A Dark Day...
Glen Herrmannsfeldt
"Mark Crispin" <M...@CAC.Washington.EDU> wrote in message
news:Pine.WNT.4.60.03...@Tomobiki-Cho.CAC.Washington.EDU...
> On Tue, 20 May 2003, nospam wrote:
(snip)
Someone wrote:
> > > OSF/1 (a.k.a. "Tru64") is another technological dead end, created by
those
> > I think I better get it straight what your meaning buy "dead end" ?
>
> As in "has no future", "doomed to be abandoned by its vendor leaving its
> users stranded", "will have no successor technology" dead end.
I recently bought a VAX architecture book ($3, used). There is a quote in
the beginning, something like the long life they expected for the
architecture.
When IBM designed S/360, the first "architecture", they were hoping for a
long life, but I don't think anyone believed it would still be going after
40 years.
-- glen
Eric Smith
> When IBM designed S/360, the first "architecture", they were hoping for a
> long life, but I don't think anyone believed it would still be going after
> 40 years.
We don't yet know with 100% certainty whether the System/360
architecture will last 40 years, since the first machine (a 360/40)
shipped about 38 years ago.
Dean Kent
news:qhvfw5p...@ruckus.brouhaha.com...
We know, with 100% certainty, that the architecture will last 40 years.
Something less than 23 months is fairly easy to predict in the market that
it is used in.
What we don't know is how many more years it will last after that. However,
there were approximately 9,000 IBM s/390 licenses in 1998, and approximately
11,000 today, from what I have been told :-).
Regards,
Dean
Dean Kent
"Dean Kent" <dk...@realworldtech.com> wrote in message
news:R%Aya.985$PY6.16...@newssvr21.news.prodigy.com...
>
> We know, with 100% certainty, that the architecture will last 40 years.
> Something less than 23 months is fairly easy to predict in the market that
> it is used in.
Oops, I let myself get confused with the 38 years old statement. OS/360
was introduced on April 7, 1964. That makes it a tad over 39 years old, if
my math is correct, so I'll revise the statement to say that something less
than 11 months is fairly easy to predict... ;-).
Regards,
Dean
>
>
Anne & Lynn Wheeler
> When IBM designed S/360, the first "architecture", they were hoping
> for a long life, but I don't think anyone believed it would still be
> going after 40 years.
ibm was planning on doing something completely different in the early
'70s called future system (FS) ... which eventually got canceled.
However, there is some contention that Amdahl left to do his own
360/370 plug compatible because it seemed like IBM was going to walk
away from the 360. In Amdahl's talk at MIT auditorium in the early
'70s ... sort of giving his business case justification ... was that
there was a $100b in software already written for the 360, and even if
IBM totally walked away from 360, the legacy software (at that time)
should provide him with viable market until the end of the century
(nearly 30 years). random fs refs:
http://www.garlic.com/~lynn/subtopic.html#futuresys
--
Anne & Lynn Wheeler | ly...@garlic.com - http://www.garlic.com/~lynn/
Internet trivia, 20th anniv: http://www.garlic.com/~lynn/rfcietff.htm
Eric Smith
> Oops, I let myself get confused with the 38 years old statement. OS/360
> was introduced on April 7, 1964.
Announced, not available.
It certainly would have been a neat trick for a customer to run OS/360 on
April 7, 1964, since the customer couldn't possibly have had a System/360
processor to run it on any sooner than spring of 1965.
Dean Kent
OK - oops again. :-). Still - there is absolutely no doubt it will be in
use, and supported, thru mid-2005.
Regards,
Dean
>
Mark Crispin
> We don't yet know with 100% certainty whether the System/360
> architecture will last 40 years, since the first machine (a 360/40)
> shipped about 38 years ago.
Are you sure? I'm pretty sure that S/360 was introduced in '63 or '64,
perhaps not with OS/360, but perhaps with BOS, TOS, or DOS.
-- Mark --
http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Paul Rubin
> OK - oops again. :-). Still - there is absolutely no doubt it will be in
> use, and supported, thru mid-2005.
There's always doubt.
Dean Kent
news:7xsmr93...@ruckus.brouhaha.com...
But is there a reasonable doubt? ;-).
Regards,
Dean
Mark Crispin
> Are you sure? I'm pretty sure that S/360 was introduced in '63 or '64,
> perhaps not with OS/360, but perhaps with BOS, TOS, or DOS.
Note too that most people date the PDP-10 architecture from 1964, although
the first PDP-6 was probably not shipped that year (and all the PDP-6
manuals I have are dated 1965).
Glen Herrmannsfeldt
When was the first machine running inside IBM. It doesn't have to ship to
exist.
One could say that the architecture existed as soon as its description was
written.
Also, I believe there were emulators running on 7090's, which should count,
too.
In any case S/360 (hardware) existed long before OS/360 (software) running
other OSs.
-- glen
David Kanter
>
> Cancelling Jupiter was, by itself, an act of suicide by Digital.
What were Jupiter and Venus? I tried looking them up on Google, but
nothing came up that was meaningful...
> You seem not to have grasped how much that infuriated many many powerful
> entities. There were people who banned Digital from their corporate
> procurement list as a result. Even among people who did not go that far,
> you did not see them go to VMS -- they went to UNIX.
This sounds like an interesting story...where can I find the full
version?
> Yes, Jupiter was in a great deal of trouble (as was Venus); and perhaps
> Venus was more important to save. Nevertheless, it was inexcusable that
> Digital failed to put in the resources to save Jupiter as well.
>
> Don't confuse short-term results with long-term results.
>
David Kanter
Rupert Pigott
Aha ! But internally IBM did have emulators before actual
shipping ... Come to think of it did any ISVs or customers
ever get to use those emulators before the actual hardware
shipped ?
Cheers,
Rupert
Heinz W. Wiggeshoff
Think OS/2 Warp.
Gene Wirchenko
[snip]
>When was the first machine running inside IBM. It doesn't have to ship to
>exist.
>
>One could say that the architecture existed as soon as its description was
>written.
>
>Also, I believe there were emulators running on 7090's, which should count,
>too.
Emulated, huh? That sure beats having to say RSN, eh?
>In any case S/360 (hardware) existed long before OS/360 (software) running
>other OSs.
Sincerely,
Gene Wirchenko
Computerese Irregular Verb Conjugation:
I have preferences.
You have biases.
He/She has prejudices.
bob smith
>>>I just stated why cancelling Jupiter was a good thing.
>>
>>Cancelling Jupiter was, by itself, an act of suicide by Digital.
>
>
> What were Jupiter and Venus? I tried looking them up on Google, but
> nothing came up that was meaningful...
the PDP10 instruction set using 100K ecl components and a series of
microcoded machines. It was approved to support 30 bit addressing (CE
or if you prefer CGB would not allow 32 bit addressing).
Venus was the next gen high end vax. I don't recall which number it
turned into 8xxx or soemthing. I know Al helenius and Trygve Fossum
wore working on it.
At the time, the differences between a variable lenght instruction and a
fixed length instruction allowed us to project a much faster machine on
the 10 side than the vax side.
//bob
Nick Spalding
The great day when IBM fell off the wagon. I was on a course at
Poughkeepsie at the time and IBM imported a train-load of the press to the
cafeteria in the school building and actually served booze to the
multitude. We didn't get to share this as we were all given a day off so
that we wouldn't be contaminated.
--
Nick Spalding
Walter Bushell
Golbal nuclear war?
>
> Regards,
> Dean
--
he who lives by the pelvis shall die by the pelvis.
Walter
Peter Flass
What's nice is that I can pick up a program written (compiled, even!) 38
years ago and run it today with no changes. This is a lesson SOME
companies will never learn.
