
> Started after DEC, had a much superior product to the VAX at that point
> in time. And dependingon what you consider their actual demise, lasted
> either 9 or 6 years shorter than DEC. Unless by "different" you actually
> meant less successful. :-)
>
I gotta disagree with you on the "much superior product to the VAX part.
When I was a grad student in the U.W. Chemistry dept, we started looking for
a 32-bit or better minicomputer with virtual memory. Prime already had a
product, the VAX was about to be born, and the 2020 was lurking in the wings
too. I took a good look at the Prime product at the time. I may be able to
count the number of compatibility mode bits on one hand, but I might have to
use all the digits to do so. The O/S looked OK, but when we sniffed VMS and
the VAX, we could tell it was a much more satisfying solution.
Sadly, the 2020 was too underpowered to support the number of users we had,
and was obviously out of the mainstream, even then, even though we had
other -10s on campus to compare it to.
Carl
Carl Appellof
"Mark Crispin" <m...@Panda.COM> wrote in message
news:alpine.OSX.1.10.0...@pangtzu.panda.com...
>
> It is like saying that DEC never built any analog circuitry, and never
> created any software, because it was the Digital Equipment Corporation.
>
> It is a fallacy to presume understanding the meaning of an acronym by
> considering only the words in its expansion.
>
It may have built analog circuitry, and certainly created software, but in
the end, is no longer a Corporation :-(
Carl
Rich Alderson
I'm sorry, I thought we were talking about someting on his web pages. I went
back and looked, and see that I was mistaken.
Bad Bob. Bad. No biscuit.
Pat Farrell
> Pat Farrell <pfar...@pfarrell.com> writes:
>> The 11's did live a long time in embedded and custom boxes.
>
OK, how many of them are other than embedded and custom boxes?
There were a huge number of OEMs that put application software on 11s,
including AMS. But we moved quickly to the vaxen in the mid-80s.
Don't the basic CPU live on, at least in spirit, in a lot of embedded
microcontrollers?
jmfbahciv
None. I don't remember hearing the word Wang until TW went
to work for them. But I was kept isolated from most outside
exposures.
/BAH
jmfbahciv
Cutler.
/BAH
jmfbahciv
> On Sun, 29 Jun 2008, Morten Reistad posted:
>> One thing that doesn't figure is why they didn't embrace
>> the "alternate software" if they were so intent on selling
>> just hardware. Unix, ITS, Tenex etc. developed, sold and
>> supported by third parties should fit the DEC policy like
>> a glove. yet, it didn't.
>
> Prior to some time in the 1970s, it did. However, at that time someone
> at DEC realized that the money was in software and services,
Not software, support. And that was Jack Shields. His goal
was to have field support services make US$5 billion/year
in [emoticon counts on fingers] I think he said 8 years when
we talked. I think we had that conversation in 1972. So it
was before he became a VP.
<snip>
/BAH
jmfbahciv
/BAH
jmfbahciv
> jmfbahciv wrote:
>> Pat Farrell wrote:
>>> They had the Vax,
>> We had PDP-11s.
>
> And they were important in 85/86 in what way?
We weren't talking 85/86. We were talking about when DEC was
thriving. You keep mapping the DEC biz of 85/86 to 75/76.
They were completely different worlds.
<snip>
/BAH
jmfbahciv
> In article <4vQ9k.9453$i55...@newsfe22.lga>,
> Pat Farrell <pfar...@pfarrell.com> writes:
>> jmfbahciv wrote:
>>> Pat Farrell wrote:
>>>> They had the Vax,
>>> We had PDP-11s.
>> And they were important in 85/86 in what way?
>> As the original expensive Vaxen were replaced with later versions,
>> everyone I knew ran running and screaming from TKB and the other curses
>> of 16 bit code.
>>
>> The 11's did live a long time in embedded and custom boxes.
>>
>> I don't remember when the microVax hit the market, I know it was mass
>> produced by late 80s. There was not much point in a general purpose 11
>> once the uVax arrived.
>
> Tell that to all the people still running PDP-11's. :-)
>
> Mentec, who took it over from DEC, only recently has stopped offering
> any apparent further support, much to the chagrin of a lot of people.
> And, it was only a few years ago they stopped developing the CPU line
> itself.
>
> bill
>
/BAH
jmfbahciv
> Christian Corti <Christi...@studserv.uni-stuttgart.de> writes:
>
>> In alt.sys.pdp11 Rich Alderson <ne...@alderson.users.panix.com> wrote:
>
>>> Perhaps Bob should put *nothing* there, but there is nothing in any Internet
>>> RFC that says he has to put *anything* of *any* particular format, or that
>>> he *must not* put something of some particular format, and he is old enough
>>> to decide for himself.
>
>> RFC 1036 chapters 2.1. and 2.2.
>> In short: you are *required* to put *something* in a *particular* format.
>
> I'm sorry, I thought we were talking about someting on his web pages. I went
> back and looked, and see that I was mistaken.
>
> Bad Bob. Bad. No biscuit.
<grin> I never thought of biscuits. Do biscuits work?
/BAH
Morten Reistad
Were there no hacker bars? User group events? other gatherings?
These are the order of the day here, and appearantly in California.
-- mrr
Chuck Harris
> On Fri, 27 Jun 2008, Chuck Harris posted:
>> I pulled the prints on the KA10, and sorry Mark, but it is binary
>> digital. The B138 adder module used to make the KA10 adder is
>> composed of
>> transistors, diodes, resistors and capacitors, but it is a discrete
>> version
>> of what is commonly known as DTL, or diode transistor logic.
>
> Look again. You're missing something really important.
You are right, I did! Thanks for pointing that out.
I spent a couple of lunchtimes analyzing the circuit, and what a
delight it is. First, it clearly meets the definition of being
purely digital. The logic levels are strictly defined voltage
regions. Since digital is a subset of analog, it is also analog,
in the same way as all digital circuits. But, it is definitely
not just binary! Parts of it are trinary, and the carry signals
are quatrinary (though only used as binary).
So, how does it work?
The CI, and CO ports are 4 state in nature. They represent the
number of binary 1's on the A,B, and CI inputs.
