Computer of the century
John Ahlstrom
> But
> what was the computer of the century?
>
--snip snip
> An
> obvious choice would be the IBM 360. It brought computing to businesses,
> for better or for worse, and gave IBM a virtual monopoly during the 60s
> and 70s. Or does some other computer deserve this distinction?
>
No one can ignore the commercial success of the 360. I think on
technological grounds, however, the B5000 might be a better choice
as the first with a number of important characteristics that then
became almost universal, to wit:
1. Commercial and Scientific data types and instructions
2. Multiprocessing
3. Protected Multi-tasking OS
4. System and User modes of operation
5. Floating IO/Processors/Channels
6. Economical high-level-language programming
7. HLL Operating System
8. Commercial Virtual Memory - Atlas beat it by some months
but was never commercially successful
--
Will we be reviled in the 9990s for not having had the
foresight to use 5 digit dates? We know the need for them
is coming. We pretend none of our apps or data bases will
still be in use by then.
Ian Joyner
> John Hendrickx wrote:
>
> > But
> > what was the computer of the century?
> >
> --snip snip
>
> > An
> > obvious choice would be the IBM 360. It brought computing to businesses,
> > for better or for worse, and gave IBM a virtual monopoly during the 60s
> > and 70s. Or does some other computer deserve this distinction?
> >
>
> No one can ignore the commercial success of the 360. I think on
> technological grounds, however, the B5000 might be a better choice
> as the first with a number of important characteristics that then
> became almost universal, to wit:
I certainly agree that the B5000 had a far reaching architecture, which has
evolved and is still present in the A Series. The OS (MCP) is still the best
in the industry, far better than Unix or NT and is worked on by a hand full
of programmers rather than the 1000s needed for these other systems.
The B5000 was given a bad wrap by Hennesy and Patterson who made silly
comparisons with the CDC 6000 at the time. I think these comments are still
in their latest edition. That the B5000 has evolved into a machine 100s or
1000s of times faster shows speed has nothing to do with stack architecture.
In fact the A Series also runs as a virtual machine on top of Intel. Perhaps
Unisys could wake up to the fact that they could as yet create the perfect
virtual machine which runs in every browser!
--
Ian Joyner
i.jo...@acm.org Eiffel for Macintosh
http://homepages.tig.com.au/~ijoyner/
http://www.object-tools.com/ot/apple.htm
Objects Unencapsulated -- The book
http://www.prenhall.com/allbooks/ptr_0130142697.html
http://www.elj.com/oue1/ (review)
Edward Reid
> No one can ignore the commercial success of the 360.
Technically, I believe OS/360 also incorporated major advances in
processing asynchronous interrupts. Part of the reason it was so late
to be delivered.
In marketing, OS/360 was the prototype of all vaporware and still one
of the outstanding examples. Early 360s were delivered with no
operating system.
In jurisprudence, this practice of vaporware was one of the major
factors leading to the interminable antitrust suit against IBM.
So really, the 360 made progress on quite a few fronts. Or perhaps it
was more erosion than progress.
> 8. Commercial Virtual Memory - Atlas beat it by some months
> but was never commercially successful
I thought it was several years: Atlas 1955, B5000 1961. But I don't
have a reference handy, except for an email someone sent me several
years ago giving these dates.
> Will we be reviled in the 9990s for not having had the
> foresight to use 5 digit dates?
"This problem will be addressed in a future paradigm shift."
(attributed iirc to a DEC response to a TR on this issue)
Edward Reid
Jeffrey S. Dutky
>
> On Tue, 21 Dec 1999 11:29:29 -0500, John Ahlstrom wrote
> > No one can ignore the commercial success of the 360.
>
> Technically, I believe OS/360 also incorporated major
> advances in processing asynchronous interrupts. Part of
> the reason it was so late to be delivered.
>
> In marketing, OS/360 was the prototype of all vaporware
> and still one of the outstanding examples. Early 360s
> were delivered with no operating system.
>
> In jurisprudence, this practice of vaporware was one of
> the major factors leading to the interminable antitrust
> suit against IBM.
>
> So really, the 360 made progress on quite a few fronts.
> Or perhaps it was more erosion than progress.
Let's not forget that Fred Brooks' experiences working
on the OS/360 project led to the seminal text in software
engineering 'The Mythical Man-Month'.
- Jeff Dutky
Tony Lima
<jahl...@cisco.com> wrote:
>John Hendrickx wrote:
>
>
>> But
>> what was the computer of the century?
>>
>--snip snip
>
>> An
>> obvious choice would be the IBM 360. It brought computing to businesses,
>> for better or for worse, and gave IBM a virtual monopoly during the 60s
>> and 70s. Or does some other computer deserve this distinction?
>>
>
>No one can ignore the commercial success of the 360. I think on
>technological grounds, however, the B5000 might be a better choice
>as the first with a number of important characteristics that then
>became almost universal, to wit:
[snip]
Second the nomination of the B5000 series, the first true
time-sharing computer. Subject to the caveat that there are
still 53.5 weeks remaining in the century and who knows what
will be invented during that period. - Tony
Alexandre Pechtchanski
[ comp.sys.unisys dropped from follow-up for obvious reasons ]
>On Tue, 21 Dec 1999 11:29:29 -0500, John Ahlstrom wrote
[ big snip ]
>> Will we be reviled in the 9990s for not having had the
>> foresight to use 5 digit dates?
>
>"This problem will be addressed in a future paradigm shift."
>(attributed iirc to a DEC response to a TR on this issue)
IIRC, the complaint was that VAX/VMS calendar counted 2000 as a leap year (as it
should), and DEC response pointed out that VMS clock will run out in year
[sorry, don't remember, but many thousand years from now]. The quoted line
dealt with the situation of clock overflow and was prefaced with congratulating
the customer for the lo-o-ong view ;-)
[ When replying, remove *'s from address ]
Alexandre Pechtchanski, Systems Manager, RUH, NY
John Ahlstrom
> 8. Commercial Virtual Memory - Atlas beat it by some months
> but was never commercially successful
Then Edward Reid wrote:
> > 8. Commercial Virtual Memory - Atlas beat it by some months
> > but was never commercially successful
>
>
According to
The History of the Development of Parallel
Computing
by Gregory V. Wilson g...@cs.toronto.edu
at http://ei.cs.vt.edu/~history/Parallel.html
The Atlas project began in 1956 and Atlas was operational in 1966
I have no information on when the 5000 project began.
I believe it was operational in 62 or 63.
Can anyone help?
JKA
--
We are having a wonderful October this
December.
John Ahlstrom
John Ahlstrom wrote:
> 8. Commercial Virtual Memory - Atlas beat it by some months
> but was never commercially successful
Then Edward Reid wrote:
> > 8. Commercial Virtual Memory - Atlas beat it by some months
> > but was never commercially successful
>
>
Actually Edward wrote:
> I thought it was several years: Atlas 1955, B5000 1961. But I don't
> have a reference handy, except for an email someone sent me several
> years ago giving these dates.
>
>
To which I responded:
> According to
> The History of the Development of Parallel
> Computing
> by Gregory V. Wilson g...@cs.toronto.edu
> at http://ei.cs.vt.edu/~history/Parallel.html
>
> The Atlas project began in 1956 and Atlas was operational in 1966
>
> I have no information on when the 5000 project began.
> I believe it was operational in 62 or 63.
>
> Can anyone help?
>
>
bill_h
> Then Edward Reid wrote:
>
> > > 8. Commercial Virtual Memory - Atlas beat it by some months
> > > but was never commercially successful
Wasn't RCA using VM around 1960?
Bill
Tucson
Mark W Brehob
> John Hendrickx wrote:
>> But
>> what was the computer of the century?
>>
> --snip snip
I have to go with the IBM PC. No other computer has had as much influence
on as many people as that series of computers. Moderate architecture
(mainly just old) but the market impact is obvious. The 360 would get my
2nd place vote. It probably made the PC possible.
Here is a different question:
Assuming a fairly free and open society, was the "Personal Computer
revolution" going to happen with or without the IBM PC? Was it an
idea who's time had come?
Mark
Gene Wirchenko
>In comp.arch John Ahlstrom <jahl...@cisco.com> wrote:
>> John Hendrickx wrote:
>
>
>>> But
>>> what was the computer of the century?
>>>
>> --snip snip
>
>I have to go with the IBM PC. No other computer has had as much influence
>on as many people as that series of computers. Moderate architecture
>(mainly just old) but the market impact is obvious. The 360 would get my
>2nd place vote. It probably made the PC possible.
IBM rode on the coattails of those who started the micro
industry.
>Here is a different question:
> Assuming a fairly free and open society, was the "Personal Computer
> revolution" going to happen with or without the IBM PC? Was it an
> idea who's time had come?
Yes. After all, there already was an 8-bit market. IBM didn't
get in until the markets had been created. If they hadn't done it,
someone else would have. I used an 8086-based system before the IBM
pc was released.
Sincerely,
Gene Wirchenko
Computerese Irregular Verb Conjugation:
I have preferences.
You have biases.
He/She has prejudices.
John Ahlstrom
> Wasn't RCA using VM around 1960?
I believe not. I believe the first RCA VM was a Spectra 70/xx (45?)
which was post 1965.
JKA
Paul DeMone
Ian Joyner wrote:
[snip]
> The B5000 was given a bad wrap by Hennesy and Patterson who made silly
> comparisons with the CDC 6000 at the time. I think these comments are still
> in their latest edition. That the B5000 has evolved into a machine 100s or
> 1000s of times faster shows speed has nothing to do with stack architecture.
And exactly what was their heresy? To sully the rarefied world of
ivory tower conjecture about instruction set encoding efficiency or
instruction fetch bandwidth requirements with grubby blue collar
details like physical implementation, and *gasp*, performance?
Stack machines were the perfect fit to brain dead compilers and expensive
flip-flops. But we are long past that phase. We can afford to perform
data flow analysis, common sub-expression reuse, and interprocedural
register allocation. We can also afford reasonably large and capacious
general purpose register files with access times comparable to the minimum
logic delay to perform basic integer arithmetic and logic functions.
Sure it is possible to speed up classical stack machine implementations
with shadow caching of the top section of the stack in multiported
registers. But why create an architecture and implemention edifice to
hide the best attributes about GPRs - high speed random access through
multiple independent ports, while adding unnecessary complication to
superscalar implementation (T9000 ring a bell?)
Setting aside the anachronistic stack architecture, what was so great
about a 48 bit architecture with tagged data words to reduce opcode
encoding size? Or an inefficient and wasteful floating point format
chosen to overlap with integer data representation? And maybe compilers
have grown up enough to decide when to do a full fledged block structured
activation stack frame with display and when to do something a bit more
lightweight. So don't build unwieldy assumptions about language
implementation into hardware.
All opinions strictly my own.