Peter Flass
> ibm was planning on doing something completely different in the early
> '70s called future system (FS) ... which eventually got canceled.
Han't lots of FS surfaced in the AS/400 etc., and some of it in the
innards of AIX?
Dean Kent
news:bafc6r$gj9$1...@freenet9.carleton.ca...
>
> "Dean Kent" <dk...@realworldtech.com> wrote in message
> news:2PCya.1004$ur7.18...@newssvr21.news.prodigy.com...
> >
>
> Think OS/2 Warp.
>
Should IBM be worried that zSeries will ship with Windows pre-installed?
(shudder!). ;-).
Even if zOS is replaced with Linux S/390 entirely - would it still not be
the same system architecture it is running on - or am I confused?
Regards,
Dean
>
Peter Flass
What's your point? Still running, and still shipping as WSEB and
eCommStation.
Del Cecchi
not on the desktop. More like inside ATMs and cash registers and stuff. As I
recall, anyway.
--
Del Cecchi
cec...@us.ibm.com
Personal Opinions Only
Mark Crispin
> Jupiter (nee Hephastus) or KC10 (c for 100K ECL) was an instantiation of
> the PDP10 instruction set using 100K ecl components and a series of
> microcoded machines. It was approved to support 30 bit addressing (CE
> or if you prefer CGB would not allow 32 bit addressing).
30-bit addressing is the maximum in the PDP-10 instruction set. You can't
have 32-bit addressing without changing the format of pointers, since you
need to fit in the global/local bit, the indirect bit, and the index bits.
Note that 30-bit addressing in the PDP-10 is equivalent to "32-bits plus
1" in a byte-addressed machine. This is because the PDP-10 addresses
36-bit words and not 8-bit bytes.
If, as I propose, you use nonets to represent Unicode in UTF-9, you have
exactly as many nonets of address in the PDP-10 address space as you do
octets on a 32-bit machine. However, UTF-9 is a much more space-efficient
means of storage than UTF-8.
Normand Bélanger
"Del Cecchi" <cec...@signa.rchland.ibm.com> a écrit dans le message de news:
bag07e$reu$1...@news.rchland.ibm.com...
I can confirm that: an ATM rebooted on me a couple of weeks ago :-O
The boot screen said OS/2.
Normand
Anne & Lynn Wheeler
> Han't lots of FS surfaced in the AS/400 etc., and some of it in the
> innards of AIX?
the lore is that some number of the people from the canceled FS
project migrated to rochester and created the s/38. as/400 is
follow-on to s/38.
FS was heavily object as part of the machine instruction level ... as
well as one level store. I'm not sure how much of AIX could be
considered FS. The as/400 maintained a fairly high-level abstraction
and was able to port from a CISC hardware to power/pc RISC hardware
with little impact to customers.
in addition to lore that amdahl left to do 360 pcm, in part
because it appeared that the company would be walking away from 360
with the FS strategy:
http://www.garlic.com/~lynn/subtopic.html#futuresys
there is also some lore that 801/risc in the early to mid 70s was a
re-action to FS ... to go as far as possible to the opposite extreme
by putting as little as possible in the hardware (as opposed to
putting as much as possible in the hardware). Higher level abstraction
could be be provided by software thru the CPr operating system and the
PL.8 programming language ... running on top of minimalist hardware.
http://www.garlic.com/~lynn/subtopic.html#801
then in some sense, as/400 running on power/pc is a convergence of the
FS extremes with high level abstraction ... and risc with minimalist
hardware.
Stephen Fuld
snip
> What we don't know is how many more years it will last after that.
However,
> there were approximately 9,000 IBM s/390 licenses in 1998, and
approximately
> 11,000 today, from what I have been told :-).
This is sourprising and is also ambiguous. You don't get a S/390 licence.
You get a machine and licence the software. But you can run several
different OSs on the S/390 (now Z series) hardware, including multiple OSs
on a single machine. And the price for the lease depends on the processing
power of the machine, among other things. So it isn't really clear what
these numbers mean.
But I am still surprised. It is not surprising that the number of Z series
mips (or equivalent) is increasing, as existing site's workload increases,
but the increasing prevelance of multi-processor Z series systems (I think
virtually all of them now) would tend to reduce the number of licences.
--
- Stephen Fuld
e-mail address disguised to prevent spam
Mel Wilson
Mark Crispin <m...@CAC.Washington.EDU> wrote:
>On Tue, 20 May 2003, Mark Crispin wrote:
>> Are you sure? I'm pretty sure that S/360 was introduced in '63 or '64,
>> perhaps not with OS/360, but perhaps with BOS, TOS, or DOS.
>
>Note too that most people date the PDP-10 architecture from 1964, although
>the first PDP-6 was probably not shipped that year (and all the PDP-6
>manuals I have are dated 1965).
A _Computers and Automation_ computer census of January,
1970 lists first installation of a PDP-6 in 10/64. -10s
first installed 12/67.
First System 360 installed was the 360/40 in 4/65. By
1970 there were 1758 360/40 installations worldwide.
Regards. Mel.
Keith R. Williams
cec...@signa.rchland.ibm.com says...
...and the s/390 (z900?) service processor, I believe.
--
Keith
Gavin Scott
> Dean Kent <dk...@realworldtech.com> wrote:
>> "Paul Rubin" <http://phr...@NOSPAM.invalid> wrote in message
>> news:7xsmr93...@ruckus.brouhaha.com...
>> > "Dean Kent" <dk...@realworldtech.com> writes:
>> > > OK - oops again. :-). Still - there is absolutely no doubt it will
>> > > be in use, and supported, thru mid-2005.
>> >
>> > There's always doubt.
>>
>> But is there a reasonable doubt? ;-).
> Golbal nuclear war?
I'm pretty sure that wouldn't do it :-)
G.
Toon Moene
> What's nice is that I can pick up a program written (compiled, even!) 38
> years ago and run it today with no changes. This is a lesson SOME
> companies will never learn.
We got an e-mail of gratitude from an old school 360 user who had
recompiled his Fortran program with g77 on the S/390 running Linux and
noticed that it generated better code than the good ol' VS Fortran compiler.
"When in doubt, recompile".
--
Toon Moene - mailto:to...@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc-g95.sourceforge.net/ (under construction)
John Ahlstrom
>
>
> ibm was planning on doing something completely different in the early
> '70s called future system (FS) ... which eventually got canceled.
> 360/370 plug compatible because it seemed like IBM was going to walk
> away from the 360. In Amdahl's talk at MIT auditorium in the early
> '70s ... sort of giving his business case justification ... was that
> there was a $100b in software already written for the 360, and even if
> IBM totally walked away from 360, the legacy software (at that time)
> should provide him with viable market until the end of the century
> (nearly 30 years). random fs refs:
> http://www.garlic.com/~lynn/subtopic.html#futuresys
>
> --
> Anne & Lynn Wheeler | ly...@garlic.com - http://www.garlic.com/~lynn/
> Internet trivia, 20th anniv: http://www.garlic.com/~lynn/rfcietff.htm
I was not present at the creation of the Amdahl corp or machines,
but my understanding is that Amdahl wanted to build a non-360,
Cray killer inside IBM, but IBM said it had to be a 360. Amdahl
left, decided that there were 100 billion little green reasons to
build a 360 and did so. This is not really contradictory to your
story, but is a little bit different.
--
Thus, science was born. The first question was,
"How do we keep things from falling down?"
Soon afterwards, we had made enough of everything.
Thus, marketing was born. The first question was,
"How do we make things fall down so people have to buy more?"