They are encoded as voltage levels:
0 = -1.2V, 1 = -2.5V, 2 = -3.5V, and 3 = -4.5V
The A and B inputs are converted internally to trinary levels:
(0+0) = +0.5V, (0+1) or (1+0) = +1.5V, and (1+1) = +3.0V
The output of the A + B trinary converter is threshold compared
to a 1.5V supply, and that signal, after some level shifting, is
compared to the CI signal to make the binary SUM output. It is
also used to selectively shift up the CO signal by 1V.
Pretty neat, mo more parts than a binary solution, very fast,
and has good safety margins on all of the logic levels.
>
> At least you identified the correct module.
Of course.
-Chuck
Mark Crispin
> I spent a couple of lunchtimes analyzing the circuit, and what a
> delight it is. First, it clearly meets the definition of being
> purely digital.
Yet, arguably, it also meets the definition of being analog, by more than
just "digital is a subset of analog". The techniques used by that circuit
would be very familiar to anyone (sadly, now mostly old farts) who had
worked with analog computers.
I deliberately use the word "arguably". It is not at all unreasonable to
claim (as you do) that it is "purely digital". But neither is it
unreasonable to claim that it is analog (although obviously not "purely
analog").
It's a hybrid; and is very, very neat!
Another cute thing is the behavior of D (Carry Insert) which changes the
effect of CI (Carry Input). To avoid leaving you in suspense: the adder
is also used for TxCx, XOR, and EQV!
It's amazing what sorts of tricks people came up with in the 1960s. It's
obvious that it was done that way for performance reasons; as you said it
is a very fast circuit for what was available at the time.
When I bought the wreckage of what was once SAIL's KA10 (sadly, it was
thoroughly toasted two decades ago), I kept the front panel, meter panel,
light screens, BBN pager, all the B138 boards, and the B167 (hey, it was
part of the adder). The rest went to a scrap metal dealer for gold
reclamation.
The BBN pager went to the Swedish hackers (when last I heard, it was part
of a working Tenex system), and over the years some B138 boards have been
given to friends as souvenirs. I still have several B138 boards and, of
course, the front panel & etc.
Pat Farrell
> Morten Reistad wrote:
>> Until about 1981-82 I would call DEC Thriving without blinking.
>> They were growing, profitable and had a strong, loyal customer
>> base in segments some salesmen would kill for.
>> One thing that doesn't figure is why they didn't embrace
>> the "alternate software" if they were so intent on selling
>> just hardware. Unix, ITS, Tenex etc. developed, sold and
>> supported by third parties should fit the DEC policy like
>> a glove. yet, it didn't.
>
> Cutler.
Don't mention that name. His crap still curses us
Chuck Harris
Mark Crispin wrote:
> On Mon, 30 Jun 2008, Chuck Harris posted:
>> I spent a couple of lunchtimes analyzing the circuit, and what a
>> delight it is. First, it clearly meets the definition of being
>> purely digital.
>
> Yet, arguably, it also meets the definition of being analog, by more
> than just "digital is a subset of analog". The techniques used by that
> circuit would be very familiar to anyone (sadly, now mostly old farts)
> who had worked with analog computers.
>
> I deliberately use the word "arguably". It is not at all unreasonable
> to claim (as you do) that it is "purely digital". But neither is it
> unreasonable to claim that it is analog (although obviously not "purely
> analog").
>
> It's a hybrid; and is very, very neat!
That is the truth. For those who prefer to think of it in an analog way,
the very first thing they do is run the A and B inputs into a 2 bit DAC
to get a trinary count of the number of 1's on the A and B inputs.
The power supply is a mess of voltages created by tapping a string of
11 diodes in various places. The different voltages are used to allow
simple one, or two transistor comparators to fire at various points.
> Another cute thing is the behavior of D (Carry Insert) which changes the
> effect of CI (Carry Input). To avoid leaving you in suspense: the adder
> is also used for TxCx, XOR, and EQV!
Yes, I saw that. The D (Carry Insert) prevents the Carry from propagating,
which changes the ADD function into an XOR function. EQV would require
some help from the outside, in that the inputs and outputs of the XOR
function need inverting. I guess that is what the B167 is for.
> It's amazing what sorts of tricks people came up with in the 1960s.
> It's obvious that it was done that way for performance reasons; as you
> said it is a very fast circuit for what was available at the time.
The speed was one of the things that amazed me. After I figured out
how it worked, I got curious about why they did what they did. So, I
ginned together a circuit for an adder using DRTL, and put it, and the
B138, into LTspice. The B138 adds in the fractions of a microsecond region,
and the DRTL the several microsecond region. I am certain that if I
spent some time, and bolstered up a few of the inverters, I could reduce
the add times down to 1 microsecond, but I would then still have a fair
ways to go to reach the B138.
When every extra transistor costs the company tens, to hundreds of thousands
of dollars, you get motivated to spend a little engineering time inventing
an adder like this. I would venture that somebody spent quite a few weeks
with this circuit splayed out on a breadboard in the lab. Just talking
Fairchild into making the D668 dual 1N3606 diodes took a pile of money.
> When I bought the wreckage of what was once SAIL's KA10 (sadly, it was
> thoroughly toasted two decades ago), I kept the front panel, meter
> panel, light screens, BBN pager, all the B138 boards, and the B167 (hey,
> it was part of the adder). The rest went to a scrap metal dealer for
> gold reclamation.
It is a shame that artifacts like these circuits are fading from our memory.
Were it not for bitsavers, I would never have had a chance to crunch the
numbers on this little marvel.
As I age, I get nostalgic feelings towards devices like this. My catch
and release program has caught and released a fair number of computers
over the years. I think my 8/E will be the last. The memories of how
exciting these machines were when they were new is seriously dampened when
you actually try and use them today. When I try and compile something
with my 8/E on dectapes, I get this weird idea that I should whip together
a little IDE drive for the 8/E so I could do some serious work.... but
then I catch myself, and remember that the 8/E did its last really serious
work about 35 years ago.
> The BBN pager went to the Swedish hackers (when last I heard, it was
> part of a working Tenex system), and over the years some B138 boards
> have been given to friends as souvenirs. I still have several B138
> boards and, of course, the front panel & etc.