--
Paul W. DeMone The 801 experiment SPARCed an ARMs race of EPIC
Kanata, Ontario proportions to put more PRECISION and POWER into
dem...@mosaid.com architectures with MIPSed results but ALPHA's well
pde...@igs.net that ends well.
we...@nospam.mediaone.net
>According to
>The History of the Development of Parallel
>Computing
> by Gregory V. Wilson g...@cs.toronto.edu
>at http://ei.cs.vt.edu/~history/Parallel.html
>The Atlas project began in 1956 and Atlas was operational in 1966
Hadn't seen that page. Odd, it doesn't mention the Bailey Meter 756,
the first commercially successful parallel system, introduced in 1962.
(It was designed for controlling power plants, one in Australia ran until
last year.) Executed off a drum, used germanium transistors. It's now
at the Australia Computer Museum.
Gotta get a Web page up for that.
-Ric Werme
--
Ric Werme | we...@nospam.mediaone.net
http://people.ne.mediaone.net/werme | ^^^^^^^ delete
Grace H J Sturgess
were severe limitations to their Algol compiler at that time. My program
blew all their compiler limits. So, they used my program as a test of their
compiler. Atlas was the most impressive machine I have ever seen. I used the
one at London University and would go down there in the morning to run my
program. The techs would release the computer to the operators. About 4-5
operators would grab a cart and start feeding paper tape into the machine.
Pretty soon the dozen or so tape drives would be whizzing around and paper
would start pouring out of 6 or more printers. Then when my program started
compiling, everything stopped as the compiler grabbed all the resources it
could. Fortunately the compile only took a couple of minutes, then the run
would die and I would get a stack trace back dump to look at.
Grace
John Ahlstrom <jahl...@cisco.com> wrote in message
news:38613509...@cisco.com...
> Previously I wrote:
>
> John Ahlstrom wrote:
>
> > 8. Commercial Virtual Memory - Atlas beat it by some months
> > but was never commercially successful
>
>
>
> Then Edward Reid wrote:
>
> > > 8. Commercial Virtual Memory - Atlas beat it by some months
> > > but was never commercially successful
> >
> >
>
> Actually Edward wrote:
>
>
> > I thought it was several years: Atlas 1955, B5000 1961. But I don't
> > have a reference handy, except for an email someone sent me several
> > years ago giving these dates.
> >
> >
>
> To which I responded:
>
> > According to
> > The History of the Development of Parallel
> > Computing
> > by Gregory V. Wilson g...@cs.toronto.edu
> > at http://ei.cs.vt.edu/~history/Parallel.html
> >
> > The Atlas project began in 1956 and Atlas was operational in 1966
> >
> > I have no information on when the 5000 project began.
> > I believe it was operational in 62 or 63.
> >
> > Can anyone help?
> >
> >
John Ahlstrom
> And exactly what was their heresy? To sully the rarefied world of
> ivory tower conjecture about instruction set encoding efficiency or
> instruction fetch bandwidth requirements with grubby blue collar
> details like physical implementation, and *gasp*, performance?
>
>
Gee Paul, I don't remember writing anything about
stacks or tags. I did forget to mention protected multi-
tasking. And Ian explicitly said, in the passage you quoted:
> That the B5000 has evolved into a machine
> 100s or
> > 1000s of times faster shows speed has nothing to do with stack
> architecture.
>
Watch this space for a different way to look at the
Hennesey-Patterson numbers. It ain't surprising
that a 100 nsec CPU coupled with a 1 microsec
memory machine with 10 CPU functional units can beat
a 1 microsec CPU with a 4 microsec memory and
1 CPU functional unit.
What's surprising is that the 6600 didn't beat the 5500 by MUCH
more.
JKA
Toon Moene
> IIRC, the complaint was that VAX/VMS calendar counted 2000 as a leap year
(as it
> should), and DEC response pointed out that VMS clock will run out in year
> [sorry, don't remember, but many thousand years from now].
Well, if your slide rule is still operational (the models I know are all Y2K
compliant), you could even calculate it yourself.
The epoch is 18 (?) November 1858, the increment is in units of 100
nanoseconds and the clock tick is kept in a signed 64-bit integer.
Success !
[ That's what *I* would call a home assignment :-) ]
--
Toon Moene (mailto:to...@moene.indiv.nluug.nl)
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
GNU Fortran: http://egcs.cygnus.com/onlinedocs/g77_news.html
Bernd Paysan
> Stack machines were the perfect fit to brain dead compilers and expensive
> flip-flops. But we are long past that phase. We can afford to perform
> data flow analysis, common sub-expression reuse, and interprocedural
> register allocation. We can also afford reasonably large and capacious
> general purpose register files with access times comparable to the minimum
> logic delay to perform basic integer arithmetic and logic functions.
No, not in general. CPUs don't just sit in large tower cases under your
desk, they now sit almost everywhere. There are many more CPUs in
embedded systems, and all those CPUs still have to be small. Most of
them are ACCU machines, and most of their ISAs are so brain-dead that
writing a compiler is really a pain. Stack computers, which can be
implemented with about the same amount of resources as simple ACCU
machines, allow to produce several times more compact code - with a much
simpler compiler - as for the latter (especially since you can pass
parameters to common subroutines much easier), and every K of ROM is
costly in these applications (a few cents price increase of a sub-dollar
chip is expensive).
Stack machines scale well up to the point where you have to go
superscalar. The doctor tells you: if it hurts, don't do that. Stack
oriented code is already a good enough target for recompilation, so you
have that as an upgrade path.
I'd also like to point out that "closing the semantic gap" ISA elements
are at least as bad for stack computers as for any other sort of CISCy
ISA: there is typically no point in doing that, and on a stack computer,
a corresponding subroutine is also about as cheap to call as the complex
microcoded operation itself. Ok, that's only true for two-stack machines
with a separate return stack, but that should be state of the art by now
;-).
--
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/
Ian Joyner
> Ian Joyner wrote:
> [snip]
> > The B5000 was given a bad wrap by Hennesy and Patterson who made silly
> > comparisons with the CDC 6000 at the time. I think these comments are still
> > in their latest edition. That the B5000 has evolved into a machine 100s or
> > 1000s of times faster shows speed has nothing to do with stack architecture.
>
> And exactly what was their heresy?
I didn't say anything about heresy. The heresy is usually seen the other way--you
are treated as a heretic if you doubt Hennessy and Patterson. But what they did
wrong was to use a dubious comparison.
> To sully the rarefied world of
> ivory tower conjecture about instruction set encoding efficiency or
> instruction fetch bandwidth requirements with grubby blue collar
> details like physical implementation, and *gasp*, performance?
Grubby blue collar?
No I see that the two can be separate, and indeed the past 30 years have shown
this with the same ISAs getting faster and faster.
> Stack machines were the perfect fit to brain dead compilers and expensive
> flip-flops.
Brain-dead? Is your argument so groundless that you have to characterize things
this way?
No. Paradoxically, correct separation of concerns as ISA design and physical
implementation as above means that other things considerations be integrated.
Indeed the B5000 was the first machine to design hardware that took into
consideration that these things were supposed to be programmed. So they
integrated software and hardware considerations. However, the ISA has survived
well--precisely because it abstracted away from the physical details. That way
programmers were protected from such details and their software investment was
protected from the last 30 years of hardware change. Contrast that with the CDC
6000, where is that software investment today?
> But we are long past that phase.
Are we? Seems like software is more a mess than ever. And my thesis here is that
this has resulted from low-level architectures programmed in low-level languages
(although some of these are dressed up like HLLs, like C and C++).
> We can afford to perform
> data flow analysis, common sub-expression reuse, and interprocedural
> register allocation. We can also afford reasonably large and capacious
> general purpose register files with access times comparable to the minimum
> logic delay to perform basic integer arithmetic and logic functions.
>
> Sure it is possible to speed up classical stack machine implementations
> with shadow caching of the top section of the stack in multiported
> registers. But why create an architecture and implemention edifice to
> hide the best attributes about GPRs - high speed random access through
> multiple independent ports, while adding unnecessary complication to
> superscalar implementation (T9000 ring a bell?)
>
> Setting aside the anachronistic stack architecture, what was so great
> about a 48 bit architecture with tagged data words to reduce opcode
> encoding size? Or an inefficient and wasteful floating point format
> chosen to overlap with integer data representation? And maybe compilers
> have grown up enough to decide when to do a full fledged block structured
> activation stack frame with display and when to do something a bit more
> lightweight. So don't build unwieldy assumptions about language
> implementation into hardware.
>
> All opinions strictly my own.
> --
> Paul W. DeMone The 801 experiment SPARCed an ARMs race of EPIC
> Kanata, Ontario proportions to put more PRECISION and POWER into
> dem...@mosaid.com architectures with MIPSed results but ALPHA's well
> pde...@igs.net that ends well.
The key to good performance is to keep the data as close to the processor as
possible. You can do that with both register-based and stack-based machines.
Optimizations therefore--for the most part--lie below the level of instruction
set architecture.
Ian Joyner
> Mark W Brehob <bre...@cse.msu.edu> wrote:
>
> >In comp.arch John Ahlstrom <jahl...@cisco.com> wrote:
> >> John Hendrickx wrote:
> >
> >
> >>> But
> >>> what was the computer of the century?
> >>>
> >> --snip snip
> >
> >I have to go with the IBM PC. No other computer has had as much influence
> >on as many people as that series of computers. Moderate architecture
> >(mainly just old) but the market impact is obvious. The 360 would get my
> >2nd place vote. It probably made the PC possible.
I wouldn't go for a machine that two kids in a garage could have invented.
Five years earlier for the two kids, it might have been a significant
achievement. But for a multi-billion dollar corporation 5 years later, I don't
think so.
Rui Pedro Mendes Salgueiro
> Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
>> IIRC, the complaint was that VAX/VMS calendar counted 2000 as a leap year
> (as it
>> should), and DEC response pointed out that VMS clock will run out in year
>> [sorry, don't remember, but many thousand years from now].
> The epoch is 18 (?) November 1858, the increment is in units of 100
> nanoseconds and the clock tick is kept in a signed 64-bit integer.
> [ That's what *I* would call a home assignment :-) ]
% bc
2^63
9223372036854775808
^D
% units
You have: 922337203685477580800 ns
You want: years
* 29227.727
/ 3.4214088e-05
So, the clock will wrap-around in the year 31086.
BTW, Linux (for Alpha)'s clock (still the only one in which time_t is
64 bits ?) will wrap-around in 292277266665. That is 292 GigaYears.
Since the remaining life span of the Sun is only ~5 GigaYears, this
should be enough.
--
http://www.mat.uc.pt/~rps/f1/ a born-again-tifoso
Mark Sandman - Morphine, RIP (July 3th, 1999, Italy)
.pt is Portugal| `Whom the gods love die young'-Menander (342-292 BC)
Europe | Villeneuve 50-82, Toivonen 56-86, Senna 60-94
Russell Crook
>
<snip VMS clock expiry calculation>
>
> So, the clock will wrap-around in the year 31086.
>
> BTW, Linux (for Alpha)'s clock (still the only one in which time_t is
> 64 bits ?)
Not so. 64-bit Solaris 7 (and beyond) kernels have 64 bit
time_t (which is what 64-bit applications use,
although legacy 32-bit applications still use a
32-bit time_t in that environment).
Don't know about other 64-bit Unixes.
> will wrap-around in 292277266665. That is 292 GigaYears.