John Ahlstrom
Dean Kent wrote:
>
> "Dean Kent" <dk...@realworldtech.com> wrote in message
> news:R%Aya.985$PY6.16...@newssvr21.news.prodigy.com...
> >
> > We know, with 100% certainty, that the architecture will last 40 years.
> > Something less than 23 months is fairly easy to predict in the market that
> > it is used in.
>
> Oops, I let myself get confused with the 38 years old statement. OS/360
> was introduced on April 7, 1964. That makes it a tad over 39 years old, if
> my math is correct, so I'll revise the statement to say that something less
> than 11 months is fairly easy to predict... ;-).
>
> Regards,
> Dean
>
Quibble, quibble, quibble - The discussion is about
the architecture S/360, not the OS OS/360. Sources
indicate both were announced Apr 7, 1964, but hardware
was shipped much earlier than OS/360 I believe - unless
PCP was considered OS/260.
Mark Crispin
> Quibble, quibble, quibble - The discussion is about
> the architecture S/360, not the OS OS/360. Sources
> indicate both were announced Apr 7, 1964, but hardware
> was shipped much earlier than OS/360 I believe - unless
> PCP was considered OS/260.
Did IBM have a working S/360 processor at that time? When was the first
working S/360 processor built?
Mike Cowlishaw
> On Wed, 21 May 2003, John Ahlstrom wrote:
>> Quibble, quibble, quibble - The discussion is about
>> the architecture S/360, not the OS OS/360. Sources
>> indicate both were announced Apr 7, 1964, but hardware
>> was shipped much earlier than OS/360 I believe - unless
>> PCP was considered OS/260.
>
> Did IBM have a working S/360 processor at that time? When was the
> first working S/360 processor built?
Well, all the early 360s were in fact interpreters of the architecture (I
believe) -- just like the initial Java VM was an interpreter of the Java
architecture. The first true hardware 360 was quite a bit later
(Model 91? Model 95?).
--
Mike Cowlishaw -- IBM UK
http://www2.hursley.ibm.com/mfcsumm.htm
Heinz W. Wiggeshoff
"Mike Cowlishaw" <m...@uk.ibm.com> wrote in message
news:baghct$ors$1...@news.btv.ibm.com...
> Mark Crispin wrote:
> > Did IBM have a working S/360 processor at that time? When was the
> > first working S/360 processor built?
>
> Well, all the early 360s were in fact interpreters of the
architecture (I
> believe) -- just like the initial Java VM was an interpreter of the
Java
> architecture. The first true hardware 360 was quite a bit later
> (Model 91? Model 95?).
The 360 model 75 was "hardwired". I don't recall how
the optional Array Processor worked. That would have
been a smokin' APL machine, though. Probably not
too useful for REXX.
B-)
Anne & Lynn Wheeler
John Ahlstrom <jahl...@cisco.com> writes:
> I was not present at the creation of the Amdahl corp or machines,
> but my understanding is that Amdahl wanted to build a non-360,
> Cray killer inside IBM, but IBM said it had to be a 360. Amdahl
> left, decided that there were 100 billion little green reasons to
> build a 360 and did so. This is not really contradictory to your
> story, but is a little bit different.
there was acs-1 .. and then acs360 for gene:
http://www.cs.clemson.edu/~mark/acs.html
Eric Smith
>>> When IBM designed S/360, the first "architecture", they were hoping for
>>> a long life, but I don't think anyone believed it would still be going
>>> after 40 years.
I wrote:
>> We don't yet know with 100% certainty whether the System/360
>> architecture will last 40 years, since the first machine (a 360/40)
>> shipped about 38 years ago.
Glen wrote:
> When was the first machine running inside IBM. It doesn't have to ship to
> exist.
>
> One could say that the architecture existed as soon as its description was
> written.
IMNSHO, that's a pointless way to look at it. In that sense, the IBM 701
architecture still exists, and there is every possiblity that both it
and the S/360 architecture will last hundreds if not thousands of years.
If you consider the end of the architecture to be when customers stop
running it, then the beginning is clearly when customers start running
it.
> In any case S/360 (hardware) existed long before OS/360 (software) running
> other OSs.
So? It's still the case that customers didn't get any S/360 hardware
until April 1965, and they didn't get OS/360 until even later.
Eric Smith
> What's nice is that I can pick up a program written (compiled, even!) 38
> years ago and run it today with no changes. This is a lesson SOME
> companies will never learn.
It's not necessarily true that you can run 360 programs on current
hardware. A fair bit of 360 software was hard-coded for the upper byte
of the PC to contain the flags (PSW). While the 370 architecture
included a compatability mode to support that, the 370/XA and later
versions of the architecture do not.
Giles Todd
wrote in alt.folklore.computers:
> I believe OS/2 is still in use, or at least it was a couple of years ago. Just
> not on the desktop. More like inside ATMs and cash registers and stuff. As I
> recall, anyway.
It was running on an ATM in Venice a few weeks ago. I watched the
damned thing reboot after serving the customers in front of me. After
about five minutes waiting for it to get to a point where it would
accept my card with nothing but lines and lines of description of what
hardware it had found, I decided to look for another ATM.
Damn. I wish I had taken a photo now.
Giles.
Antonio Carlini
> Jupiter (nee Hephastus) or KC10 (c for 100K ECL) was an instantiation of
> the PDP10 instruction set using 100K ecl components and a series of
> microcoded machines. It was approved to support 30 bit addressing (CE
> or if you prefer CGB would not allow 32 bit addressing).
I gues that CGB is Gordon Bell, but who was CE?
> Venus was the next gen high end vax. I don't recall which number it
> turned into 8xxx or soemthing. I know Al helenius and Trygve Fossum
> wore working on it.
Venus was the VAX 8600. I believe they had to downgrade the
speed to get it out at all (and it was pretty late
by then). When they eventually managed to up the speed,
the result came out as the VAX 8650 (aka Morningstar).
There was a rumour of a VAX 8670, but I have no idea
if it was ever more than a rumour.
Antonio
--
---------------
Antonio Carlini arca...@iee.org
bob smith
Antonio Carlini wrote:
> bob smith wrote:
>
>> Jupiter (nee Hephastus) or KC10 (c for 100K ECL) was an instantiation
>> of the PDP10 instruction set using 100K ecl components and a series of
>> microcoded machines. It was approved to support 30 bit addressing (CE
>> or if you prefer CGB would not allow 32 bit addressing).
>
>
> I gues that CGB is Gordon Bell, but who was CE?
CGB, is indeed Chester Gordon Bell. CE is Central Engineering, and the
board under bell.
>
>> Venus was the next gen high end vax. I don't recall which number it
>> turned into 8xxx or soemthing. I know Al helenius and Trygve Fossum
>> wore working on it.
>
>
> Venus was the VAX 8600. I believe they had to downgrade the
> speed to get it out at all (and it was pretty late
> by then). When they eventually managed to up the speed,
> the result came out as the VAX 8650 (aka Morningstar).
> There was a rumour of a VAX 8670, but I have no idea
> if it was ever more than a rumour.
>
> Antonio
Thanks, I was not sure of the numbering, I just recall it was late, had
a HUGE staff compared to 2080, and had a target manufacturing price for
the CPU of 50K.
bob
Peter da Silva
Peter Flass <peter...@yahoo.com> wrote:
>What's nice is that I can pick up a program written (compiled, even!) 38
>years ago and run it today with no changes. This is a lesson SOME
>companies will never learn.
Exec/8?
--
Rev. Peter da Silva, ULC. 29.6852N 95.5770W WWFD?