One wonders what the state of "Antique computing" will be in 20 years
time. Deep pockets, like Paul Allen, have saved quite a lot, but more
and more is getting lost to the scrappers every day.
-Chuck
Mark Crispin
> numbers on this little marvel.
I have original KA10 prints when I bought SAIL's KA10. Sadly, I haven't a
clue as to what happened to SAIL's PDP-6 prints; they probably went along
with the world's last working PDP-6 to the Computer Museum (and
destruction) in 1984.
However, I do have the source paper tape for the CONS diagnostic, which
may be the last extact document for how the CONS instruction worked on
SAIL's PDP-6. It is on my list someday to get that read.
> As I age, I get nostalgic feelings towards devices like this. My catch
> and release program has caught and released a fair number of computers
> over the years.
It's been mostly catch, and not release for me. Besides three 2020s (two
working), I have a fully stuffed (including EAE) PDP-8/f, along with a
bunch of peripherals including LA120, GIGI, VT100 upgraded to VT125 (not
working),...
> When I try and compile something
> with my 8/E on dectapes, I get this weird idea that I should whip together
> a little IDE drive for the 8/E so I could do some serious work.... but
> then I catch myself, and remember that the 8/E did its last really serious
> work about 35 years ago.
I had the same idea, but eventually gave up and got an sbc6120 instead.
I installed a small laptop IDE that I retired years ago as "too small",
but of course is hugely more than a PDP-8 can access. So, I have 8
4095-block disks, and innumerable backup images... ;-)
What I'd really like to do is run TSS/8 on it, but sadly the 6120 doesn't
have timesharing mode. So I'm stuck with OS/8.
Elliott Roper
<pfar...@pfarrell.com> wrote:
That's a bit unfair. At one point Dave epitomised the "fix-up-messes
without permission/being noticed" ethic that is being discussed with
such approval in this thread.
In 1972/3 RSX-11D was dying of bloat. It had no chance of ever running
in the littler PDP-11's coming out. Cutler snuck off on a sickie and
came back with 11-M. I was working for DEC CSS (special systems) in
Sydney at the time. 11-M saved our bacon. It saved mine. His 'crap'
includes some of the best PDP-11 assembler ever written. He taught me
to program properly by the method of SYSGEN taking so long I had time
to read his source. I made quite a decent living nicking his ideas and
techniques, as well a those of a few others around with him, like Stu
Wecker and John Gilbert, and the RT-11 lot, I'm getting old and
forgetting names...
DEC thrived in those days because of the unofficial networks of really
good people, many of whom were not paid to be doing the things they
did. It was a truly fabulous company to work for. There was a
'competence ethic' that gently defied institutionalised authority, and
kept it honest. Lots of what Barb says about 'training your boss' rings
really true to me. I think it was what kept DEC going, even while that
idiot Robert Palmer was going about saying "Our product is our share
price" and changing the colour of the logo. I never saw the US side of
it, but the dumbest thing they ever did was boot Ken Olsen. He might
have leaned too much toward hardware, he might have helped miss the toy
computer trend, but there was no way he would have run DEC into the
ground as quickly as the bozos that booted him did. The competence
ethic seemed to flow from Ken. It flowed out with him. I had left DEC
and was running my own business OEM-ing VMS stuff, but I can still
remember how gutted I and my old friends from CSS were. Flogging DEC
kit always felt a little dirty after they did that.
Ah, I descended into nostalgic sincerity for a second. Sorry.
OK, there is evidence that Dave Cutler went downhill at Microsoft, and
was by no means universally liked as a person, but not all his stuff is
crap. No way.
--
To de-mung my e-mail address:- fsnospam$elliott$$
PGP Fingerprint: 1A96 3CF7 637F 896B C810 E199 7E5C A9E4 8E59 E248
JackRubin
>
> - Show quoted text -
Chuck and Mark -
Thanks so much for taking a deep breath and returning to the subject
(more or less) at hand rather than flaming - we who lurk in awe have
learned a tremendous amount from your discussion, both technically and
historically.
best,
Jack
jmfbahciv
Sure. But these businesses were far away from each other w.r.t.
local watering places. We would find a bar that didn't have
our competition in it so we could talk shop.
/BAH
jmfbahciv
He is a fact of today's computing life. To get around
the bugs and/or non-features requires an understanding
of his OS philosophy and thinking styles. Don't you
do that with other software that isn't yours?
/BAH
jmfbahciv
Those are familiar names :-) Cutler was a fantastic bit coder. His
personality quirks caused problems which had long-range side effects.
AFAIK, I'm the only one who successfully said no to him and made it
stick.
Every extremely talented person needs somebody who can say no to them.
There were a couple of people whose personality was so strong most
people feared them and couldn't disagree, let alone say no.
>
> DEC thrived in those days because of the unofficial networks of really
> good people, many of whom were not paid to be doing the things they
> did. It was a truly fabulous company to work for. There was a
> 'competence ethic' that gently defied institutionalised authority, and
> kept it honest. Lots of what Barb says about 'training your boss' rings
> really true to me. I think it was what kept DEC going, even while that
> idiot Robert Palmer was going about saying "Our product is our share
> price" and changing the colour of the logo. I never saw the US side of
> it, but the dumbest thing they ever did was boot Ken Olsen.
You are assuming that the BoD meant to stay in business; the only
hypothesis that fits the facts is that they did not intend to.
> He might
> have leaned too much toward hardware, he might have helped miss the toy
> computer trend, but there was no way he would have run DEC into the
> ground as quickly as the bozos that booted him did. The competence
> ethic seemed to flow from Ken. It flowed out with him. I had left DEC
> and was running my own business OEM-ing VMS stuff, but I can still
> remember how gutted I and my old friends from CSS were. Flogging DEC
> kit always felt a little dirty after they did that.
> Ah, I descended into nostalgic sincerity for a second. Sorry.
It's a good story.
>
> OK, there is evidence that Dave Cutler went downhill at Microsoft,
I don't think he went downhill; most of his stuff was overruled
and the monitor resorted back to holey-ness.
> and
> was by no means universally liked as a person, but not all his stuff is
> crap. No way.
His code isn't crap. He had [emoticon grasps for a good word but
fails] objectivity problems w.r.t. thinking about a computer system
which includes soft/hardware and the customers' needs and wants.