> Since the remaining life span of the Sun is only ~5 GigaYears, this
> should be enough.
For most things :->
>
> --
> http://www.mat.uc.pt/~rps/f1/ a born-again-tifoso
> Mark Sandman - Morphine, RIP (July 3th, 1999, Italy)
> .pt is Portugal| `Whom the gods love die young'-Menander (342-292 BC)
> Europe | Villeneuve 50-82, Toivonen 56-86, Senna 60-94
--
Russell Crook, Systems Engineer, Computer Systems
Sun Microsystems of Canada Inc. 19 Allstate Parkway, Suite 305
Markham, Ontario, Canada L3R 5A4 rmc...@Canada.Sun.com
Tel: +1-905-943-4625 Fax: +1-905-943-4601
Not speaking officially for Sun (or anyone else, for that matter).
In theory, there's no difference between theory and practice;
in practice, there is.
Alexandre Pechtchanski
>Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
>
>> IIRC, the complaint was that VAX/VMS calendar counted 2000 as a leap year
>(as it
>> should), and DEC response pointed out that VMS clock will run out in year
>> [sorry, don't remember, but many thousand years from now].
>
>Well, if your slide rule is still operational (the models I know are all Y2K
>compliant), you could even calculate it yourself.
>
>The epoch is 18 (?) November 1858, the increment is in units of 100
>nanoseconds and the clock tick is kept in a signed 64-bit integer.
Ugh, thanks, I guess... I've forgotten epoch start during the three years I
don't administer VMS anymore...
>Success !
>
>[ That's what *I* would call a home assignment :-) ]
;-)
Peter Seebach
Rui Pedro Mendes Salgueiro <r...@rena.mat.uc.pt> wrote:
>BTW, Linux (for Alpha)'s clock (still the only one in which time_t is
>Since the remaining life span of the Sun is only ~5 GigaYears, this
>should be enough.
Oh, *MY* code won't still be running, sure. That's what everyone says.
Short-term thinkers. Bah!
-s
--
Copyright 1999, All rights reserved. Peter Seebach / se...@plethora.net
C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon!
Consulting & Computers: http://www.plethora.net/
Get paid to surf! No spam. http://www.alladvantage.com/go.asp?refid=GZX636
John Stott
>Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
>
>> IIRC, the complaint was that VAX/VMS calendar counted 2000 as a leap year
>(as it
>> should), and DEC response pointed out that VMS clock will run out in year
>> [sorry, don't remember, but many thousand years from now].
>
>Well, if your slide rule is still operational (the models I know are all Y2K
>compliant), you could even calculate it yourself.
>
>The epoch is 18 (?) November 1858, the increment is in units of 100
>nanoseconds and the clock tick is kept in a signed 64-bit integer.
To quote Digital:
This base time of Nov. 17, 1858 has since been used by TOPS-10,
TOPS-20, and VAX/VMS. Given this base date, the 100 nanosecond
granularity implemented within VAX/VMS, and the 63-bit absolute time
representation (the sign bit must be clear), VMS should have no
trouble with time until:
31-JUL-31086 02:48:05.47
At this time, all clocks and time-keeping operations within VMS will
suddenly stop, as system time values go negative.
--
John P. Stott jps...@src.wisc.edu
Synchrotron Radiation Center http://www.src.wisc.edu
University of Wisconsin-Madison http://www.wisc.edu
Doug Siebert
>BTW, Linux (for Alpha)'s clock (still the only one in which time_t is
>64 bits ?) will wrap-around in 292277266665. That is 292 GigaYears.
>Since the remaining life span of the Sun is only ~5 GigaYears, this
>should be enough.
I think all the 64 bit versions of Unix have a 64 bit time_t. I know
HP-UX does. I just checked IRIX and it appears that it deliberately
uses int for time_t on 64 bit versions. Oh well, IRIX probably won't be
around in 2038 anyway. Considering how SGI is racing faster than anyone
to port all their useful proprietary Unix technology over to Linux, it
probably won't even take until 2008.
HP-UX 11's release notes mention about the 64 bit time_t but say that
due to standards, various functions like asctime() are only good up
through 9999 -- because the standards state it provides the return value
in a 26 character string. I really hope anyone who is concerned about
a Y10K problem due to this is just kidding...
--
Douglas Siebert Director of Computing Facilities
douglas...@uiowa.edu Division of Mathematical Sciences, U of Iowa
I'm not too interested in caller ID. But caller IQ, I'll pay a lot for that!
John Savard
>Will we be reviled in the 9990s for not having had the
>foresight to use 5 digit dates? We know the need for them
>is coming. We pretend none of our apps or data bases will
>still be in use by then.
That's a little *too* far off to worry about. By then, computers will
be using flexible-length data types, and will be intelligent enough to
see if the code they're executing makes sense...
John Savard (jsavard<at>ecn<dot>ab<dot>ca)
http://www.ecn.ab.ca/~jsavard/crypto.htm
Dennis Yelle
>
> On Tue, 21 Dec 1999 16:29:29, John Ahlstrom <jahl...@cisco.com>
> wrote:
>
> > John Hendrickx wrote:
> >
> >
> > > But
> > > what was the computer of the century?
> > >
>
> programmers/engineers that had one of these plastic pups as their
> first computers.
Is that the one with 3 bits of RAM and about 36 bits of ROM?
I had one of those.
Dennis Yelle
P. S.
I misread "plastic pups" as "plastic pumps" the first time I read
the message above. "plastic pumps" describes the device, because
the "clock" was a plastic thing that one moved in and out by hand
to clock the thing into the next state.
--
Want to get paid for using the internet?
If so, go to: http://alladvantage.com/go.asp?refid=BAL536
Keith R. Williams
wrote:
> John Hendrickx wrote:
>
>
> > But
> > what was the computer of the century?
> >
Well, I was thinking of the DigiComp-1. I know a lot of
programmers/engineers that had one of these plastic pups as their
first computers.
----
Keith
Victor A. Garcia
1974), we asked the question of the year 2000 problem to our boss, ( he had
a PhD. in Mathematics), guess what he say to us:
" Why worry about something so far ahead, they will have solved it by
then"
Yeah, sure..........
"John Savard" <jsa...@domain.ctry> wrote in message
news:3863ab2f...@news.prosurfr.com...
Gaven Miller
news:83vns3$8bv$1...@rena.mat.uc.pt...
> In comp.arch Toon Moene <to...@moene.indiv.nluug.nl> wrote:
> > Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
> >> IIRC, the complaint was that VAX/VMS calendar counted 2000 as a
leap year
> > (as it
> >> should), and DEC response pointed out that VMS clock will run out
in year
> >> [sorry, don't remember, but many thousand years from now].
>
> > nanoseconds and the clock tick is kept in a signed 64-bit integer.
Pardon my ignorance, but *why* was that date chosen as the epoch?
Did it correspond to something of importance to a VAX/VMS designer?
Or was the answer more mundane?
--
Due to the overwhelming amount of junkmail I have received in the past,
I can only be contacted via an intermediary : gem at gem win co nz. I
would like to apologise to the genuine respondents that this may
inconvenience.
Casper H.S. Dik - Network Security Engineer
dsie...@icaen.uiowa.edu (Doug Siebert) writes:
>I think all the 64 bit versions of Unix have a 64 bit time_t. I know
>HP-UX does. I just checked IRIX and it appears that it deliberately
>uses int for time_t on 64 bit versions. Oh well, IRIX probably won't be
>around in 2038 anyway. Considering how SGI is racing faster than anyone
>to port all their useful proprietary Unix technology over to Linux, it
>probably won't even take until 2008.
Digital Unix uses a 32 bit time_t as well, AFAIK.
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
Dan Foster
Gaven Miller <g.sw...@the.gluepot.com> wrote:
>> In comp.arch Toon Moene <to...@moene.indiv.nluug.nl> wrote:
>>
>> > The epoch is 18 (?) November 1858, the increment is in units of 100
>> > nanoseconds and the clock tick is kept in a signed 64-bit integer.
>
>Pardon my ignorance, but *why* was that date chosen as the epoch?
The reason is perfectly explained in this archived article:
(all on one line, naturally)
It's slightly too involved to summarize it here, so it's best to read it
via the web. Has to do with tracking of the Julian calendar in an 18 bit
field.
-Dan
donald tees
>John Ahlstrom <jahl...@cisco.com> wrote, in part:
>
>>Will we be reviled in the 9990s for not having had the
>>foresight to use 5 digit dates? We know the need for them
>>is coming. We pretend none of our apps or data bases will
>>still be in use by then.
>
>That's a little *too* far off to worry about. By then, computers will
>be using flexible-length data types, and will be intelligent enough to
>see if the code they're executing makes sense...
>
You mean like a human mind is intelligent enough to cure its own mental
illness?
Terry Kennedy
> http://x40.deja.com/=dnc/[ST_rn=ps]/getdoc.xp?AN=119573783.1&search=thread&CONTEXT=946118770.890896395&HIT_CONTEXT=946118751.889651233&HIT_NUM=4&hitnum=3
> (all on one line, naturally)
Here's a dirty little secret about Deja archives: You only need the stuff
up through the "AN=" field, so:
http://x40.deja.com/=dnc/[ST_rn=ps]/getdoc.xp?AN=119573783.1
In fact, the "/[ST_rn=ps]" is redundant, and if you don't mind the Deja
border fluff, you can even leave off the "/=dnc", which gets you down to:
http://x40.deja.com/getdoc.xp?AN=119573783.1
Now, the ".1" at the end of that refers to article segment 1, so you can
make this even shorter by leaving it off (and that gets you the whole
article as well, another benefit). Further, it's not a good idea to quote
the hostname that Deja gives you on a search - it's better to use the
generic www.deja.com. So, that gives us:
http://www.deja.com/getdoc.xp?AN=119573783
Lastly, the fact that the article in question was posted in 11/95 and
the article number also begins with 1195 is just a coincidence.
Terry Kennedy http://www.tmk.com
te...@tmk.com Jersey City, NJ USA
Andi Kleen
> [[ PLEASE DON'T SEND ME EMAIL COPIES OF POSTINGS ]]
>
> dsie...@icaen.uiowa.edu (Doug Siebert) writes:
>
> >I think all the 64 bit versions of Unix have a 64 bit time_t. I know
> >HP-UX does. I just checked IRIX and it appears that it deliberately
> >uses int for time_t on 64 bit versions. Oh well, IRIX probably won't be
> >around in 2038 anyway. Considering how SGI is racing faster than anyone
> >to port all their useful proprietary Unix technology over to Linux, it
> >probably won't even take until 2008.
>
> Digital Unix uses a 32 bit time_t as well, AFAIK.
I'm pretty sure it does not. Linux/alpha uses a compatible ABI to DU,
and it has 64bit time_t. It also uses 64bit timeval structs, which
causes interesting problems with tcpdump [tcpdump directly puts timevals
into its dump files].
-Andi
--
This is like TV. I don't like TV.
Brian Inglis
<g.sw...@the.gluepot.com> wrote:
>"Rui Pedro Mendes Salgueiro" <r...@rena.mat.uc.pt> wrote in message
>news:83vns3$8bv$1...@rena.mat.uc.pt...