"Be conservative in what you generate, and liberal in what you accept"
-- Matthew 10:16 (l.trans)
Anne & Lynn Wheeler
oh, and some slight drift
Computing's dying breed
By Tom Foremski
Published: May 20 2003 18:30 | Last Updated: May 20 2003 18:30
Once, pundits predicted that the mainframe computer was a dinosaur
heading for extinction. These huge classics of the corporate computing
world were reaching the end of the line and would give way to the new
generation of server-based systems.
..snip..
But IT workers with mainframe experience are getting older. A study by
the Meta Group last year found that 55 per cent were over 50, compared
with fewer than 10 per cent of those with Unix or Windows NT server
skills.
..snip..
Mark Crispin
> >> Jupiter (nee Hephastus) or KC10 (c for 100K ECL) was an instantiation
> >> of the PDP10 instruction set using 100K ecl components and a series of
> >> microcoded machines. It was approved to support 30 bit addressing (CE
> >> or if you prefer CGB would not allow 32 bit addressing).
> > I gues that CGB is Gordon Bell, but who was CE?
> CGB, is indeed Chester Gordon Bell. CE is Central Engineering, and the
> board under bell.
Neither of them had anything to do with 30-bit vs. 32-bit addressing. The
PDP-10 pointer word format limits it to 30 bits (but that's still more
memory than a 32-bit byte-oriented machine).
On the other hand, I would believe a limit of 27-bit addressing. It is
relatively straightforward to extend from 23-bit addressing (the limit in
the KL) to 27 bits. Instead of a 5-bit section table, you have a full
page (9-bit) section table, and page numbers become 18 bits instead of 14
bits. There's a fair amount of surgery needed to the virtual memory
stuff, but it's relatively straightforward.
Getting those last three bits, though, are a nightmare. To start with,
only 18 bits are available for the page number in a storage address; bits
12-17 hold the storage medium. So that breaks immediate page pointers and
indirect page pointers. It also breaks the page fail word which only has
27 bits.
I understand that XKL did this; they would have had to redefine how the
pager works and then change lots of stuff in the TOPS-20 kernel.
Rich Alderson
> Antonio Carlini wrote:
>> Venus was the VAX 8600. I believe they had to downgrade the
>> speed to get it out at all (and it was pretty late
>> by then). When they eventually managed to up the speed,
>> the result came out as the VAX 8650 (aka Morningstar).
>> There was a rumour of a VAX 8670, but I have no idea
>> if it was ever more than a rumour.
> Thanks, I was not sure of the numbering, I just recall it was late, had
> a HUGE staff compared to 2080, and had a target manufacturing price for
> the CPU of 50K.
Shortly before the cancellation, the 2080 had been renumbered as the 4050;
I have a marketing materials binder with the DECSYSTEM-4050 name right there
on the front for all to see.
Grump. I never thought to scan it till just now; have to wait till I can get
it out of the storage locker.
--
Rich Alderson ne...@alderson.users.panix.com
"You get what anybody gets. You get a lifetime." --Death, of the Endless
bob smith
Bingo, but when that same sort of arguement was presented to the
committee of consulting engineers, Bell's staff, the answer was no.
Yes, would have broken Tops 20. Yes, would have been a departure from
PDP10 architecture in a subtle but majore way. Yes, it would have been a
discriminator. Yes, we did use that ploy to get 30 bits approved. We
did spend a lot of time looking at how we could support the extended
addressing in fortran.
I hope some of this discussion spurs some ofthe lurkers to pipe in here.
It was a less than pleasent experience having to stop work and prepare a
briefing or presenation for a new set of CE crew while trying to keep a
deadline. It was not pleasent rejustifying over and over, the decisions
made. It was fun when Ted Hess wore a suit and I wore jeans and a tee
shirt. That kind of threw them for a minute.
hack
Not true: 360-style 24-bit mode is still supported, even under the new
64-bit z/Architecture. 370/XA added 31-bit mode, and z/Arch added 64-bit
mode, so now there are three modes user-level programs can run in, and
mode switching is via user-level ("problem-state") instruction. In 24-bit
mode all user-level instructions that were present on S/360 still work
exactly the same way. Supervisor-state (OS-level) compatibility is another
story -- but stability is still pretty good: one major change every ten
years, roughly.
The mode-switching instructions were designed to permit user-level glue
code to mediate new programs calling old programs, and vice versa, as
long as certain rules are observed (e.g. parameters must be "below the
line" as they say).
Compatibility means old programs still work, but they can't necessarily
exploit new facilities (e.g. larger address space). The point is, they
work *at least* as well as they ever did (probably faster by a few orders
of magnitude), and some programs (those written with a bit of care and
forethought, e.g. using only clean addresses) could even exploit the
larger address space available with 370/XA. 64-bit exploitation is
another story, however, because it takes a new set of instructions to
use all 64 bits of the bigger registers (unlike PowerPC, where it is
somewhat easier to exploit 64-bit addressing with old programs).
Compatibility even extends to OS interfaces. There is a basic set of
services, dating back to OS level 21 (early 70s I believe), that many
compilers and run-time environments kept using until the 1990s, and
that are still available in "OS simulation" environments in CMS and
other non-OS/360-derived operating systems (like mine). This means
that compilers from the early 70s, and the code they generate, can
still be used 30 years later on current machinery. I don't know what
the comparable status is in z/OS today.
Michel.
Eric Smith
> What's nice is that I can pick up a program written (compiled, even!) 38
> years ago and run it today with no changes. This is a lesson SOME
> companies will never learn.
I wrote:
> It's not necessarily true that you can run 360 programs on current
> hardware. A fair bit of 360 software was hard-coded for the upper byte
> of the PC to contain the flags (PSW). While the 370 architecture
> included a compatability mode to support that, the 370/XA and later
> versions of the architecture do not.
ha...@watson.ibm.com (hack) writes:
> Not true: 360-style 24-bit mode is still supported, even under the new
> 64-bit z/Architecture. 370/XA added 31-bit mode, and z/Arch added 64-bit
370 had a bit for System/360 compatibility mode. That feature went away in
370/XA. The only references to 360 in the z/Architecture POO are:
* p. xix: EBCDIC character code originated with System/360
* pp. xix, 2-6, 13-2, 13-3, 15-12, 17-9, 17-10, 17-17, X-40: System/360,
System/370 parallel I/O channel interfaces
* p. 1-1: description of evolution of the architecture
* p. 7-147: the use of the Test And Set instruction for multiprocessor
synchronization is provided for System/360 compatibility.
* p. 19-12: IEEE traps use data-exception, which was one of the original
15 exceptions from the System/360 architecture
There is a 24-bit address mode, entered by the SAM24 instruction, but it
does NOT change the PSW format for full System/360 compatibility. As
far as I can tell, it only causes bits 0-7 of 32-bit values to be
ignored when used as addresses. Thus it appears to me that any
System/360 software that depends on the PSW format will fail to work
correctly on the 390 or z/Architecture. I'm not sure how much problem
state software has these dependencies, but I know that at least some of
it does.
Glen Herrmannsfeldt
"Eric Smith" <eric-no-s...@brouhaha.com> wrote in message
news:qhaddfl...@ruckus.brouhaha.com...
> Peter Flass <peter...@yahoo.com> writes:
> > What's nice is that I can pick up a program written (compiled, even!) 38
> > years ago and run it today with no changes. This is a lesson SOME
> > companies will never learn.
(snip)
> There is a 24-bit address mode, entered by the SAM24 instruction, but it
> does NOT change the PSW format for full System/360 compatibility. As
> far as I can tell, it only causes bits 0-7 of 32-bit values to be
> ignored when used as addresses. Thus it appears to me that any
> System/360 software that depends on the PSW format will fail to work
> correctly on the 390 or z/Architecture. I'm not sure how much problem
> state software has these dependencies, but I know that at least some of
> it does.