/BAH
Elliott Roper
<jmfbahciv@aol> wrote:
> Elliott Roper wrote:
<snip>
> > DEC thrived in those days because of the unofficial networks of really
> > good people, many of whom were not paid to be doing the things they
> > did. It was a truly fabulous company to work for. There was a
> > 'competence ethic' that gently defied institutionalised authority, and
> > kept it honest. Lots of what Barb says about 'training your boss' rings
> > really true to me. I think it was what kept DEC going, even while that
> > idiot Robert Palmer was going about saying "Our product is our share
> > price" and changing the colour of the logo. I never saw the US side of
> > it, but the dumbest thing they ever did was boot Ken Olsen.
>
> You are assuming that the BoD meant to stay in business; the only
> hypothesis that fits the facts is that they did not intend to.
Isn't that called corruption?
<snip>
> > OK, there is evidence that Dave Cutler went downhill at Microsoft,
>
> I don't think he went downhill; most of his stuff was overruled
> and the monitor resorted back to holey-ness.
You are absolutely right. I could have put that much better than I did.
I guess that means you are no longer the only person ever to say no to
David N Cutler? ;-)
> > and
> > was by no means universally liked as a person, but not all his stuff is
> > crap. No way.
>
> His code isn't crap. He had [emoticon grasps for a good word but
> fails] objectivity problems w.r.t. thinking about a computer system
> which includes soft/hardware and the customers' needs and wants.
Heh! That sounds like me at my arrogant best! Half my customers'
nicknames were spelt the same as salad ingredients, and customer care
instructions included regular watering.
Pat Farrell
>> Don't mention that name. His crap still curses us
>
> He is a fact of today's computing life. To get around
> the bugs and/or non-features requires an understanding
> of his OS philosophy and thinking styles. Don't you
> do that with other software that isn't yours?
Just as his myopic decisions cursed DEC. While RSX was probably OK, he
drove VMS to be RSX on steroids rather than a decent OS, and he got
promoted, moved west and drove NT to be RSX on steroids again.
Why or why couldn't someone as smart as he, from LCG, have had that
path. We would not be cursed today with the demon spawn OS that most of
the world curses daily.
Pat Farrell
>>> Cutler.
>> Don't mention that name. His crap still curses us
>
> That's a bit unfair. At one point Dave epitomised the "fix-up-messes
> without permission/being noticed" ethic that is being discussed with
> such approval in this thread.
His success with RSX-11M was commendable. He brought the spirit of 11M
to VMS, not so good. He kept the good stuff of Tenex and Tops-20 out of
VMS, to the death of DEC. He brought the spirit of VMS to NT, very bad.
Since NT begat W2000, and XP and Vista, the roots of many of the truely
evil design decisions that we are cursed with today go straight to his
shoulders.
VMS was never what it could have been, because of his NIH approach. NT,
the same
Pat Farrell
> Sure. But these businesses were far away from each other w.r.t.
> local watering places. We would find a bar that didn't have
> our competition in it so we could talk shop.
I only went in and out of the area for a decade or so, but there
appeared to be very little cross fertilization, and very little cross
hiring. In other areas, Silicon Valley, DC during the Internet boom,
folks would have beers, make friendships, and get hired all the time.
That didn't appear to the the culture in 128/495
Elliott Roper
<pfar...@pfarrell.com> wrote:
> Elliott Roper wrote:
> >>> Cutler.
> >> Don't mention that name. His crap still curses us
> >
> > That's a bit unfair. At one point Dave epitomised the "fix-up-messes
> > without permission/being noticed" ethic that is being discussed with
> > such approval in this thread.
>
>
> His success with RSX-11M was commendable. He brought the spirit of 11M
> to VMS, not so good.
The very low-level I/O of early VMS was very M-ish and pure Cutler.
Their paths diverged after that. I can't believe Cutler would have done
kernel ASTs. VMS got kernel thread scheduling long after he was at
Microsoft. FIles-11 made the jump, but that is not Cutler's work. And
there was not a lot wrong with FIles-11 in its day.
> He kept the good stuff of Tenex and Tops-20 out of
> VMS, to the death of DEC.
Hmm, lots of the process scheduling in VMS looked like it came from
Tops-10. It was miles ahead of the simple stuff in M. I never saw much
of Tenex or Tops-20 in VMS or anywhere. So I guess you might be able to
blame him for that. The port and class driver design was not very RSX
either. I think that came after Cutler was off VMS, yet you can see
that in NT. The HAL takes that to a neat and logical place. Is that an
example of NIH?
> He brought the spirit of VMS to NT, very bad.
The result was not that hot, yet there were and are a lot of VMS
refugees at Redmond. How can we tell if it would have been even worse
without 'em?
> Since NT begat W2000, and XP and Vista, the roots of many of the truely
> evil design decisions that we are cursed with today go straight to his
> shoulders.
Some of the worst is there because he was over-ruled. Like the ones
that brought us Shatter (interprocess comms through a hidden window)
and Blaster (nicking that horrible OSF unix RPC mechanism) And of
course the celebrated shoving the general graphics driver subsystem
across the whole user/kernel divide on the transition from 3.5.1 to
NT4.
Scary as it seems, it could be his shoulders were not strong enough to
hold back MS from doing short-sighted stuff.
> VMS was never what it could have been, because of his NIH approach. NT,
> the same.
Well he got Blaster for doing RPC the unix way. He successfully did HAL
which was port and class done right. The NT file system has a bit of
files-11, a bit of BSD. I don't see a lot of NIH
But then, can we blame him for the registry?
Off with his head!
Just for interest. What happened to all the magic in Tenex and Tops-20?
Has it all faded away?
Chuck Harris
> On Mon, 30 Jun 2008, Chuck Harris posted:
>> Were it not for bitsavers, I would never have had a chance to crunch the
>> numbers on this little marvel.
>
> I have original KA10 prints when I bought SAIL's KA10. Sadly, I haven't
> a clue as to what happened to SAIL's PDP-6 prints; they probably went
> along with the world's last working PDP-6 to the Computer Museum (and
> destruction) in 1984.