>> In comp.arch Toon Moene <to...@moene.indiv.nluug.nl> wrote:
>> > Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
>> >> IIRC, the complaint was that VAX/VMS calendar counted 2000 as a
>leap year
>> > (as it
>> >> should), and DEC response pointed out that VMS clock will run out
>in year
>> >> [sorry, don't remember, but many thousand years from now].
>>
>> > The epoch is 18 (?) November 1858, the increment is in units of 100
>> > nanoseconds and the clock tick is kept in a signed 64-bit integer.
>
>Pardon my ignorance, but *why* was that date chosen as the epoch?
>
>Did it correspond to something of importance to a VAX/VMS designer?
>
>Or was the answer more mundane?
VMS uses the Modified Julian Day MJD = JD - 2400000.5, Julian Day
adjusted to have a recent epoch (1858-11-17 00:00:00 UTC) and
Midnight start time, which is more convenient for terrestrial use
(Julian Day is used by astronomers and has a Noon start time,
thus the .5 day adjustment).
Thanks. Take care, Brian Inglis Calgary, Alberta, Canada
--
Brian_...@CSi.com (Brian dot Inglis at SystematicSw dot ab dot ca)
use address above to reply
David Wragg
> I'm pretty sure it does not. Linux/alpha uses a compatible ABI to DU,
> and it has 64bit time_t. It also uses 64bit timeval structs, which
> causes interesting problems with tcpdump [tcpdump directly puts timevals
> into its dump files].
Yes, but from arch/alpha/kernel/osf_sys.c:
/* Translations due to the fact that OSF's time_t is an int. Which
affects all sorts of things, like timeval and itimerval. */
...
Jeffrey S. Dutky
>
> John Ahlstrom <jahl...@cisco.com> wrote, in part:
>
> >Will we be reviled in the 9990s for not having had the
> >foresight to use 5 digit dates? We know the need for them
> >is coming. We pretend none of our apps or data bases will
> >still be in use by then.
>
> That's a little *too* far off to worry about. By then,
> computers will be using flexible-length data types, and
> will be intelligent enough to see if the code they're
> executing makes sense...
Wouldn't this imply a solution to the halting problem?
Wouldn't it also violate Godel's Theorem?
- Jeff Dutky
Ariel Scolnicov
(The two are essentially equivalent). No, it wouldn't imply a
solution, as long as you qualify what you mean. Someone making this
claim had better add that computers will *often* be intelligent enough
to see if the code they're executing makes sense. Note that we can
solve any finite number of `traps' to correct (non-halting, say)
execution simply by encoding a test that disallows them.
The halting equivalent would be a program that can determine whether
*some* programs halt (some functional compilers can do this, for a
restricted set of programs; alternatively, a program which tests if
the input contains "n GOTO n" will succeed in determining for *some*
programs if they halt).
The G\"odel equivalent is to take your arithmetic axioms, and add some
axioms which decide some G\"odel propositions (for the original set of
axioms). The new set of axioms can decide these propositions, and is
consistent (if arithmetic is consistent).
However, in each case, knowing the program / set of axioms, we can
build a program / proposition which will trip it up...
[Hey! It's on topic for afc -- as 1930s computer folklore! Followups
set]
--
Ariel Scolnicov
John Winters
Rui Pedro Mendes Salgueiro <r...@rena.mat.uc.pt> wrote:
[snip]
>BTW, Linux (for Alpha)'s clock (still the only one in which time_t is
>64 bits ?) will wrap-around in 292277266665. That is 292 GigaYears.
>Since the remaining life span of the Sun is only ~5 GigaYears, this
>should be enough.
Golly, won't you have a red face when the time comes around.
John
--
John Winters. Wallingford, Oxon, England.
The Linux Emporium - the source for Linux CDs in the UK
See http://www.linuxemporium.co.uk/
Natarajan Krishnaswami
> John Savard wrote:
> > That's a little *too* far off to worry about. By then,
> > computers will be using flexible-length data types, and
> > will be intelligent enough to see if the code they're
> > executing makes sense...
>
> Wouldn't this imply a solution to the halting problem?
No, because they will (probably :-) all be space-bounded. For any
linear bounded automata (TM that using up to k*n-length tape for an
n-length input), the halting problem (will the LBA halt on a given
input) is decidable: there are finitely-many reachable configurations
(card(alphabet)^(k*n)), so any nonhalting sequence of configurations
must repeat. The uniform halting problem for LBAs (will the LBA halt on
any configuration?) is still undecidable, as it's equivalent to Post's
correspondence problem[1].
But our computers, since they have finite memory, are not unrestricted
LBAs: since there are finitely many possible *initial* configurations,
we can simply enumerate them and check if they halt, so the uniform
halting problem is also decidable. >:-)
[1] see Caron, A-C: "Linear-Bounded Automata and Rewrite Systems:
Influence of Initial Configuration on Decision Properties",
Proceedings of the Colloquium on Trees In Algebra and Programming,
Brighton, Lecture Notes in Computer Science 493, pp 74-89, 1991.
<N/>
--
you have been evaluated. you have a negative reference count. prepare
to be garbage collected. persistence is futile.
-- Erik Naggum
Andy Isaacson
> In article <83vns3$8bv$1...@rena.mat.uc.pt>,
> Rui Pedro Mendes Salgueiro <r...@rena.mat.uc.pt> wrote:
> >BTW, Linux (for Alpha)'s clock (still the only one in which time_t is
> >64 bits ?) will wrap-around in 292277266665. That is 292 GigaYears.
> >Since the remaining life span of the Sun is only ~5 GigaYears, this
> >should be enough.
>
>
> Short-term thinkers. Bah!
For a very interesting take on this, read Vernor Vinge's latest book
_A Deepness in the Sky_ (now out in paperback!). The hero is a
"Programmer-Archeologist". Here's a relevant snippet from the book:
(page 225 in my paperback edition, chapter 17)
... Take the Traders' method of timekeeping. The frame
corrections were incredibly complex -- and down at the very
bottom of it was a little program that ran a counter. Second by
second, the Qeng Ho counted from the instant that a human had
first set foot on Old Earth's moon. But if you looked at it still
more closely... the starting instant was actually about fifteen
million seconds later, the 0-second of one of Humankind's first
computer operating systems.
let that be a lesson to you... your code may still be running 8
thousand years^W^W^W250 gigaseconds from now. :)
-andy
Andy Newman
>let that be a lesson to you... your code may still be running 8
>thousand years^W^W^W250 gigaseconds from now. :)
No way. We'll hit Vernor's singularity and disappear before that.
--
Another Andy .... Chuck Berry lied about the promised land
Terje Mathisen
>
> Alexandre Pechtchanski <alex*@*rockvax.rockefeller.edu> wrote:
>
> > IIRC, the complaint was that VAX/VMS calendar counted 2000 as a leap year
> (as it
> > should), and DEC response pointed out that VMS clock will run out in year
> > [sorry, don't remember, but many thousand years from now].
>
> compliant), you could even calculate it yourself.
>
> nanoseconds and the clock tick is kept in a signed 64-bit integer.
OK!
That explains why NT also has a 64-bit time counter, with 100ns
resolution.
What's really stupid though is the fact that up to at least NT4 sp5,
there was no way to query the system time and get a reply with better
precision than the typical 10ms timer interrupt.
To add insult to injury, all the timer interrupt code is hidden down in
the HAL, so you cannot even write a simple device driver to hook into
the irq code, to allow easy interpolation between timer ticks.
This is standard on all (?) other os's supported by Network Time
Protocol (NTP), and absolutely required for a good ntp time server.
Terje
--
- <Terje.M...@hda.hydro.com>
Using self-discipline, see http://www.eiffel.com/discipline
"almost all programming can be viewed as an exercise in caching"
Jeffrey S. Dutky
>
> Jeffrey S. Dutky wrote:
> > John Savard wrote:
> > > That's a little *too* far off to worry about. By then,
> > > computers will be using flexible-length data types, and
> > > will be intelligent enough to see if the code they're
> > > executing makes sense...
> >
> > Wouldn't this imply a solution to the halting problem?
>
> No, because they will (probably :-) all be space-bounded.
> the LBA halt on a given input) is decidable: there are
> sequence of configurations must repeat. The uniform
> halting problem for LBAs (will the LBA halt on any
> configuration?) is still undecidable, as it's equivalent
> to Post's correspondence problem[1].
>
> But our computers, since they have finite memory, are not
> unrestricted LBAs: since there are finitely many possible
> *initial* configurations, we can simply enumerate them and
> check if they halt, so the uniform halting problem is also
> decidable. >:-)
Doesn't the presence of external I/O expand the state space
for our computers, possibly eliminating the finiteness due
to the size of memory?
If the state transitions are based partly on some external
input (S(memory,input) rather than simply S(memory)) and the
external imput is based on some unknown quantity outside of
the computer's memory (such as radiation source or a user's
whim) then the state space for the input could be quite
large, possibly infinite.
- Jeff Dutky
John Ahlstrom
> John Ahlstrom <jahl...@cisco.com> wrote, in part:
>
> >Will we be reviled in the 9990s for not having had the
> >foresight to use 5 digit dates? We know the need for them
> >is coming. We pretend none of our apps or data bases will
> >still be in use by then.
>
> That's a little *too* far off to worry about. By then, computers will
> be using flexible-length data types, and will be intelligent enough to
> see if the code they're executing makes sense...
>
> John Savard (jsavard<at>ecn<dot>ab<dot>ca)
> http://www.ecn.ab.ca/~jsavard/crypto.htm
By flexible-length data types do you mean something
like a 'data mark' or maybe 'word mark' at the end
of each data item? ;)
--
Plans are useless; planning is essential.
D D Eisenhower
Natarajan Krishnaswami
> Doesn't the presence of external I/O expand the state space
> for our computers, possibly eliminating the finiteness due
> to the size of memory?
Yes, but I conveniently ignoring was that. It's not much fun if your
transition function is no functional. ;-)
OTOH, we're talking 100th century tech; they'll probably just roll out
new universes to do the simulations, or something. :-)
Keith Thompson
[...]
> It's slightly too involved to summarize it here, so it's best to read it
> via the web. Has to do with tracking of the Julian calendar in an 18 bit
> field.
<pick type="nit">
It's based on the Julian Day system, not on the Julian calendar.
The Julian Day is the number of days since noon, January 1, 4713 B.C.;
it's commonly used by astronomers. The Modified Julian Day uses a
base of November 17, 1858, which is Julian Day 2,400,000.
The Julian calendar is the calendar system instituted by Julius
Caesar, and replaced by the current Gregorian calendar in 1582 (or
later in some countries).
The similarity of the names is mostly coincidental (the Julian Day
system was named by its inventor, Joseph Scaliger, after his father,
Julius Caesar Scaliger). Just to add to the confusion, it was
invented in 1582, the same year the Gregorian calendar was first
introduced.
Finally, some people have misused the term "Julian day" to refer to a
date format such as YYYYDDD, where YYYY is the year and DDD is the day
of the year. This is just plain wrong.