The place where 360 problem state software sees the PSW is in the address
supplied by BAL and BALR. I believe that in 24 bit mode the appropriate
flag bits are still supplied in the high byte. One effect of this
compatibility is incompatability with the newer modes. BAL sets the high
bit (it is the instruction length code for four byte instructions) which, if
used in newer instructions sets 31 bit mode. Old software won't use the
newer instructions, so it won't have that problem.
-- glen
Charles Shannon Hendrix
> Computing's dying breed
> By Tom Foremski
> Published: May 20 2003 18:30 | Last Updated: May 20 2003 18:30
>
> Once, pundits predicted that the mainframe computer was a dinosaur
> heading for extinction. These huge classics of the corporate computing
> world were reaching the end of the line and would give way to the new
> generation of server-based systems.
That's basically what we were taught in school: "Don't waste your time
with it any more."
I started out on IBM mainframes, but the new school I moved to had none.
> But IT workers with mainframe experience are getting older. A study by
> the Meta Group last year found that 55 per cent were over 50, compared
> with fewer than 10 per cent of those with Unix or Windows NT server
> skills.
Which leads to the question: will the mainframe companies and shops
start trying to get younger workers into the mainframe job market?
If you are interested, how do you get into it?
There are mainframes in most any city, but it seems there are never any
openings in shops that run them.
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
jmfb...@aol.com
They may not be able to.
>
>It was a less than pleasent experience having to stop work and prepare a
>briefing or presenation for a new set of CE crew while trying to keep a
>deadline. It was not pleasent rejustifying over and over, the decisions
>made. It was fun when Ted Hess wore a suit
[emoticon's jaw drops to the floor] Wow! Did you take a picture?
> .. and I wore jeans and a tee
>shirt. That kind of threw them for a minute.
For only a minute? CE was brain dead by 1975 when we moved to Marlboro.
At the time, my conclusion was that CMU was an infection.
/BAH
Subtract a hundred and four for e-mail.
jmfb...@aol.com
Rich Alderson <ne...@alderson.users.panix.com> wrote:
>bob smith <sfm...@bellatlantic.net> writes:
>
>> Antonio Carlini wrote:
>
>>> Venus was the VAX 8600. I believe they had to downgrade the
>>> speed to get it out at all (and it was pretty late
>>> by then). When they eventually managed to up the speed,
>>> the result came out as the VAX 8650 (aka Morningstar).
>>> There was a rumour of a VAX 8670, but I have no idea
>>> if it was ever more than a rumour.
>
>> Thanks, I was not sure of the numbering, I just recall it was late, had
>> a HUGE staff compared to 2080, and had a target manufacturing price for
>> the CPU of 50K.
>
>Shortly before the cancellation, the 2080 had been renumbered as the 4050;
I don't remember that. AFAICT, there were at least five "corporations"
in charge of that machine and none of them worked for DEC even though
that's what was stamped on their paychecks.
What a fucking-idiot idea; especially when thinking about the name
of the 2020. It was a nice _small_ machine (where small was
simultaneous user count rather than footprint).
>I have a marketing materials binder with the DECSYSTEM-4050 name right
there
>on the front for all to see.
>
>Grump. I never thought to scan it till just now; have to wait till I can
get
>it out of the storage locker.
>
Good grief! And the put it in print?
Jan C. Vorbrüggen
> > years ago and run it today with no changes. This is a lesson SOME
> > companies will never learn.
>
> We got an e-mail of gratitude from an old school 360 user who had
> recompiled his Fortran program with g77 on the S/390 running Linux and
> noticed that it generated better code than the good ol' VS Fortran > compiler.
Hah! Who has source code for his applications from thirty years ago, raise
your hand....what, nobody?
Jan
Nick Spalding
I have half a dozen boxes of cards in my attic dating from between 40 and
35 years ago.
--
Nick Spalding
jmfb...@aol.com
jmfb...@aol.com
Morten Reistad
Not from 30 years ago, but almost.
I still have the source from my first programs in 1975, on a
programmable calculator. That makes it 28.
I have the decks I submitted to the course on the 1130.
I have most of what I did on Tops20 (1978-1986) on tape.
I even have most of my e-mail intact. MH even understands
tops20 mailboxes (pre-rfc732) so I can read them. I started
using e-mail regularly in 1979.
I have two holes in this; my term in the Navy 1984-85 (they used
e-mail extensively then, but we were not allowed to save any of it)
and the corporate monstrosity KPNQwest that righly went under.
The M$ mail system made it too difficult to save all of this.
I do have copies of the important bits though.
-- mrr
Heinz W. Wiggeshoff
"Jan C. Vorbrüggen" <jvorbr...@mediasec.de> wrote in message
news:3ECCDE83...@mediasec.de...
> Hah! Who has source code for his applications from thirty years ago,
raise
> your hand....what, nobody?
I hope you've had time to reflect on the, umm,
silliness of that comment in this newsgroup.
I have programs on tape or listings going back
to circa 1968.
Mark Crispin
�悲茵�怦�葹�齒�竇�竢粤�肬�蓍�瘰韭蜒癆蜿銖�胙闕��蜥��奛癇�瘍鍖�鱇蜩��尞��葹鈔���癆��鈿硼糯�
�粹逾�葹���鋹�齒�竇�竢粤�胙闕��蜥��奛癇�瘍鍖�碯��葹��齒辣�胙闕
恩�奛癇�瘍鍖�瘤�頏續��逡竏�瘡�閹�蜚��癆�蜩�俺�奛癇�闌�闥�鈬�鬯
�葹��髟蜚��聽��齒�竇�竢粤�肅跂��癆�癇�胙闕�矼肬鱚�外�奛癇�瘍鍖
褫齡�鈿�迚鈬�
⑬�浴鳬�⑬
蔗�痕��聿�癈蓍鈑�隨繖�迺�噬蜈釿�粹纉�鈿�纃纈艱�胙闕���鈑��鞜鶯�竟跚�笂��闥�頤碎蜒�粤矚��
jchausler
"Jan C. Vorbrüggen" wrote:
> Hah! Who has source code for his applications from thirty years ago, raise
> your hand....what, nobody?
Well, let's see. I've got some floppies (both 8 inch
and 5 1/4 inch) from my 6809 system from about
1980. I've got some audio cassettes and paper
tape and 5 1/4 inch floppies for my 6800 from
about 1977. I've got some paper tape from a
DG Nova from about 1973. (However, I don't
have a Nova on which to run those.)
All this is source code and I have the executables
on this media as well and all of it is still readable
and I have the equipment on which to read it, all
operational.
Now, I also have 20+ half inch magnetic tapes,
most 556 BPI from a Univac 1108, and some
DECtapes from both a PDP-9 and a PDP-10
and even a couple of 1 inch magnetic tapes from
a Bendix G-20, all with sources and executables
on them, and all from the late 60's. For these,
however, I no longer have access to equipment
which can be used to read them and so cannot
claim to "still have source code" for those.
I do have a few short source decks of punch
cards from the late 60's as well and although
they're "human readable", I don't have a
card reader and so don't claim to have those
sources either. Actually with some of that
"magnetic developer" or whatever it is, one
could say the magnetic tapes are "human
readable", but I think I'll take a pass on that...
Chris
AN GETTO$;DUMP;RUN,ALGOL,TAPE
$$
Toon Moene
> I wrote:
>>We got an e-mail of gratitude from an old school 360 user who had
>>recompiled his Fortran program with g77 on the S/390 running Linux and
>>noticed that it generated better code than the good ol' VS Fortran > compiler.