That certainly was one of the great tragedies of the 80's. I would feel
more badly about it, if I wasn't associated with a computer scrapper back
then, and I am sure that I was involved in far worse atrocities with government
surplus machines. On the bright side, we did sell a lot of DEC machines
to hobbyist's for slightly above the scrap rate. The problem was that nobody
wanted all of the 8's that came through as scrap. There were hundreds, if
not thousands of them. Everyone wanted the VAXen. They were users with
small budgets, not collectors.
> However, I do have the source paper tape for the CONS diagnostic, which
> may be the last extact document for how the CONS instruction worked on
> SAIL's PDP-6. It is on my list someday to get that read.
>
>> As I age, I get nostalgic feelings towards devices like this. My catch
>> and release program has caught and released a fair number of computers
>> over the years.
>
> It's been mostly catch, and not release for me. Besides three 2020s
> (two working), I have a fully stuffed (including EAE) PDP-8/f, along
> with a bunch of peripherals including LA120, GIGI, VT100 upgraded to
> VT125 (not working),...
My 8/E is similarly stuffed. It has the EAE, plus the VC8-E point plot
hardware, a bunch of parallel I/O boards, TD8/E and a TU56D... I wish
I had more core, though. It only has 16K, and the ROM. I am using an
LA180 as the console. It probably has more processing power on board
than the 8/E.
>
>> When I try and compile something
>> with my 8/E on dectapes, I get this weird idea that I should whip
>> together
>> a little IDE drive for the 8/E so I could do some serious work.... but
>> then I catch myself, and remember that the 8/E did its last really
>> serious
>> work about 35 years ago.
>
> I had the same idea, but eventually gave up and got an sbc6120 instead.
> I installed a small laptop IDE that I retired years ago as "too small",
> but of course is hugely more than a PDP-8 can access. So, I have 8
> 4095-block disks, and innumerable backup images... ;-)
I think I remember Vince Slyngstad developing an IDE interface for the 8/E.
I have tons of 128K x 8 CMOS RAM chips, I ought to stuff a couple on
a board with a battery and fill in my remaining core.
> What I'd really like to do is run TSS/8 on it, but sadly the 6120
> doesn't have timesharing mode. So I'm stuck with OS/8.
A TSS/8 system was my introduction to computers. The system was owned
by a local leasing company that leased TTY's, and modems, and any hardware
that you could afford. They liked the 8/I in their lobby to have its
lights blinking all the time, so they leased dialup access at a cheap rate to
schools. As I recall, the passwords were combinations of 3 control characters.
-Chuck
Carl Appellof
"Mark Crispin" <m...@Panda.COM> wrote in message
news:alpine.OSX.1.10.0...@pangtzu.panda.com...
> Another cute thing is the behavior of D (Carry Insert) which changes the
> effect of CI (Carry Input). To avoid leaving you in suspense: the adder
> is also used for TxCx, XOR, and EQV!
>
> It's amazing what sorts of tricks people came up with in the 1960s. It's
> obvious that it was done that way for performance reasons; as you said it
> is a very fast circuit for what was available at the time.
>
Some -10 hardware or FS guy once told me the reason for all the seemingly
redundant instructions in the -10's instruction set was the fact that it was
easier that way, and used fewer circuits. As long as you're sharing
circuits to compute different things, you might as well make the instruction
set symmetric, even if two different instructions accomplish the same thing.
There's a certain beauty to the -10 instruction set that I find lacking in
many others.
Carl
glen herrmannsfeldt
(snip)
> Yet, arguably, it also meets the definition of being analog, by more
> than just "digital is a subset of analog". The techniques used by that
> circuit would be very familiar to anyone (sadly, now mostly old farts)
> who had worked with analog computers.
There were stories some years ago about using four state I/O pins
to get two bits in/out each IC pin. It does start to confuse
the line between analog and digital.
You could build an adder with two D/A converters, an analog
adder (probably two resistors in this case) and an A/D converter.
It sounds like the KA-10 didn't go quite that far, though.
As the bits increase, I believe it gets slower than most
digital binary adders.
-- glen
Al Kossow
> I have original KA10 prints when I bought SAIL's KA10. Sadly, I haven't
> a clue as to what happened to SAIL's PDP-6 prints; they probably went
> along with the world's last working PDP-6 to the Computer Museum (and
> destruction) in 1984.
>
The Computer Museum NEVER received SAIL's PDP-6
It went to DEC after DECUS, and was put into storage there. The Museum
had already split from DEC and there would have been a formal written
transfer and written record. By 1984, The Computer Museum was a separate
institution from DEC.
I have searched the accession records, and all of the Museum newsletters
and there is no mention of the PDP-6 anywhere.
When Compaq cleared the which held DEC's collection (not everything went
to the Computer Museum) in the late 90's and donated it to the Computer
Museum History Center, SAIL's fast memory box was included in that donation
(the cabinet matches the one in the picture from DECUS), so if it was
scrapped it was NOT done by the Museum.
I have no idea why someone at DEC would have only saved the fast memory box,
or if, in fact, the system had survived to the late 90's and disappeared at
the hands of Compaq/HP.
Mark Crispin
>> He kept the good stuff of Tenex and Tops-20 out of
>> VMS, to the death of DEC.
> Hmm, lots of the process scheduling in VMS looked like it came from
> Tops-10. It was miles ahead of the simple stuff in M. I never saw much
> of Tenex or Tops-20 in VMS or anywhere.
I agree. Most TOPS-10 customers that made the transition to VMS found
that it was not a particularly difficult transition, once they got over
the hurdle of the assembly language programs being left behind.
VMS was an upgrade for TOPS-10 customers. Although early VAXen were
lower-end machines than the KL10, the 8600 changed that. VAX clusters was
a major advance (TOPS-20 CFS might have matched the functionality with
further development that never happened), certainly more attuned to
conditions in the 1980s than TOPS-10 SMP.
Mark Crispin
> I have tons of 128K x 8 CMOS RAM chips, I ought to stuff a couple on
> a board with a battery and fill in my remaining core.
People do that with their SBC6120. The nice thing is that it is
effectively a modern PDP-8 with far less electricity consumption (you can
run it from a battery.