</pick>
--
Keith Thompson (The_Other_Keith) k...@cts.com <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://www.sdsc.edu/~kst>
"Oh my gosh! You are SO ahead of your time!" -- anon.
Doug Siebert
>let that be a lesson to you... your code may still be running 8
>thousand years^W^W^W250 gigaseconds from now. :)
I hereby promise (and note I didn't put the X-no-archive: header line
in) that if any code I've written is still running when Y10K hits,
and it has a problem due to dependance on 4 digit dates, and I'm still
alive*, just ring me up (or whatever the form of communication is) and
I'll fix it for free. You might want to point me to the archive of the
archive of the archive of the Dejanews archive of this posting, since
I expect there's a chance my memory might be a bit rusty for 8000 year
old promises at that time. 800 year old promises, no worries, but 8000
years, that's actually a pretty damn long time!
* NOTE: Offer is void if I'm just a head in a jar, ala _Futurama_ :)
Bruce Hoult
> Finally, some people have misused the term "Julian day" to refer to a
> date format such as YYYYDDD, where YYYY is the year and DDD is the day
> of the year. This is just plain wrong.
Where "some people" appears to mean every mainframce programmer I've ever
met (which mostly means Burroughs (as was) ones).
-- Bruce
Nick Spalding
> By flexible-length data types do you mean something
> like a 'data mark' or maybe 'word mark' at the end
> of each data item? ;)
Come back 1401, all is forgiven.
--
Nick Spalding
Nick Spalding
And IBM.
--
Nick Spalding
Joseph Yuska
Let me throw in CDC here also. All flavors of their operating systems
that I'm familiar with.
Joe Yuska
Philippe Tusler
Bernd Paysan
>
> On Mon, 27 Dec 1999 10:43:30 -0500, Jeffrey S. Dutky <du...@bellatlantic.net> wrote:
> > Doesn't the presence of external I/O expand the state space
> > for our computers, possibly eliminating the finiteness due
> > to the size of memory?
>
> Yes, but I conveniently ignoring was that. It's not much fun if your
> transition function is no functional. ;-)
Well, AFAIK our universe is not more than a finite state automata, too.
At least we suppose that it has a finite size, and a finite number of
elementary particles in it, all of them with a finite state (at least
the observable state is finite). So any output that just goes down into
our universe doesn't change the characteristic of the memory: it's still
finite.
> OTOH, we're talking 100th century tech; they'll probably just roll out
> new universes to do the simulations, or something. :-)
That's the real problem. No finite state machine can be simulated within
itself - you can only simulate a smaller state machine. Infinite state
machines like TMs are much better at simulating themselves, but then
have halting-problems. However, since it's not possible to create
infinite tapes, one could as well solve the halting problem with an
infinite fast TM, with the infinity of speed being (in seconds) 2 to the
power of the infinity of memory size in bits. Everything that doesn't
halt within a second is repetitive. Thus with a memory of |N| bits, you
just need a speed of |P(N)| Hz. Exercise for the reader: How fast must a
TM go to solve the halting problem, when the tape can be addressed using
real numbers?
All these Gödel, Turing and others really proved is that there isn't
just one kind of infinity, as in "you can't build something bigger" or
"you can't go beyond". There is always a larger infinity to which your
infinity looks like a zero, you just must invent a new way to count
things.
And if we are still on exercises to the reader: I've a nice proof that
|P(N)|=|N|, with the LaTeX encoding of sets in P(N). I'm still waiting
for an element of P(N) where the algorithmic description can't be
encoded as LaTeX document. And it's easy to understand that all LaTeX
documents are just elements of N.
--
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/
amoli...@visi-dot-com.com
Bernd Paysan <bernd....@gmx.de> wrote:
> [ snip ]
>All these Gödel, Turing and others really proved is that there isn't
>just one kind of infinity, as in "you can't build something bigger" or
>"you can't go beyond". There is always a larger infinity to which your
>infinity looks like a zero, you just must invent a new way to count
>things.
Well, Goedel proved something a little deeper than THAT, I think!
Thinking on it for a bit, he proved either something about finitely
generated ideals in certain Boolean algebras, or something about
the nature of Truth, depending on how you want to look at it.
He actually gets most of his credit (as, I think, does Turing)
not for what he proved but for what he conceived of as provable.
As with most Really Clever Thoughts it was one of those things that
was so outlandish as to never have occured to anyone.
bill_h
>
> John Savard wrote:
>
> > John Ahlstrom <jahl...@cisco.com> wrote, in part:
> >
> > >Will we be reviled in the 9990s for not having had the
> > >foresight to use 5 digit dates? We know the need for them
> > >is coming. We pretend none of our apps or data bases will
> > >still be in use by then.
> >
> > That's a little *too* far off to worry about. By then, computers will
> > be using flexible-length data types, and will be intelligent enough to
> > see if the code they're executing makes sense...
> >
> > John Savard (jsavard<at>ecn<dot>ab<dot>ca)
> > http://www.ecn.ab.ca/~jsavard/crypto.htm
>
> like a 'data mark' or maybe 'word mark' at the end
> of each data item? ;)
What about OTHER solar systems or planets? What if you need
more 'days' then three digits gives you?
We gotta re-think this whole DATE thing......it's too Earth-specific!
Anyway, why should EARTH be the designated Universal Time Base?
It doesn't even go back to DAY_00000000000000000000000000000001
Seems it got started up sometime later.
Bill
Tucson
Gene Wirchenko
Add Hewlett-Packard's 2000F BASIC to list.
Sincerely,
Gene Wirchenko
Computerese Irregular Verb Conjugation:
I have preferences.
You have biases.
He/She has prejudices.
jmfb...@aol.com
Joseph Yuska <jyu...@bbn.com> wrote:
>Nick Spalding wrote:
>>
>> Bruce Hoult wrote:
>>
>> > In article <yecogbb...@king.cts.com>, Keith Thompson
<k...@cts.com> wrote:
>> >
>> > > Finally, some people have misused the term "Julian day" to refer to
a
>> > > date format such as YYYYDDD, where YYYY is the year and DDD is the
day
>> > > of the year. This is just plain wrong.
>> >
>> > Where "some people" appears to mean every mainframce programmer I've
ever
>> > met (which mostly means Burroughs (as was) ones).
>>
>> And IBM.
>> --
>> Nick Spalding
>
>
>Let me throw in CDC here also. All flavors of their operating systems
>that I'm familiar with.
TOPS-10 used it for their system accounting until I got rid of it. I used
real dates. Leap years really messed up the data. We ran under
the year 1984 to test our DATE75 software (yes, children there was
a date problem before Y2K). 1975 wasn't a leap year and 1984 was.
Damn that scifi writer anyway :-).
/BAH
Subtract a hundred and four for e-mail.
Anne & Lynn Wheeler
from posting earlier this year:
http://www.garlic.com/~lynn/99.html#24
1.In 1969, Continental Airlines was the first (insisted on being the
first) customer to install PARS. Rushed things a bit, or so I hear. On
February 29, 1972, ALL of the PARS systems cancelled certain
reservations automatically, but unintentionally. There were (and still
are) creatures called "coverage programmers" who deal with such
situations.
2.A bit of "cute" code I saw once operated on a year by loading a
byte of packed data into a register (using INSERT CHAR), then used LA
R,1(R) to bump the year. Got into a bit of trouble when the year 196A
followed 1969. I guess the problem is not everyone is aware of the odd
math in calendars. People even set up new religions when they discover
new calendars (sometimes).
--
--
Anne & Lynn Wheeler | ly...@adcomsys.net, ly...@garlic.com
http://www.garlic.com/~lynn/ http://www.adcomsys.net/lynn/
Eric Chomko
: Mark W Brehob <bre...@cse.msu.edu> wrote:
[...]
: >I have to go with the IBM PC. No other computer has had as much influence
: >on as many people as that series of computers. Moderate architecture
: >(mainly just old) but the market impact is obvious. The 360 would get my
: >2nd place vote. It probably made the PC possible.
: IBM rode on the coattails of those who started the micro
: industry.
No question about it. But given that ISA came out of simply a manufacter
deciding that THAT was their buss; can anyone else boast of such a thing?
: >Here is a different question:
: > Assuming a fairly free and open society, was the "Personal Computer
: > revolution" going to happen with or without the IBM PC? Was it an
: > idea who's time had come?
: Yes. After all, there already was an 8-bit market. IBM didn't
: get in until the markets had been created. If they hadn't done it,
They always waited for someone else to get cut by the bleeding edge of
technology. Heck, their 5100 system used an Intel 8080 way back before
S-100 had been adopted.
: someone else would have. I used an 8086-based system before the IBM
: pc was released.
And I own a SWTPC 6800 that I got 4 years before the PC came out! Your
point?
CotC really is one that has had the most appeal to the masses. Trust me,
I'm no fan of the PC. In fact, when I got one with a business partner back
in the early 80s, I loathed the thing. The decision was strictly based up
getting the most popular platform at the time.
Eric
Del Cecchi
Eric Chomko <cho...@IDT.NET> writes:
|> In alt.folklore.computers Gene Wirchenko <ge...@shuswap.net> wrote:
|> : Mark W Brehob <bre...@cse.msu.edu> wrote:
|> [...]
|> : >I have to go with the IBM PC. No other computer has had as much influence
|> : >on as many people as that series of computers. Moderate architecture
|> : >(mainly just old) but the market impact is obvious. The 360 would get my
|> : >2nd place vote. It probably made the PC possible.
|>
|> : IBM rode on the coattails of those who started the micro
|> : industry.
|>
|> No question about it. But given that ISA came out of simply a manufacter
|> deciding that THAT was their buss; can anyone else boast of such a thing?
|>
|> : >Here is a different question:
|> : > Assuming a fairly free and open society, was the "Personal Computer
|> : > revolution" going to happen with or without the IBM PC? Was it an
|> : > idea who's time had come?
|>
|> : Yes. After all, there already was an 8-bit market. IBM didn't
|> : get in until the markets had been created. If they hadn't done it,
|>
|> They always waited for someone else to get cut by the bleeding edge of
|> technology. Heck, their 5100 system used an Intel 8080 way back before
|> S-100 had been adopted.
This is not correct. Where on earth did you come up with that idea?
|>
|> : someone else would have. I used an 8086-based system before the IBM
|> : pc was released.
|>
|> And I own a SWTPC 6800 that I got 4 years before the PC came out! Your
|> point?
|>
|> CotC really is one that has had the most appeal to the masses. Trust me,
|> I'm no fan of the PC. In fact, when I got one with a business partner back
|> in the early 80s, I loathed the thing. The decision was strictly based up
|> getting the most popular platform at the time.
|>
|> Eric
--
Del Cecchi
cecchi@rchland
Eric Chomko
[...]
: To quote Digital:
: This base time of Nov. 17, 1858 has since been used by TOPS-10,
: TOPS-20, and VAX/VMS. Given this base date, the 100 nanosecond
: granularity implemented within VAX/VMS, and the 63-bit absolute time
: representation (the sign bit must be clear), VMS should have no
: trouble with time until:
: 31-JUL-31086 02:48:05.47
: At this time, all clocks and time-keeping operations within VMS will
: suddenly stop, as system time values go negative.