>
>
> Hah! Who has source code for his applications from thirty years ago, raise
> your hand....what, nobody?
http://www.fortran.com/ibm3.jpg
(Well, you have to put a C in column one on the line starting with
FREQUENCY and add an END statement after the STOP 77777, but otherwise
it just compiles with g77 :-)
--
Toon Moene - mailto:to...@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc-g95.sourceforge.net/ (under construction)
Dean Kent
Does 22 years win anything? ;-)
Regards,
Dean
Peter da Silva
Christopher C. Stacy
> Hah! Who has source code for his applications from thirty years ago,
> raise your hand....what, nobody?
I have some paper listings (and paper tape!) source code I wrote in
1973, but I wasn't prolific until around 1979. Much of the stuff
from 1979 forward I have online on the PC I'm writing this note from.
I missed the context of the question, but I would be surprised if
lots and lots of people didn't have source code from 30 years ago.
Michael Meissner
Only if you have a system hooked up to the network with a working card reader.
Having media without a reader is kind of useless. I have DAT-3 tapes that I
can no longer read, since my DAT-3 drive bit the dust for the 3rd time after
the warranty expired.
--
Michael Meissner
email: mrm...@the-meissners.org
http://www.the-meissners.org
Robert Bonomi
/me raises hand
I've got some hard-copy listings from 1967.
I've got mag-tapes from 1975
Peter Flass
I see lots of stuff. Heck, I've got the PL/I(F) compiler from some
many-odd years ago and it runs just fine. I re-assembled it when I
moved it to VM, but I didn't have to do anything to run it on OS/390.
There's tons of PD utilities, etc. available of varying age.
bob smith
>
>>.. and I wore jeans and a tee
>>shirt. That kind of threw them for a minute.
>
>
> For only a minute? CE was brain dead by 1975 when we moved to Marlboro.
> At the time, my conclusion was that CMU was an infection.
Carl Lowenstein
I have the first major Fortran program I wrote, the listing is bound
as an appendix of my Ph.D. thesis. Forty years old at the end of this
month. A few years ago in a fit of idleness I used the visual/manual
OCR to produce a copy of it. Worked just fine, although the compiler
objected to an extra pair of parentheses in an implied DO-loop.
carl
--
carl lowenstein marine physical lab u.c. san diego
clow...@ucsd.edu
Robert Wessel
> 370 had a bit for System/360 compatibility mode. That feature went away in
> 370/XA. The only references to 360 in the z/Architecture POO are:
>
> * p. xix: EBCDIC character code originated with System/360
>
> * pp. xix, 2-6, 13-2, 13-3, 15-12, 17-9, 17-10, 17-17, X-40: System/360,
> System/370 parallel I/O channel interfaces
>
> * p. 1-1: description of evolution of the architecture
>
> * p. 7-147: the use of the Test And Set instruction for multiprocessor
> synchronization is provided for System/360 compatibility.
>
> * p. 19-12: IEEE traps use data-exception, which was one of the original
> 15 exceptions from the System/360 architecture
>
> There is a 24-bit address mode, entered by the SAM24 instruction, but it
> does NOT change the PSW format for full System/360 compatibility. As
> far as I can tell, it only causes bits 0-7 of 32-bit values to be
> ignored when used as addresses. Thus it appears to me that any
> System/360 software that depends on the PSW format will fail to work
> correctly on the 390 or z/Architecture. I'm not sure how much problem
> state software has these dependencies, but I know that at least some of
> it does.
The PSW is not normally directly available to problem state programs,
so the various format changes are mostly invisible. A few OS services
will return a native (untranslated) PSW (usually exception handling
stuff), and you can find some PSWs in dumps, and a few in low core,
but few applications use those.
24-bit programs mostly work just fine on 64-bit zOS.
Supervisor (privileged) code is a whole different story, of course.
Jan C. Vorbrüggen
> silliness of that comment in this newsgroup.
> I have programs on tape or listings going back
> to circa 1968.
Relax, relax. Take the post with a large grain of salt and turn on
your detector for sarcasm.
While I'm surprised and delighted how many here have various media with
(possibly) some source et al. on them, the fact remains that at least in
business environments, a lot of applications only exist as (patched)
executables, likely without any specification (especially for the patches).
And not all OSes are as exemplary as VMS or VMS in allowing old code to
execute on newer versions (SunOS/Solaris, anyone?).
Jan
Jan C. Vorbrüggen
Punch cards have the advantage that, like paper, they can be read optically.
Just scan them in and use some image processing to decode them.
Anybody written a virtual card reader driver for an S/360 emulator based
on this approach?
> Having media without a reader is kind of useless.
Quite. Wasn't that NASA sending 'round a plea for help in finding a
266 bpi (or somesuch) tapedrive in order to try to read some archived tapes?
Jan
Heinz W. Wiggeshoff
> > I hope you've had time to reflect on the, umm,
> > silliness of that comment in this newsgroup.
> > I have programs on tape or listings going back
> > to circa 1968.
>
> Relax, relax. Take the post with a large grain of salt and turn on
> your detector for sarcasm.
Hmm - seems I forgot the B-) at the end of that followup.
Heinz W. Wiggeshoff
"jchausler" <jcha...@earthlink.net> wrote in message
news:3ECD18DC...@earthlink.net...
...
> Actually with some of that
> "magnetic developer" or whatever it is, one
> could say the magnetic tapes are "human
> readable", but I think I'll take a pass on that...
Ya, my eyes are going to hell too.
Nick Spalding
It's not only the reader, for the older stuff I have I would need a 1410 or
7010. There is a guy working on a 1410 emulator to whom I sent all my old
1410/7010 documentation in 1996 but I haven't heard from him for about a
year.
--
Nick Spalding
Sander Vesik
But you can also use an optical scanner ...
--
Sander
+++ Out of cheese error +++
jmfb...@aol.com
I'd have to visit CA to get the SPS cards to work.
/BAH
jmfb...@aol.com
If you think about it, this is amazing. I wonder if old C progs will
be as hearty in 20 years.
Christopher C. Stacy
Jan> While I'm surprised and delighted how many here have various media with
Jan> (possibly) some source et al. on them, the fact remains that at least in
Jan> business environments, a lot of applications only exist as (patched)
Jan> executables, likely without any specification (especially for the patches).
While that may be true for 30 year old programs, it was also often
true for those same programs when they were one or two years old!
jmfb...@aol.com
Even a week old. That was another reason why we fought tooth and
nail whenever somebody tried to strip sources out of our distribution
tapes. When I finally got control of how the distribution tapes were
made my hard and fast rule was that any binary had to be generated
just before those tapes were made. The only exceptions were the
binaries I wasn't allowed to have sources for (gag..what an awful
sentence!).
/BAH
Morten Reistad
>In article <bak42o$jij$1...@news1.ucsd.edu>,
> c...@deeptow.ucsd.edu (Carl Lowenstein) wrote:
>>In article <krnpcvcr34mop0rn2...@4ax.com>,
>>Nick Spalding <spal...@iol.ie> wrote:
>>>Jan C. Vorbrüggen wrote, in <3ECCDE83...@mediasec.de>:
>>>
>>>>>> What's nice is that I can pick up a program written (compiled, even!)38
>>>>>> years ago and run it today with no changes. This is a lesson SOME
>>>>>> companies will never learn.
>>>>> We got an e-mail of gratitude from an old school 360 user who had
>>>>> recompiled his Fortran program with g77 on the S/390 running Linux
>>>>> and noticed that it generated better code than the good ol' VS Fortran
>>>> compiler.