Sadly, kits are no longer available, but with enough energy you can roll
your own:
http://www.sparetimegizmos.com/Hardware/SBC6120-2.htm
You may want to add an IOB6120 (sadly, kits also no longer available):
http://www.jkearney.com/sbc6120/iob6120.htm
John Sauter
>
> Those are familiar names :-) Cutler was a fantastic bit coder. His
> personality quirks caused problems which had long-range side effects.
> AFAIK, I'm the only one who successfully said no to him and made it
> stick.
>
I know of a second case. One of Mr. Cutler's projects was a PL/I
compiler for VAX/VMS. I discovered that the compiler was emitting code
which did not fully construct descriptors. It built enough of the
descriptor that calls to PL/I modules would always succeed, but when
calling entry points written in other languages, that depended on fields
that PL/I code did not care about, those fields were allowed to be
uninitialized, and since they were on the stack, that meant garbage.
I talked to the people around Mr. Cutler about the problem, but couldn't
get any traction. Finally, I attended a phase review meeting, and
raised the issue there, as one which should, until fixed, prevent the
project from proceeding past that milestone. Mr. Cutler was present at
the meeting, and was vehement in his disapproval of my position. He
wasn't willing to waste a microsecond or a byte of code to make PL/I
compatible with other (necessarily lesser) languages. After all, why
would anyone want to code in anything other than PL/I?
My boss, Tom Hastings, got involved after the meeting, and the result
was that code emitted by the PL/I compiler generated descriptors that
were compatible with other VAX languages.
John Sauter (John_...@systemeyescomputerstore.com)
John Sauter
Al, I have been working with Geo Wiederhold on the history of early
computers at Stanford. You can read our current effort here:
<http://infolab.stanford.edu/pub/voy/museum/computers.html>
There is a paragraph on the SAIL PDP-6; would you care to expand it?
John Sauter (John_...@systemeyescomputerstore.com)
Chuck Harris
> On Tue, 1 Jul 2008, Chuck Harris posted:
>> I have tons of 128K x 8 CMOS RAM chips, I ought to stuff a couple on
>> a board with a battery and fill in my remaining core.
>
> People do that with their SBC6120. The nice thing is that it is
> effectively a modern PDP-8 with far less electricity consumption (you
> can run it from a battery.
The battery was to give the CMOS memory core like data retention.
The truth is, if all you want to do is run the software, the SimH
emulator does a dandy job on the laptop of your choice. I would imagine
that it could be made to run on my little Palm TX... faster than realtime.
The experience is a little more authentic with the rack fans roaring,
the dectape going "skitch, skitch, skitch, ..." the LA180 going "yick,
yick, yick, yick..." and the blinken lights a blinken.
-Chuck
Mark Crispin
> Mark Crispin wrote:
>> On Tue, 1 Jul 2008, Chuck Harris posted:
>>> I have tons of 128K x 8 CMOS RAM chips, I ought to stuff a couple on
>>> a board with a battery and fill in my remaining core.
>> People do that with their SBC6120. The nice thing is that it is
>> effectively a modern PDP-8 with far less electricity consumption (you can
>> run it from a battery.
> The battery was to give the CMOS memory core like data retention.
Ah, you misunderstood my comment. I wasn't referring to the battery on
the optional CMOS RAM board; I was referring to the face that a 9V radio
battery will power a complete SBC6120.
> The truth is, if all you want to do is run the software, the SimH
> emulator does a dandy job on the laptop of your choice. I would imagine
> that it could be made to run on my little Palm TX... faster than realtime.
Yes, but you won't get lights and switches that way. Most people with an
SBC6120 got that option.
Mine is in a picture-frame bezel that can be hung on a wall or sat on a
desktop. The only wires going out the back are DC input (which I supply
with a wall-wart) and a USB cable (my SBC6120's serial port is connected
to an el cheapo serial/USB converter).
> The experience is a little more authentic with the rack fans roaring,
> the dectape going "skitch, skitch, skitch, ..." the LA180 going "yick,
> yick, yick, yick..." and the blinken lights a blinken.
I could connect it to my LA120 (removing the serial/USB converter) but
instead I have it going to a USB input on my Mac, from which I can access
it with xterm. The disk does make a satisfying noise with access,
although nothing could match the DECtape experience I admit. As for
blinking lights, I do have them...and on my TOPS-20 system too!
Pat Farrell
>> He kept the good stuff of Tenex and Tops-20 out of
>> VMS, to the death of DEC.
> Hmm, lots of the process scheduling in VMS looked like it came from
> Tops-10. It was miles ahead of the simple stuff in M. I never saw much
> of Tenex or Tops-20 in VMS or anywhere. So I guess you might be able to
> blame him for that.
Yes, that's the point. Tenex/Tops-20 was where the research world was
taking operating systems. TOPS-10 was too old to be current. Building on
it was not good.
> Scary as it seems, it could be his shoulders were not strong enough to
> hold back MS from doing short-sighted stuff.
This could be true. MS made a lot of decisions then that reflected the
world of PCs and PC lans. NT did a good job of competing with Netware of
the time.
It completely missed the stuff that Tenex and TOPS-20/AN sites had been
doing for a decade.
> Just for interest. What happened to all the magic in Tenex and Tops-20?
> Has it all faded away?
Pretty much so, sadly.
I think Working Sets occasionally gets a bit of life, and some memory
mapped IO ideas (tenex and Multics) are around still.
When Tenex was hot, and early Tops-20 work, it was where huge parts of
the OS research was focused. But by the early 80s, the KL was too slow,
so the university folks moved elsewhere.
Mark Crispin
> Yes, that's the point. Tenex/Tops-20 was where the research world was
> taking operating systems. TOPS-10 was too old to be current. Building on
> it was not good.
Agreed. However, by 1979 the focus of the research world had shifted to
UNIX; mostly because KL10s were too expensive and a VAX/780 was competant
enough.
> [NT] completely missed the stuff that Tenex and TOPS-20/AN sites had
> been doing for a decade.
To be fair, NT got some things right compared to its contemporaries; but
in general you are correct.
>> Just for interest. What happened to all the magic in Tenex and Tops-20?
>> Has it all faded away?
> Pretty much so, sadly.
Yup. Following the utter victory of the Sith, a small sect of Jedi
preserved what it could, and went into hiding awaiting a better day.
They are still around, but even the youngest are well past the zenith of
their lives, and many old masters have passed.