Talk about overkill! Man, they should have at least biased the dates to
allow for more days in the past. Especially since a steady stream of the
VAXs I've seen are headed for the scrap yard, being replaced by large
SGIs.
Eric
Eric Chomko
: John Savard wrote in message <3863ab2f...@news.prosurfr.com>...
[...]
: >That's a little *too* far off to worry about. By then, computers will
: >be using flexible-length data types, and will be intelligent enough to
: >see if the code they're executing makes sense...
: You mean like a human mind is intelligent enough to cure its own mental
: illness?
That's what shrinks tell us, right?
Eric
fungus
Eric Chomko wrote:
>
> : You mean like a human mind is intelligent enough to cure
> : its own mental illness?
>
> That's what shrinks tell us, right?
>
No.
What a shrink tells you is that *his* mind is intelligent
to sure *your* mental illness
(and that it'll be $25 a session, thankyou very much)
--
<\___/>
/ O O \
\_____/ FTB.
Keith R. Williams
> They always waited for someone else to get cut by the bleeding edge of
> technology. Heck, their 5100 system used an Intel 8080 way back before
> S-100 had been adopted.
Nope. The 5100 used a proprietary processor (IIRC, Palm IV
emulating a 360).
----
Keith
Eric Chomko
: In article <84dnsu$g...@nnrp1.farm.idt.net>,
: Eric Chomko <cho...@IDT.NET> writes:
[...]
: |> They always waited for someone else to get cut by the bleeding edge of
: |> technology. Heck, their 5100 system used an Intel 8080 way back before
: |> S-100 had been adopted.
: This is not correct. Where on earth did you come up with that idea?
Well documented. They did not make the first commercial machines, that
would be Eckert and Mauchly and UNIVAC though the "THINK"-meisters Watson
had already been around for years. They claimed to have VM years after
Burroughs
pioneered it. The PC coming out in August 1981, was years after Altair,
Apple, etc. Want more? Heck, IBM is late getting into Internet sevices.
Timesharing, channel I/O, the list is long. Do us a favor, tell us one
significant invention (not innovation) that IBM made that would be
considered a first.
The worst and most costly however was OS/2. A great OS but about three
years too late.
Eric
Eric Chomko
On Wed, 29 Dec 1999 19:37:02, Eric Chomko <cho...@IDT.NET> wrote: : : >
They always waited for someone else to get cut by the bleeding edge of : >
technology. Heck, their 5100 system used an Intel 8080 way back before : >
S-100 had been adopted.
: Nope. The 5100 used a proprietary processor (IIRC, Palm IV
: emulating a 360).
Both the IBM 5100 and 5110? I have a book the 5100, I'll check.
Knowing IBM they wouldn't divulge that info. Does anyone own one that
could check? I actually worked on both an IBM 5100 and a 5110 at one time.
Eric
Martyn Richard Jones
Best wishes for the season and special greetings to all Univac gang members,
wherever we may be.
As for computer of the century, I would have though that this would be
obvious ---- UNIVAC 1100, naturally.
Anyone who pretends to know anything about the real history of computer
engineering and design will have to recognize the innovation that was
brought to the fore with the UNIVAC 1100 family throughout its long and
illustrious evolution.
After the Sperry Univac 1100 & Sperry 2200 everything else is just kids
stuff.
Although I must admit I also had some fun with .... Univac 90/30,
System/80, Mapper/10, System/11, BC/7, UDS 2000, V77, Mapper/5, Sperry (TI)
Explorer, the Sperry PC and the vintage 1985 U's.
Best regards,
Martyn R Jones
Director, Iniciativa
Del Cecchi
>
> In alt.folklore.computers Del Cecchi <cec...@signa.rchland.ibm.com> wrote:
> : In article <84dnsu$g...@nnrp1.farm.idt.net>,
> : Eric Chomko <cho...@IDT.NET> writes:
> [...]
> : |> technology. Heck, their 5100 system used an Intel 8080 way back before
> : |> S-100 had been adopted.
>
>
> Well documented. They did not make the first commercial machines, that
> would be Eckert and Mauchly and UNIVAC though the "THINK"-meisters Watson
> had already been around for years. They claimed to have VM years after
> Burroughs
> pioneered it. The PC coming out in August 1981, was years after Altair,
> Apple, etc. Want more? Heck, IBM is late getting into Internet sevices.
> Timesharing, channel I/O, the list is long. Do us a favor, tell us one
> significant invention (not innovation) that IBM made that would be
> considered a first.
>
> The worst and most costly however was OS/2. A great OS but about three
> years too late.
>
> Eric
I was referring to the notion that the 5100 had an 8080 in it. As for
inovations, bidirectional printing, the cursor, the dram, and the first
series of computers in which architecture was separated from
implementation, the s/360.
OS/2 failed for a variety of reasons. Late was only one of them.
del cecchi
Bill Todd
Gene Wirchenko <ge...@shuswap.net> wrote in message
news:386bddfa...@news.shuswap.net...
...
> I think the idea of Computer of the Century to be faintly
> ridiculous. There have been many computers out there. Were I to
> judge, I'd be looking for computers that were innovative in some
> important way that was, as a result, used in later computers. Whether
> such a system was used by ten or ten million people would be
> irrelevant. I don't think that one system could be selected while
> still being fair to the others.
I like that criterion, even though I suspect it rules out a lot of machines
near and dear to us geeks. My immediate reaction is that the Xerox Alto may
well have had the greatest effect on the way computers are used today, by
virtue of having had the greatest effect on the user interface - even though
I'm one of those who considers GUIs mere window dressing for the 'important'
stuff underneath...
- bill
Ron Hunsinger
> Joseph Yuska <jyu...@bbn.com> wrote:
>
> >Nick Spalding wrote:
> >>
> >> Bruce Hoult wrote:
> >>
> >> > In article <yecogbb...@king.cts.com>, Keith Thompson
<k...@cts.com> wrote:
> >> >
> >> > > Finally, some people have misused the term "Julian day" to refer to a
> >> > > date format such as YYYYDDD, where YYYY is the year and DDD is the day
> >> > > of the year. This is just plain wrong.
> >> >
> >> > Where "some people" appears to mean every mainframce programmer I've ever
> >> > met (which mostly means Burroughs (as was) ones).
> >>
> >> And IBM.
> >> --
> >> Nick Spalding
> >
> >
> >Let me throw in CDC here also. All flavors of their operating systems
> >that I'm familiar with.
>
> Add Hewlett-Packard's 2000F BASIC to list.
I think if you look more closely, you'll see that all these systems call
the YYDDD format a Julian DATE. (Date, not day.) Julian date and Julian day
are two entirely different animals.
-Ron Hunsinger
Gene Wirchenko
[snip]
>: someone else would have. I used an 8086-based system before the IBM
>: pc was released.
>
>And I own a SWTPC 6800 that I got 4 years before the PC came out! Your
>point?
The IBM pc's CPU was the 8088 (a cut-down version of the 8086).
The point is that, again, IBM wasn't first or even close.
>CotC really is one that has had the most appeal to the masses. Trust me,
Why?
>I'm no fan of the PC. In fact, when I got one with a business partner back
>in the early 80s, I loathed the thing. The decision was strictly based up
>getting the most popular platform at the time.
I think the idea of Computer of the Century to be faintly
ridiculous. There have been many computers out there. Were I to
judge, I'd be looking for computers that were innovative in some
important way that was, as a result, used in later computers. Whether
such a system was used by ten or ten million people would be
irrelevant. I don't think that one system could be selected while
still being fair to the others.
Sincerely,
Eric Chomko
: Hello,
: Best wishes for the season and special greetings to all Univac gang members,
: wherever we may be.
: As for computer of the century, I would have though that this would be
: obvious ---- UNIVAC 1100, naturally.
Ah, I touched that you would say that, as I am fond of the machine. Put
you're touched for saying it considering Exec 8.
: Anyone who pretends to know anything about the real history of computer
: engineering and design will have to recognize the innovation that was
: brought to the fore with the UNIVAC 1100 family throughout its long and
: illustrious evolution.
Yes, we computer geeks agree wholeheartedly.
: After the Sperry Univac 1100 & Sperry 2200 everything else is just kids
: stuff.
You still pray the the '@' sign? Tell us do you love EXEC 8?
: Although I must admit I also had some fun with .... Univac 90/30,
: System/80, Mapper/10, System/11, BC/7, UDS 2000, V77, Mapper/5, Sperry (TI)
: Explorer, the Sperry PC and the vintage 1985 U's.
Varian 77? R U nutz? The friggen buttons were crap. If the Sperry PC was
so great why didn't they make a 386 and 486 version?
Eric
: Best regards,
Keith R. Williams
wrote:
> Eric Chomko <cho...@IDT.NET> wrote:
>
> [snip]
>
> >: someone else would have. I used an 8086-based system before the IBM
> >: pc was released.
> >
> >And I own a SWTPC 6800 that I got 4 years before the PC came out! Your
> >point?
>
> The IBM pc's CPU was the 8088 (a cut-down version of the 8086).
Cut down? Sure it sported an 8-bit bus, but in the scheme of
things this really didn't matter much. This was a cost savings
that allowed the sixteen-bit "architecture" (I know this is .arch
;-).
> The point is that, again, IBM wasn't first or even close.
No, but the PC1 legitimized the PC.
> >CotC really is one that has had the most appeal to the masses. Trust me,
>
> Why?
Well, everyone has their vote. The masses have more votes? ;-)
> >I'm no fan of the PC. In fact, when I got one with a business partner back
> >in the early 80s, I loathed the thing. The decision was strictly based up
> >getting the most popular platform at the time.
>
> I think the idea of Computer of the Century to be faintly
> ridiculous. There have been many computers out there. Were I to
> judge, I'd be looking for computers that were innovative in some
> important way that was, as a result, used in later computers. Whether
> such a system was used by ten or ten million people would be
> irrelevant. I don't think that one system could be selected while
> still being fair to the others.
Oh, please! There should be no man-of-the-century? (WinstonC
would be my choice) There shoud be no Woman of the century? (Is
my wife listening?). ...no inventor, no invention, no
humanitarion... Please, spare me.
Everyone has their favorites and no one is forcing you to agree
with them. Get over it!
Then agian, the press is getting to me too with this tripe.
----
Keith
Hank Murphy
But
let's put it in perspective: what would be the car of the century? What
would be the airplane of the century?
I think that early machines which popularized the technology have a greater
claim than technical wonders which were limited in their scope. So, the
Ford Model T probably has a greater claim to car of the century than the
Ferrari Testarossa. Similarly, the DC-3 is arguably the plane of the
century, or possibly the 707, even though not very many of either are
still in commercial use.
With that criteria, I would vote for either the 360 or perhaps the PC.
Of course, in fifty years, we may elevate the Palm to computer of the
century, depending on where the technology goes...
Bill Todd wrote in message <84hat4$6il$1...@pyrite.mv.net>...
>
>Gene Wirchenko <ge...@shuswap.net> wrote in message
>news:386bddfa...@news.shuswap.net...
>
>...