[snip]
>If you think about it, this is amazing. I wonder if old C progs will
>be as hearty in 20 years.
I bet they will. There is a drive towards long-term stability
descending on the whole technology industry, and at least the
process control people will keep C alive.
Then there is unix. (lower case u, meaning the whole set of systems).
That will surely survive more than 20 years. The odds of running
an intel x86 binary 20 years after the fact is also good.
But I would expect it to be reasonably hopeless to run Windows
2000 programs. MS cannot even run their own software from 6 years
back on W2k!
-- mrr
Joe Morris
I can go one step further back: I've still got the listings (no cards)
from the first class assignment from a class in 7090 assembler (FAP)
programming in 1962. The program worked correctly the first time I
submitted it (it wasn't particularly complex) -- then on the remaining
runs (total of 4 runs were possible IIRC) I tried to tweak it to make it
more elegant.
None of the runs with the "enhanced" design worked.
Joe Morris
Peter da Silva
>>But you can also use an optical scanner ...
>It would be faster to retype it in (for the FORTRAN programs).
If you have a listing and can retype it, I'd count that as "having source".
Mark Crispin
> >If you think about it, this is amazing. I wonder if old C progs will
> >be as hearty in 20 years.
> I bet they will. There is a drive towards long-term stability
> descending on the whole technology industry, and at least the
> process control people will keep C alive.
C, yes; old C programs, no.
Speaking as someone who has spent a lot of his life in the past 15 years
writing portable C code, I can assert that it is *very* difficult to write
truly portable C code (where "portable" means "will compile and run on any
systems with a C compiler").
A *lot* of code these days assumes that all the world is Linux or Solaris.
A lot of code in the past asssumed that all the world is a VAX. Almost
all C code assumes that memory is allocated and addressed in 8-bit bytes.
It is extremely rare to be able to compile a C program written for a 1980s
UNIX on Linux without modification. GNU autoconf doesn't help; it just
adds further complexity and obfuscation (speaking as someone who has
wasted many many days doing battle with that noisome pile of reptile
excrement).
It's easier to compile a program, written in assembly language for the
PDP-6 for a 2-series monitor, on TOPS-20 release 7 and have it run.
> But I would expect it to be reasonably hopeless to run Windows
> 2000 programs. MS cannot even run their own software from 6 years
> back on W2k!
I have had little trouble running all sorts of software on WinXP. Even
ancient DOS programs work. When problems arise, it's typically with
programs that used hacks that depending upon some undocumented part of the
OS instead of the official interface, and those hacks broke.
-- Mark --
http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Peter da Silva
>If you think about it, this is amazing. I wonder if old C progs will
>be as hearty in 20 years.
C is inherently less portable than Fortran because Fortran simply didn't
make a lot of hardware and OS details visible at all. Still, most 20+
year old C code is going to be happy running under the PDP-11 emulator. :)
Charlie Gibbs
(Morten Reistad) writes:
>According to <jmfb...@aol.com>:
>
>>In article <bak42o$jij$1...@news1.ucsd.edu>,
>>c...@deeptow.ucsd.edu (Carl Lowenstein) wrote:
>>
>>>In article <krnpcvcr34mop0rn2...@4ax.com>,
>>>Nick Spalding <spal...@iol.ie> wrote:
>>>
>>>>Jan C. Vorbrüggen wrote, in <3ECCDE83...@mediasec.de>:
>>>>
>>>>>> We got an e-mail of gratitude from an old school 360 user who
>>>>>> had recompiled his Fortran program with g77 on the S/390 running
>>>>>> Linux and noticed that it generated better code than the good ol'
>>>>>> VS Fortran compiler.
>[snip]
>>If you think about it, this is amazing. I wonder if old C progs will
>>be as hearty in 20 years.
>
>I bet they will. There is a drive towards long-term stability
>descending on the whole technology industry, and at least the
>process control people will keep C alive.
I re-wrote my bread-and-butter software package from BASIC into C
in 1992, and I continue upgrading it to this day. I'm compiling
versions for MS-DOS, 16-bit Windows, 32-bit Windows, SCO UNIX, and
Linux from the same source base (and I use the libraries in separate
utility programs for my Amiga). I think C programs are going to be
around for a while.
>Then there is unix. (lower case u, meaning the whole set of systems).
>That will surely survive more than 20 years. The odds of running
>an intel x86 binary 20 years after the fact is also good.
Worst case, I see no reason why the C compiler on, say, my MS-DOS
box should suddenly quit running. As long as I have a box that can
run MS-DOS, I can run the compiler I have (which might be old, but
which does the job quite well, thankyouverymuch). The same goes
for that 386 that's running SCO 3.2v4.2, which is still grinding
out programs that keep customers satisfied.
>But I would expect it to be reasonably hopeless to run Windows
>2000 programs. MS cannot even run their own software from 6 years
>back on W2k!
This, of course, is by design. On the other hand, my MS-DOS programs
still run there, and my 32-bit Windows programs run on everything from
Win95 through XP. But then, my motives are slightly different... :-)
--
/~\ cgi...@kltpzyxm.invalid (Charlie Gibbs)
\ / I'm really at ac.dekanfrus if you read it the right way.
X Top-posted messages will probably be ignored. See RFC1855.
/ \ HTML will DEFINITELY be ignored. Join the ASCII ribbon campaign!
Terje Mathisen
I mostly agree, except that a _lot_ of those hacks was needed just to
get useful speed, meaning that they eventually became part of the more
or less documented 'real' interface.
I.e. amazingly many low-level Dos programs (even TSRs!) still run.
Terje
--
- <Terje.M...@hda.hydro.com>
"almost all programming can be viewed as an exercise in caching"
bfra...@jetnet.ab.ca
>>I have the first major Fortran program I wrote, the listing is bound
>>as an appendix of my Ph.D. thesis. Forty years old at the end of this
>>month. A few years ago in a fit of idleness I used the visual/manual
>>OCR to produce a copy of it. Worked just fine, although the compiler
>>objected to an extra pair of parentheses in an implied DO-loop.
>
>
> If you think about it, this is amazing. I wonder if old C progs will
> be as hearty in 20 years.
>
What is C in 20 years? C^3 will be out by then.
Any how many of the old Fortran programs out there
still use Sense Switches?
Ben.
Charles Richmond
fixes had to be patches to the binaries, because *no* source
could be found for the programs. IMHO, this happened partly
because keeping and filing the source code would be an
overhead function, *not* a profit center. As overhead, it is
ripe to have its budget cut out from under it.
--
+----------------------------------------------------------------+
| Charles and Francis Richmond richmond at plano dot net |
+----------------------------------------------------------------+
Peter da Silva
Mark Crispin <m...@CAC.Washington.EDU> wrote:
>Speaking as someone who has spent a lot of his life in the past 15 years
>writing portable C code, I can assert that it is *very* difficult to write
>truly portable C code (where "portable" means "will compile and run on any
>systems with a C compiler").
It's a challenge, though if you stick to using stdio and pretend you're
writing for a PDP-11 running Version 7 it's not *that* tough.
I've got some code I originally wrote at Berkeley in 1980 that's been
dragged through MS-DOS to Xenix-286, had a Tcl interpreter plugged into
the side, debugged on the Amiga, and then brought back to Solaris and
Microport. I lost track of it after that, but I suspect it'd be pretty happy
on Tru64 and OSX.
The biggest problem isn't writing portable code, it's taking what someone
without a lot of experience *thought* was portable and porting it.
>A *lot* of code these days assumes that all the world is Linux or Solaris.