Slowly, painfully slowly, some of the magic is being rediscovered, at
great expense to those that went to such effort to forget it (especially
as what was once considered obvious is now patented proprietary
technology).
> I think Working Sets occasionally gets a bit of life, and some memory
> mapped IO ideas (tenex and Multics) are around still.
To different degrees, Linux and Solaris have aspects some of these, but
not to the degress that it was in Tenex/TOPS-20 and Multics.
Morten Reistad
Mark Crispin <m...@Panda.COM> wrote:
>On Tue, 1 Jul 2008, Elliott Roper posted:
>>> He kept the good stuff of Tenex and Tops-20 out of
>>> VMS, to the death of DEC.
>> Hmm, lots of the process scheduling in VMS looked like it came from
>> Tops-10. It was miles ahead of the simple stuff in M. I never saw much
>> of Tenex or Tops-20 in VMS or anywhere.
>
>I agree. Most TOPS-10 customers that made the transition to VMS found
>that it was not a particularly difficult transition, once they got over
>the hurdle of the assembly language programs being left behind.
PPOE, and several others, took the transition to Primos instead.
Three reasons: Mature, thought out OS (mostly a cheap rip-off from
Multics), Ringnet, and an upgrade path; even if you had to do MP.
DEC could have upped this one with ease, if they ever saw the
need. We never heard a pip. That is why I ask if there was ever
cross-company communication.
>VMS was an upgrade for TOPS-10 customers. Although early VAXen were
>lower-end machines than the KL10, the 8600 changed that. VAX clusters was
>a major advance (TOPS-20 CFS might have matched the functionality with
>further development that never happened), certainly more attuned to
>conditions in the 1980s than TOPS-10 SMP.
When was the 8600 launced ? The teflon brain strikes again.
-- mrr
Morten Reistad
Pat Farrell <pfar...@pfarrell.com> wrote:
>Elliott Roper wrote:
>>> He kept the good stuff of Tenex and Tops-20 out of
>>> VMS, to the death of DEC.
>It completely missed the stuff that Tenex and TOPS-20/AN sites had been
>doing for a decade.
>
>
>> Just for interest. What happened to all the magic in Tenex and Tops-20?
>> Has it all faded away?
>
>Pretty much so, sadly.
Linux has picked up a lot of stuff from Tops20/Tenex. Internals are
closer to tops20 than to a classic unix.
>I think Working Sets occasionally gets a bit of life, and some memory
>mapped IO ideas (tenex and Multics) are around still.
The pager lies below everything in Linux. Everything except a tiny bit
of the kernel runs on to of it. The "on behalf of the user" modes are
more or less copied directly.
Linux swapping is just implemented as "good enough to satisfy posix".
Everything relies on paging.
Memory mapping has had one, huge facelift though, with the support
of EPFs and dynamic linking. This means that Linux utilises segments
intensively in the internals; but present a flat "projection" to
the user programs. Below this is a plain, paging-based memory mapper.
Bash has had extensions for looking like the comnd% parser. With a
good description file, it feels just right.
IPC is glued to the old, SYSV ipc.
>When Tenex was hot, and early Tops-20 work, it was where huge parts of
>the OS research was focused. But by the early 80s, the KL was too slow,
>so the university folks moved elsewhere.
They have been doing Linux for a decade now. The effort gone into
2.6 is really stellar.
-- mrr
Elliott Roper
Crispin <m...@Panda.COM> wrote:
Mark and Pat, That turned into a very interesting thread.
I went looking for twenex internals stuff on the web without much
success. But I did find the teco and ddt song!
Sigh!
I guess I'll have to crank up a copy of SimH and do it for real.
Simon Clubley
>
> His success with RSX-11M was commendable. He brought the spirit of 11M
> to VMS, not so good. He kept the good stuff of Tenex and Tops-20 out of
> VMS, to the death of DEC. He brought the spirit of VMS to NT, very bad.
> Since NT begat W2000, and XP and Vista, the roots of many of the truely
> evil design decisions that we are cursed with today go straight to his
> shoulders.
>
> VMS was never what it could have been, because of his NIH approach. NT,
> the same
>
[I have a PDP-11/VAX/Alpha background, but have never used the 36-bit
operating systems, so I'm curious about the lessons that could have been
learnt from those operating systems if Cutler had bothered to look at them.]
What would you have taken from the 36-bit world and placed into VMS and why ?
For the features that you identified, what is it about these features that
made them a better way of doing things than the approach that Cutler took ?
Thanks,
Simon.
--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world
jmfbahciv
> jmfbahciv wrote (excerpted):
>
>>
>> Those are familiar names :-) Cutler was a fantastic bit coder. His
>> personality quirks caused problems which had long-range side effects.
>> AFAIK, I'm the only one who successfully said no to him and made it
>> stick.
>>
>
> I know of a second case. One of Mr. Cutler's projects was a PL/I
> compiler for VAX/VMS. I discovered that the compiler was emitting code
> which did not fully construct descriptors. It built enough of the
> descriptor that calls to PL/I modules would always succeed, but when
> calling entry points written in other languages, that depended on fields
> that PL/I code did not care about, those fields were allowed to be
> uninitialized, and since they were on the stack, that meant garbage.
>
> I talked to the people around Mr. Cutler about the problem, but couldn't
> get any traction. Finally, I attended a phase review meeting, and
> raised the issue there, as one which should, until fixed, prevent the
> project from proceeding past that milestone. Mr. Cutler was present at
> the meeting, and was vehement in his disapproval of my position. He
> wasn't willing to waste a microsecond or a byte of code to make PL/I
> compatible with other (necessarily lesser) languages. After all, why
> would anyone want to code in anything other than PL/I?
That sounds like Cutler's style of thinking.
>
> My boss, Tom Hastings,
Yep. He would be able to fight that one.
got involved after the meeting, and the result
> was that code emitted by the PL/I compiler generated descriptors that
> were compatible with other VAX languages.
> John Sauter (John_...@systemeyescomputerstore.com)
Good for you.
/BAH
jmfbahciv
And the smells. Have you ever discovered that one?
/BAH
jmfbahciv
> Mark Crispin wrote:
>> On Mon, 30 Jun 2008, Chuck Harris posted:
>>> Were it not for bitsavers, I would never have had a chance to crunch the
>>> numbers on this little marvel.