>
>> I think the idea of Computer of the Century to be faintly
>> ridiculous. There have been many computers out there. Were I to
>> judge, I'd be looking for computers that were innovative in some
>> important way that was, as a result, used in later computers. Whether
>> such a system was used by ten or ten million people would be
>> irrelevant. I don't think that one system could be selected while
>> still being fair to the others.
>
jmfb...@aol.com
Right. And now I'm completely mixed up about the correct
terminology :-). I not claiming that you're wrong; it's just
that the heuristic that I used to remember which is which
seems to have become unglued. :-). I'm so glad I got rid
of them in TOPS10.
gla...@glass2.lexington.ibm.com
Failed? What do you mean 'failed'? It was (and still is) a perfectly
acceptable operating system (although it's almost impossible to find
new applications for it). It may have been surpassed by other operating
systems, but I think 'failed' is too strong of a word.
As for OS/2 being late, don't forget that originally, it was a joint
project between IBM and Microsoft. Additionally, there were some
hardware constraints in the 80286 which made life difficult.
Dave
P.S. Standard Disclaimer: I work for them, but I don't speak for them.
P.P.S. This post was created by an application running under IBM's OS/2
operating system.
gla...@glass2.lexington.ibm.com
>On Fri, 31 Dec 1999 01:10:16, ge...@shuswap.net (Gene Wirchenko)
>wrote:
>
>> Eric Chomko <cho...@IDT.NET> wrote:
>>
>> [snip]
>>
>> >: someone else would have. I used an 8086-based system before the IBM
>> >: pc was released.
>> >
>> >And I own a SWTPC 6800 that I got 4 years before the PC came out! Your
>> >point?
>>
>> The IBM pc's CPU was the 8088 (a cut-down version of the 8086).
>
>Cut down? Sure it sported an 8-bit bus, but in the scheme of
>things this really didn't matter much. This was a cost savings
>that allowed the sixteen-bit "architecture" (I know this is .arch
>;-).
>
>> The point is that, again, IBM wasn't first or even close.
>
>No, but the PC1 legitimized the PC.
>
>> >CotC really is one that has had the most appeal to the masses. Trust me,
>>
>> Why?
>
>Well, everyone has their vote. The masses have more votes? ;-)
>
>> >I'm no fan of the PC. In fact, when I got one with a business partner back
>> >in the early 80s, I loathed the thing. The decision was strictly based up
>> >getting the most popular platform at the time.
>>
>> ridiculous. There have been many computers out there. Were I to
>> judge, I'd be looking for computers that were innovative in some
>> important way that was, as a result, used in later computers. Whether
>> such a system was used by ten or ten million people would be
>> irrelevant. I don't think that one system could be selected while
>> still being fair to the others.
>
>would be my choice) There shoud be no Woman of the century? (Is
>my wife listening?). ...no inventor, no invention, no
>humanitarion... Please, spare me.
>
>Everyone has their favorites and no one is forcing you to agree
>with them. Get over it!
>
>Then agian, the press is getting to me too with this tripe.
>
>----
> Keith
Man of the century? What are the qualifications? Hitler will
probably be remembered longer than anyone else from this century.
Does that make him 'Man of the Century'?
Oh, no, did I just invoke Godwin's law? Is the usefulness of this
thread at an end? :*)
Seriously, though, I think that there are many candidates for
'Computer of the Century', depending upon the qualifications.
There's the first computer (and that's a whole can of worms itself),
the first popular mainframe, the first microcomputer, the first
really popular PC, etc.
Gene Wirchenko
>In article <38696d12...@news.shuswap.net>, ge...@shuswap.net wrote:
>
>> Joseph Yuska <jyu...@bbn.com> wrote:
>>
>> >Nick Spalding wrote:
>> >>
>> >> Bruce Hoult wrote:
>> >>
>> >> > In article <yecogbb...@king.cts.com>, Keith Thompson
><k...@cts.com> wrote:
>> >> >
>> >> > > Finally, some people have misused the term "Julian day" to refer to a
>> >> > > date format such as YYYYDDD, where YYYY is the year and DDD is the day
>> >> > > of the year. This is just plain wrong.
>> >> >
>> >> > Where "some people" appears to mean every mainframce programmer I've ever
>> >> > met (which mostly means Burroughs (as was) ones).
>> >>
>> >> And IBM.
>> >> --
>> >> Nick Spalding
>> >
>> >
>> >Let me throw in CDC here also. All flavors of their operating systems
>> >that I'm familiar with.
>>
>> Add Hewlett-Packard's 2000F BASIC to list.
>
>I think if you look more closely, you'll see that all these systems call
>the YYDDD format a Julian DATE. (Date, not day.) Julian date and Julian day
>are two entirely different animals.
You're right about the HP 2000F I suggested.
It's a confusing differentiation though as a Julian day is a
date.
Gene Wirchenko
>On Fri, 31 Dec 1999 01:10:16, ge...@shuswap.net (Gene Wirchenko)
>wrote:
[snip]
>> The IBM pc's CPU was the 8088 (a cut-down version of the 8086).
>
>Cut down? Sure it sported an 8-bit bus, but in the scheme of
>things this really didn't matter much. This was a cost savings
>that allowed the sixteen-bit "architecture" (I know this is .arch
>;-).
That's what I meant!
>> The point is that, again, IBM wasn't first or even close.
>
>No, but the PC1 legitimized the PC.
>
>> >CotC really is one that has had the most appeal to the masses. Trust me,
>>
>> Why?
>
>Well, everyone has their vote. The masses have more votes? ;-)
>
>> >I'm no fan of the PC. In fact, when I got one with a business partner back
>> >in the early 80s, I loathed the thing. The decision was strictly based up
>> >getting the most popular platform at the time.
>>
>> I think the idea of Computer of the Century to be faintly
>> ridiculous. There have been many computers out there. Were I to
>> judge, I'd be looking for computers that were innovative in some
>> important way that was, as a result, used in later computers. Whether
>> such a system was used by ten or ten million people would be
>> irrelevant. I don't think that one system could be selected while
>> still being fair to the others.
>
>Oh, please! There should be no man-of-the-century? (WinstonC
>would be my choice) There shoud be no Woman of the century? (Is
>my wife listening?). ...no inventor, no invention, no
>humanitarion... Please, spare me.
Exactly. I spoke about this at Toastmasters tonight: idea
pollution. The idea that only the number one is worth anything and
that often, the so-called number one is very arbitrarily selected.
>Everyone has their favorites and no one is forcing you to agree
>with them. Get over it!
Favourites and the <x> of the Century are two very different
things.
>Then agian, the press is getting to me too with this tripe.
Idea pollution.
James MacDonald
gla...@glass2.lexington.ibm.com wrote:
> Seriously, though, I think that there are many candidates for
> 'Computer of the Century', depending upon the qualifications.
> There's the first computer (and that's a whole can of worms itself),
> the first popular mainframe, the first microcomputer, the first
> really popular PC, etc.
Remember, really popular != really good; just look at Outlook Depressed :-)
Personally, I'd have to vote for the NeXT cube: it was certainly one of the
most revolutionary computers ever developed; what the Mac should have been.
--
James MacDonald; Acorn/NeXT/Ally McBeal
We are John Cage of Borg. Assimilation troubles us; we have to take a moment.
Poughkeepsie.
bill_h
> In <386C03...@ibm.net>, Del Cecchi <dce...@ibm.net> writes:
.....
> >OS/2 failed for a variety of reasons. Late was only one of them.
> Failed? What do you mean 'failed'? It was (and still is) a perfectly
> acceptable operating system (although it's almost impossible to find
> new applications for it). It may have been surpassed by other operating
> systems, but I think 'failed' is too strong of a word.
>
> As for OS/2 being late, don't forget that originally, it was a joint
> project between IBM and Microsoft. Additionally, there were some
> hardware constraints in the 80286 which made life difficult.
''constraints''??? You mean, it (the 80286) could run in PROTECTED
MODE?
OR, do you mean, Microsoft didn't have ANYBODY who knew how to PROGRAM
the damn thing in protected mode? Both statements are true.
Microsoft 'discovered' the protected mode of Intel's present (286) and
FUTURE (386 and up) chips about the same time they totally shafted IBM
with the 286 assembly language version of OS/2, which, naturally, was
about as portable as lead when thinking about time to move it over to
the new cpu's. MS knew this; IBM apparently didn't, or didn't care.
They got up to speed and produced Windows 3.0, and the rest is history.
Maybe IBM got what they deserved - trusting Microsoft is ALWAYS a
mistake.
Bill
Tucson, AZ
Gene Wirchenko
>In <386C03...@ibm.net>, Del Cecchi <dce...@ibm.net> writes:
>>Eric Chomko wrote:
[snip]
>>OS/2 failed for a variety of reasons. Late was only one of them.
>>
>>del cecchi
>
>Failed? What do you mean 'failed'? It was (and still is) a perfectly
>acceptable operating system (although it's almost impossible to find
>new applications for it). It may have been surpassed by other operating
>systems, but I think 'failed' is too strong of a word.
I think failed is correct as to marketing. To rework a quote of
former Canadian Prime Minister Pierre Trudeau referring to then
opposition leader Joe Clark ("Joe Who?"): "OS/Who?"
I won't argue the technical aspects.
I briefly followed a group on FIDONet on OS/2. Said group's
posts can be fairly accurately summarized as:
95+%
"OS/2 is technically superior to MS-DOS. Those people using
MS-DOS who haven't switched are idiots."
5-%
"Uh, guys, they already have investments in their existing
systems which are working well enough for them. Besides, where are
the apps?"
It seems the OS/2 advocates missed the point. The idea isn't to
have the most technically advanced OS for your platform. It's to get
the work done. Granted MS-DOS is full of kludges and bandaids and
such rot, but the work got done. The most marvellous OS in the world
doesn't do you any good if it doesn't have the apps that you need
running on it.
>As for OS/2 being late, don't forget that originally, it was a joint
>project between IBM and Microsoft. Additionally, there were some
>hardware constraints in the 80286 which made life difficult.
Announced as the OS we'd all be using. That was bad marketing.
I decided I didn't want to be dealing with such arrogance.
[snip]
>P.P.S. This post was created by an application running under IBM's OS/2
>operating system.
*This* post was created by an application running under
Microsoft's Windows 95 operating system. Like it makes much of a
difference either way.
I'm leaning towards a UNIX-like OS for personal use. MS is
arrogant, too.
Gene Wirchenko
>In message <84ikiv$nno$2...@ausnews.austin.ibm.com>
> gla...@glass2.lexington.ibm.com wrote:
>
>> Seriously, though, I think that there are many candidates for
>> 'Computer of the Century', depending upon the qualifications.
>> There's the first computer (and that's a whole can of worms itself),
>> the first popular mainframe, the first microcomputer, the first
>> really popular PC, etc.
>
>Remember, really popular != really good; just look at Outlook Depressed :-)
I prefer to call it "Look out!".
[snip]
Brian Inglis
>In alt.folklore.computers Del Cecchi <cec...@signa.rchland.ibm.com> wrote:
>: In article <84dnsu$g...@nnrp1.farm.idt.net>,
>: Eric Chomko <cho...@IDT.NET> writes:
>[...]