A lot of code these days seems to think all the world is Red Hat 7.2 or
later. :P
>It is extremely rare to be able to compile a C program written for a 1980s
>UNIX on Linux without modification.
It's extremely rare to be able to take just about *any* code from the early
'80s in any language and run it today. Fortran is actually an exception.
I've had so much fun over the past 20 years with languages like PL/I, PL/M,
Cobol, Pascal, and Modula. The most portable code I've found has been in
Forth, Fortran, C, and PL/M. The least portable has been Forth, Modula,
Pascal, and PL/M.
Hilights include taking code I'd written in Forth on a PDP-11 and running it
unchanged on an 1802, 6502, 8080, and HP 1000, and developing a mechanical
translator to take an enormous pool of PL/M to ugly but maintainable C.
Lowlights include being unable to write a program that would run on two
different Modula implementations on the same computer, and dealing with a
bunch of PL/M code that did the equivalent of:
void *TRN1;
typedef TRN1a struct {... lots of stuff...};
typedef TRN1b struct {... lotsmorestuff...};
...
TRN1 = something complex;
do something with ((TRN1a *)TRN1)->elements;
TRN1 = TRN1 + sizeof (TRN1a);
do something with ((TRN1b *)TRN1)-> elements;
...
TRN1 = TRN1 - sizeof (TRN1a);
do something with ((TRN1a *)TRN1)->elements;
because the author apparently ran into a size limit in the structure and
didn't want to use two pointers.
>I have had little trouble running all sorts of software on WinXP. Even
>ancient DOS programs work. When problems arise, it's typically with
>programs that used hacks that depending upon some undocumented part of the
>OS instead of the official interface, and those hacks broke.
I wish that kind of behaviour was less common.
arargh...@not.at.enteract.com
<san...@haldjas.folklore.ee> wrote:
<snip>
>But you can also use an optical scanner ...
Do you know of such a program?
I have 3 scanners, and 100,000 cards.
But no program.
--
Arargh at [drop the 'http://www.' from ->] http://www.arargh.com
Basic Compiler Samples Page: http://www.arargh.com/basic/basic.html
To reply by email, change the domain name, and remove the garbage.
Charles Shannon Hendrix
> While that may be true for 30 year old programs, it was also often
> true for those same programs when they were one or two years old!
I guess that's one advantage of scripts.
Most of my oldest sources are scripts, because you can't run them
without the sources.
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
Charles Shannon Hendrix
> If you think about it, this is amazing. I wonder if old C progs will
> be as hearty in 20 years.
Yep, they will.
I run old C code all the time.
Now, if you go back to a certain point, C was actually different. It's
evolution was not steady at first.
I think the problem on the future with C will also exist for FORTRAN
and other languages. The industry is too caught up in featuritis, and
running old code is often not considered to be one of the features.
Hopefully standards will stop this for languages, but your code depends
on so many external things, I think portability problems are mostly
outside of the languages.
Peter da Silva
Charles Shannon Hendrix <csh...@SPAM.widomaker.com> wrote:
>Now, if you go back to a certain point, C was actually different. It's
>evolution was not steady at first.
I worked on code that still had "=+", but I never used a compiler that
didn't accept "+=".
>I think the problem on the future with C will also exist for FORTRAN
>and other languages. The industry is too caught up in featuritis, and
>running old code is often not considered to be one of the features.
I wrote a lot of code that turned out to depend on DEC F4P extensions.
Bad hacker, no biscuit.
>Hopefully standards will stop this for languages, but your code depends
>on so many external things, I think portability problems are mostly
>outside of the languages.
Standards help, but sometimes following the standard doesn't help as much
as knowing the traditions.
Mark Crispin
> I wrote a lot of code that turned out to depend on DEC F4P extensions.
> Bad hacker, no biscuit.
You mean F40, don't you? The crufty old Fortran compiler that generated
code that even a baby compiler writer could outdo?
Of course, today we find people writing code that depends on GCC
extensions.
David Powell
>"Jan C. Vorbrüggen" wrote:
>>
>> > I hope you've had time to reflect on the, umm,
>> > silliness of that comment in this newsgroup.
>> > I have programs on tape or listings going back
>> > to circa 1968.
>>
>> Relax, relax. Take the post with a large grain of salt and turn on
>> your detector for sarcasm.
>>
>> While I'm surprised and delighted how many here have various media with
>> (possibly) some source et al. on them, the fact remains that at least in
>> business environments, a lot of applications only exist as (patched)
>> executables, likely without any specification (especially for the patches).
>> And not all OSes are as exemplary as VMS or VMS in allowing old code to
>> execute on newer versions (SunOS/Solaris, anyone?).
>>
>This is part of what made the Y2K fixes so challenging. Some
>fixes had to be patches to the binaries, because *no* source
>could be found for the programs. IMHO, this happened partly
>because keeping and filing the source code would be an
>overhead function, *not* a profit center. As overhead, it is
>ripe to have its budget cut out from under it.
I expected to be rid of PDP-8 assembler before the OS/8 date expired
on 01-JAN-78. It didn't work out. Had to wait a few years before
Qbus stuff could replace omnibus stuff. Even with sources for most,
fixing another 8 year's of date was a challenge. Almost every
location used, many used several times over, all I needed was a
different literal to fix things. Actually, it was easier to look at
the binary, and find a fix, than to source edit, and spend days
repaging the thing.
I didn't have any Y2K problems with my own code. :) Once bitten...
Well, almost ... a few, not really all that many, almost... Only one
left that I know about, PDP-8 real-time stuff reprogrammed to i8088
hardware that I didn't expect to outlast the century.
Regards,
David P.
Peter da Silva
>> I wrote a lot of code that turned out to depend on DEC F4P extensions.
>> Bad hacker, no biscuit.
>You mean F40, don't you? The crufty old Fortran compiler that generated
>code that even a baby compiler writer could outdo?
No. "Fortran four plus".
Charles Shannon Hendrix
Mark Crispin wrote:
> C, yes; old C programs, no.
You mean outside of external dependencies?
I know there is a lot of C code out there which isn't portable, but then
I've seen a lot of non-portable code in other languages.
Usually it is assumptions about what you are writing it on, and few
people make much effort to abstract things.
Sometimes you just don't have time, but then some things are pretty
simple.
> Speaking as someone who has spent a lot of his life in the past 15 years
> writing portable C code, I can assert that it is *very* difficult to write
> truly portable C code (where "portable" means "will compile and run on any
> systems with a C compiler").
Agreed. Maybe you should write a book...
> A *lot* of code these days assumes that all the world is Linux or Solaris.
> A lot of code in the past asssumed that all the world is a VAX. Almost
> all C code assumes that memory is allocated and addressed in 8-bit bytes.
OK, this is what I meant above about assumptions.
It goes beyond this. Plenty of C code I've worked with assumes it
storing data in Oracle, printing on 80 column devices, and other things
like that.
People even do the database assumption in languages like Perl, where it
is fairly easy to eliminate specific dependencies.
> It is extremely rare to be able to compile a C program written for a 1980s
> UNIX on Linux without modification. GNU autoconf doesn't help; it just
> adds further complexity and obfuscation (speaking as someone who has
> wasted many many days doing battle with that noisome pile of reptile
> excrement).
It's great when it works, impossible when it doesn't.
I assume you noted the snafu among recent versions breaking all kinds of
stuff.
That's something the GNU crowd does that drives me nuts, and the open
source world needs to work on as well: breaking stuff, often for no good
reason.
I'd love to have something like autoconf that acutally works.
But for now, I just use a README that tells the potential user to
look at some files which control system dependent stuff.
I'm probably as guilty as the next for not making this robust enough.