>>
>> I have original KA10 prints when I bought SAIL's KA10. Sadly, I
>> haven't a clue as to what happened to SAIL's PDP-6 prints; they
>> probably went along with the world's last working PDP-6 to the
>> Computer Museum (and destruction) in 1984.
>
> That certainly was one of the great tragedies of the 80's. I would feel
> more badly about it, if I wasn't associated with a computer scrapper back
> then, and I am sure that I was involved in far worse atrocities with
> government
> surplus machines. On the bright side, we did sell a lot of DEC machines
> to hobbyist's for slightly above the scrap rate. The problem was that
> nobody
> wanted all of the 8's that came through as scrap. There were hundreds, if
> not thousands of them. Everyone wanted the VAXen. They were users with
> small budgets, not collectors.
<snip>
At one point, Digital dumped all manufactured "outdated" items. If you
find that landfill....
/BAH
jmfbahciv
> In article <pa6dnb1kSvrruvfV...@rcn.net>, jmfbahciv
> <jmfbahciv@aol> wrote:
>
>> Elliott Roper wrote:
> <snip>
>>> DEC thrived in those days because of the unofficial networks of really
>>> good people, many of whom were not paid to be doing the things they
>>> did. It was a truly fabulous company to work for. There was a
>>> 'competence ethic' that gently defied institutionalised authority, and
>>> kept it honest. Lots of what Barb says about 'training your boss' rings
>>> really true to me. I think it was what kept DEC going, even while that
>>> idiot Robert Palmer was going about saying "Our product is our share
>>> price" and changing the colour of the logo. I never saw the US side of
>>> it, but the dumbest thing they ever did was boot Ken Olsen.
>> You are assuming that the BoD meant to stay in business; the only
>> hypothesis that fits the facts is that they did not intend to.
>
> Isn't that called corruption?
>
> <snip>
>>> OK, there is evidence that Dave Cutler went downhill at Microsoft,
>> I don't think he went downhill; most of his stuff was overruled
>> and the monitor resorted back to holey-ness.
>
> You are absolutely right. I could have put that much better than I did.
> I guess that means you are no longer the only person ever to say no to
> David N Cutler? ;-)
I just learned of another instance and it required a Tom Hastings. :-)
There was an easier way to get around a disagreement of this sort.
Cutler was "gone" by then. I still don't know why that IP was handed
over; the BoD was trying to get as much cash as possible.
>>> and
>>> was by no means universally liked as a person, but not all his stuff is
>>> crap. No way.
>> His code isn't crap. He had [emoticon grasps for a good word but
>> fails] objectivity problems w.r.t. thinking about a computer system
>> which includes soft/hardware and the customers' needs and wants.
>
> Heh! That sounds like me at my arrogant best!
Most engineers couldn't grasp the customer POV. We figured out ways
to cope with that. I'd guess that 98% of hardware engineers had
zero idea w.r.t. software and how people used computers. (This was
the 70s and early 80s.)
> Half my customers'
> nicknames were spelt the same as salad ingredients, and customer care
> instructions included regular watering.
<grin> What! No fertilizer? Obviously, you had DDT to spray
the bit bugs.
/BAH
jmfbahciv
> jmfbahciv wrote:
>>> Don't mention that name. His crap still curses us
>> He is a fact of today's computing life. To get around
>> the bugs and/or non-features requires an understanding
>> of his OS philosophy and thinking styles. Don't you
>> do that with other software that isn't yours?
>
> Just as his myopic decisions cursed DEC.
No. That's not how things worked.
> While RSX was probably OK, he
> drove VMS to be RSX on steroids rather than a decent OS,
A lot of that was Husvedt(sp?). If a PDP-10 could do it,
this guy would code so that VMS would never do it. I cheered
when he cracked his head open. Cutler wasn't that bad.
> and he got
> promoted,
WTF are you talking about? He didn't get promoted in this instance.
> moved west and drove NT to be RSX on steroids again.
Whatever. You weren't there. You never worked with him nor
for him.
>
> Why or why couldn't someone as smart as he, from LCG, have had that
> path.
They were dead or dying. The others were working on VMS.
>We would not be cursed today with the demon spawn OS that most of
> the world curses daily.
There are -20 people (last I heard) working on MS' OSes. The
same thing would have happened. MS' tradeoffs are based on
distribution. But you won't understand this.
/BAH
jmfbahciv
> Elliott Roper wrote:
>>>> Cutler.
>>> Don't mention that name. His crap still curses us
>> That's a bit unfair. At one point Dave epitomised the "fix-up-messes
>> without permission/being noticed" ethic that is being discussed with
>> such approval in this thread.
>
>
> His success with RSX-11M was commendable. He brought the spirit of 11M
> to VMS, not so good. He kept the good stuff of Tenex and Tops-20 out of
> VMS, to the death of DEC.
That wasn't Cutler. That was Husvedt's doing.
>He brought the spirit of VMS to NT, very bad.
It was? No, what was bad was MS' insistence of "keeping" the DOS
mentally and inserting it into the heart of Cutler's code. That
was a battle he lost at MS.
> Since NT begat W2000, and XP and Vista, the roots of many of the truely
> evil design decisions that we are cursed with today go straight to his
> shoulders.
>
> VMS was never what it could have been, because of his NIH approach
What are you talking about. It became a very good OS and extremely
useful for users.
<snip>
/BAH
jmfbahciv
> jmfbahciv wrote:
>> Sure. But these businesses were far away from each other w.r.t.
>> local watering places. We would find a bar that didn't have
>> our competition in it so we could talk shop.
>
> I only went in and out of the area for a decade or so, but there
> appeared to be very little cross fertilization,
Anybody who went to work for competition had to sign paperwork
that prevented carrying work from a former job to the new job.
>and very little cross
> hiring. In other areas, Silicon Valley, DC during the Internet boom,
> folks would have beers, make friendships, and get hired all the time.
Yea, and all they did was play the Roman Circus games instead of
concentrating on getting real work done.
>
> That didn't appear to the the culture in 128/495
We were serious about our business. We had competition. That
is good for maintaining the souls of machines.
/BAH
>