>: |> They always waited for someone else to get cut by the bleeding edge of
>: |> technology. Heck, their 5100 system used an Intel 8080 way back before
>: |> S-100 had been adopted.
>
>: This is not correct. Where on earth did you come up with that idea?
>
>Well documented. They did not make the first commercial machines, that
>would be Eckert and Mauchly and UNIVAC though the "THINK"-meisters Watson
>had already been around for years. They claimed to have VM years after
>Burroughs
>pioneered it. The PC coming out in August 1981, was years after Altair,
>Apple, etc. Want more? Heck, IBM is late getting into Internet sevices.
>Timesharing, channel I/O, the list is long. Do us a favor, tell us one
>significant invention (not innovation) that IBM made that would be
>considered a first.
IBM always referred to Virtual Memory as Virtual Storage (VS).
Virtual Machine (VM) operation was definitely one of their
inventions -- running any of their OSes under a "hypervisor",
including another instance of itself.
Their 360/370/390 architecture standardization was inventive, as
was the use of that assembler for everything from microcode to
communication controllers.
Their HLLs include ForTran, PL/I, APL and REXX.
They have been most inventive in the disk drive arena:
Winchesters, Thin Film heads and the new Giant? Magneto-Resistive
heads, multiple heads per arm to reduce seek time, among others
IIRC.
Scanning Tunneling Microscope was another recent invention of
theirs.
And those are just what I can come up with off the top of my
head. Please feel free to correct me if any of these were based
on the ideas of others elsewhere.
>The worst and most costly however was OS/2. A great OS but about three
>years too late.
>
>Eric
Thanks. Take care, Brian Inglis Calgary, Alberta, Canada
--
Brian_...@CSi.com (Brian dot Inglis at SystematicSw dot ab dot ca)
use address above to reply
Victor A. Garcia
Date '
Julian day has been an obsolete term since the late 1580's when they
implement the Gregorian calendar.
Victor A. Garcia
OS, it does not matter how good OS/2, as an OS is or was, it FAILED, not
users, no applications... KAPUT.
If people were logical about computers, and only use the best machines, then
everybody at home will be using the Commodore Amiga, the most advance PC
ever built, and people at small outlets will still be using CTOS machines,
first and really Plug-and-Play network System, guess where both of then are
right now.....
For better or for worse, Marketing rules......
> >OS/2 failed for a variety of reasons. Late was only one of them.
> >
> >del cecchi
>
> Failed? What do you mean 'failed'? It was (and still is) a perfectly
> acceptable operating system (although it's almost impossible to find
> new applications for it). It may have been surpassed by other operating
> systems, but I think 'failed' is too strong of a word.
>
> As for OS/2 being late, don't forget that originally, it was a joint
> project between IBM and Microsoft. Additionally, there were some
> hardware constraints in the 80286 which made life difficult.
>
> Dave
Victor A. Garcia
I'll vote for him.
Pete Mullins
Victor A. Garcia <vgar...@tampabay.rr.com> wrote in message
news:lChb4.5236$kG2....@typhoon2.tampabay.rr.com...
> You IBM'ers are worse than us , yes it FAILED, people run applications,
not
> OS, it does not matter how good OS/2, as an OS is or was, it FAILED, not
> users, no applications... KAPUT.
Ah! At least this post is relevant to comp.sys.unisys.
b...@mycomputer.example.net
: Julian day has been an obsolete term since the late 1580's when they
: implement the Gregorian calendar.
The real Julian Day is not obsolute in astronomy. Today's date is
2451545 (as of 2000 January 1 12:00 noon UTC).
--
Night is the shadow of the Earth.
Colin Zealley
"Brian Inglis" <Brian.do...@SystematicSw.ab.ca> wrote in message
news:p4lq6sk75e52cnj4e...@4ax.com...
[large snip]
> IBM always referred to Virtual Memory as Virtual Storage (VS).
> Virtual Machine (VM) operation was definitely one of their
> inventions -- running any of their OSes under a "hypervisor",
> including another instance of itself.
>
I think you may find that the concept of the "virtual machine was actually
implemented by ICL in the UK, with the 2900 series, before IBM had it
working. But I'm not sure of the dates - I might have this wrong.
Colin Zealley
[more snip]
Hank Murphy
a keyboard to an OS/2 system, and get screen and printed output from OS/2.
You just don't know it.
Most ATMs still run with OS/2 (Interbold).
Many in-flight seat systems (BE Aerospace) run with OS/2.
Many emergency dispatch systems still run OS/2 (Southern California Gas
Company, L.A. County Fire, to name two).
So, where human life and money are involved, OS/2 is still a good option.
Brian Inglis
CP/67 (1968) predates my entry into computing; I followed the ICL
2900 humming and hawing during my early years.
Your comment implies that another copy of VME/B or VME/K could be
run under itself for testing: which OS was it? how was it
implemented? any pointers?
Martyn Richard Jones
This (the commercialisation of the technology) must have been roughly around
1976. I remember we used an ICL 1904S* with George III up until then at
which point we switched to 2900's. with G3 VME.
Victor A. Garcia
Redmond, Burroughs developed the MicroA on an OS/2 platform, it's still
running some of our A7's, and just last year we installed more then 400 WS
and servers for an Alabama Bank. (This year, we will upgrade then to NT,
they want Office).
But as a general use OS, it's dead.
"Hank Murphy" <hmu...@earthlink.net> wrote in message
news:Q3tb4.818$Ec5....@newsread1.prod.itd.earthlink.net...
Charles Richmond
>
> And what is wrong with that, compared to Joseph Stalin, he was a nice man.
> I'll vote for him.
> >
> > Man of the century? What are the qualifications? Hitler will
> > probably be remembered longer than anyone else from this century.
> > Does that make him 'Man of the Century'?
> >
>
Stalin is so *bad*! The Nazis are estimated to have killed 6 million Jews.
Stalin's purges killed 20 million people!!! Although many were Jews, they
all were *people*!!! To give some idea, there are only *two* states in the US
that have 20 million people or more (California and Texas). In fact, the
population of Texas is almost exactly 20 million. Imagine that someone had
the *entire* population of Texas executed. That gives you some idea of the
real charm of Joseph Stalin...all the charm of a poisonous snake!!!
--
+-------------------------------------------------------------+
| Charles and Francis Richmond <rich...@plano.net> |
+-------------------------------------------------------------+
David M. Razler
| Victor A. Garcia wrote:
| >
| > And what is wrong with that, compared to Joseph Stalin, he was a nice man.
| > I'll vote for him.
| > >
| > > Man of the century? What are the qualifications? Hitler will
| > > probably be remembered longer than anyone else from this century.
| > > Does that make him 'Man of the Century'?
| > >
| >
| I must agree with Victor...it is *not* that Hitler is so nice; it's that
| Stalin is so *bad*! The Nazis are estimated to have killed 6 million Jews.
| Stalin's purges killed 20 million people!!! Although many were Jews, they
| all were *people*!!! To give some idea, there are only *two* states in the US
| that have 20 million people or more (California and Texas). In fact, the
| population of Texas is almost exactly 20 million. Imagine that someone had
| the *entire* population of Texas executed. That gives you some idea of the
| real charm of Joseph Stalin...all the charm of a poisonous snake!!!
The Nazis killed roughly 10 million people in death camps and forced labor
camps, of which Jews accounted for about 6 million <more or less> or about
2/3rds of the world Jewish population. Other notable targets included 1-2
million Romany ('Gypsies'), those 'Aryans' with birth defects or mental
illness, homosexuals, Communists, Socialists, labor leaders.
Millions more civilians were murdered (directly, and deliberately, as opposed
to those standing where the bombs fell) by those carrying out Hitler's
genocide/thought police orders in less better-death-through-technology
environments, and an exact count is difficult to come by.
As for Josef Stalin: His actual record is rather pitiful in comparison, with
the purges resulting in *maybe* 2 million murders or so, either directly or in
institutions of forced labor and death.
The big numbers ascribed to Stalin (granted, a very evil man on my list too),
come from a very questionable claim concerning food distribution in The
[former] Ukranian SSR, now an independent state again.
It is argued by some professional anti-Stalinists and Ukranian nationals that
the famine suffered in the SSR was due to a deliberate re-routing of food to
"starve Stalin's enemies to death." Similar claims are made regarding Georgia
and other areas of the old USSR, which suffered famine under Stalin.
However, the evidence simply is not very compelling. While food was moved
about, it appears that there was no grand plan to kill off a region or people.
Nor is there an accurate count of those killed. (The Nazis were fantastic
bookkeepers - one reason we can say how many were officially (as opposed to
informally) murdered down to a few hundred thousand.
One might further argue that, given the USSR's economy, ravaged by both World
War II and I forget how many double-digit million war casulties, combined
Stalin's misdirection of economic resources to re-arm, control lands captured
during WW II and develop atomic weapons, etc. would have led to the same
number of deaths-by-starvation throughout the country. There were also
problems caused due to the escalation of forced collectivization (begun by
Lenin) that led to a general drop in food production that had been running at
an all-time high (in spite of collectivization) during the Great Patriotic War
(aka WW II).
Stalin was a murderous dictator - but he was not genocidal. The purges were
aimed at individuals or relatively small groups he feared, accounting for his
relatively low official death toll.
Hitler was both, and, I hope, will retain the record for genocide in both
number and probably percentage of given populations.
Given the current behavior of people at large, I fear the percentage of a
given population number may fall, if it has not already been topped.
dmr
David M. Razler
david....@worldnet.att.net
Daniel P. B. Smith
<bre...@cse.msu.edu> wrote:
> Here is a different question:
> Assuming a fairly free and open society, was the "Personal Computer
> revolution" going to happen with or without the IBM PC? Was it an
> idea who's time had come?
Oh, it was, absolutely. All IBM contributed was the value of their name
in validating the concept. Once Visicalc was out, the success of
personal computers was a foregone conclusion. Even IBM was surprised by
the market acceptance. (Indeed, some believe that IBM corporate had
expected it to fail--"OK, the best way to prove there isn't a market for
this thing is to let them build it. Don't interfere with them at all,
other then demanding they use a slow processor that won't cannibalize
the DataMaster. Give them enough rope and they'll hang themselves...")
Both IBM and Lotus were remarkable at making a very real but
nevertheless _incremental_ improvements and turning them into
blockbuster commercial successes.
Anyway, at the time the IBM PC came out _all_ the big companies were
coming out with big-name ho-hum incremental-improvement-at-best boxes.
Xerox was out a few months before IBM; Digital, a few months later
(hence the unfairness of Monday-morning quarterbacking about their not
realizing the importance of "IBM compatibility.") Wang came out with
theirs in the same time frame, too. And Apple screwed up with the Apple
///, which was both a stupid design and initially had reliability
problems. If it hadn't been IBM, it would have been Xerox or Digital or
Wang... or, quite possibly, Apple with the Mac in 1984.
--
Daniel P. B. Smith
current email address: dpbs...@bellatlantic.net
"Lifetime forwarding address:" dpbs...@alum.mit.edu