System 1022 Database System
System 1022 Guy
to it now? Does CompuServe still run any 36bit hardware? Thanks.
Pat Farrell
> Does anyone know if it still runs anywhere? Does CCA have the rights
> to it now? Does CompuServe still run any 36bit hardware? Thanks.
What's a CCA?
Compuserve was bought by AOL long ago. Maybe 1993?
For a while, they had some legacy 36 bit systems, I think mostly Foonley or
home grown log after DEC stopped dealing with them.
Up until about 10 years ago, a company (forget the name) was running a bunch
of our old KLs in our old space, 1777 North Kent Street. We sold them the
hardware, space, contracts, etc.
I started collecting some folklore on it on
http://www.pfarrell.com/technotes/sys1022.html
corrections, extensions, etc. welcome
--
Pat
Sarr J. Blumson
: Does anyone know if it still runs anywhere? Does CCA have the rights
: to it now? Does CompuServe still run any 36bit hardware? Thanks.
CCS Still offers the VAX version, System 1032, as a product:
http://www.cca-int.com/prodinfo/s1032.html
That's Computer Corporation of America, Pat :-)
--
--------
Sarr Blumson sarr.b...@alum.dartmouth.org
http://www-personal.umich.edu/~sarr/
jmfb...@aol.com
"Sarr J. Blumson" <sa...@defender.gpcc.itd.umich.edu> wrote:
>System 1022 Guy <qqdc...@gmail.com> wrote:
>: Does anyone know if it still runs anywhere? Does CCA have the rights
>: to it now? Does CompuServe still run any 36bit hardware? Thanks.
>
>CCS Still offers the VAX version, System 1032, as a product:
>
>http://www.cca-int.com/prodinfo/s1032.html
>
>That's Computer Corporation of America, Pat :-)
>
bits from 36 to 32?
Pat Farrell
> System 1022 Guy <qqdc...@gmail.com> wrote:
> : Does anyone know if it still runs anywhere? Does CCA have the rights
> : to it now? Does CompuServe still run any 36bit hardware? Thanks.
>
> CCS Still offers the VAX version, System 1032, as a product:
>
> http://www.cca-int.com/prodinfo/s1032.html
>
> That's Computer Corporation of America, Pat :-)
Well, I googled CCA and got three zillion hits.
None were obvious a good home.
So they bought it from AOL? Perhaps when AOL picked up Compuserve, they spun
off silly stuff they didn't want?
Wow, I see that they are still peddling Model 204. That is a name from the
past.
--
Pat
System 1022 Guy
I was hired by Andrew Garland (co-author of 1022)(Software House) back
in 1981 coming from BBN where I had worked on ARPANET. By the end
(1989) I was sys architect for the last devel version (I think 121?) of
1022 at CompuServe Data Technologies.
BTW on previous posts on the origin of the name 1022: I was perhaps one
the closest to Andy and 1022, and he NEVER told me the origin other
than he just thought it sounded good. It seemed to be a sensitive
subject with him, so I never asked again.
I think CCA got 1022 when they got 1032 from AOL-CDT, but I don't
really know.
I had no idea that the 20 had been revived until last week. I never
thought I'd be able to show my son what I done with a big chunk of my
life.
This may sound really sick, but I still have dreams in which I'm doing
half word manipulation, JSYS's, and variable sized bytes(oxymoron-try
explaining that to a UNIX) in MACRO-10.
Well now I feel like a Blue's Brother trying to put the band (SH 1022
devels) back together to save the orphanage(1022).
I always felt like I had some unfinished business. I wanted to
implement alot of tricks with SMAP and extended sections. EG: We had a
PMAP cache for file I/O(like PA1050) in extended sections. I wanted to
make it a SMAP cache. Another one: 1022 loads LINK in another section
and use it to load user functions dynamically from REL files.
We had customer contracts that provided that the customer had access to
an escrow tape in the event if we ever stopped supporting 1022. So does
anyone know where one of the escrow tapes is? (BTW: The only time the
source was given out: I had to quietly provide a source tape to NSA.
Now what do you think THEY wanted it for????
Any help finding Bill Alexander or Phil Jensen? I've just found P
Duff. Thanks.
System 1022 Guy
Pat Farrell wrote:
> I started collecting some folklore on it on
> http://www.pfarrell.com/technotes/sys1022.html
> corrections, extensions, etc. welcome
>
On your page:
--------------------------------------------------------------------------
While System 1022 ran on Tops-20, it was fundamentally a TOPS-10
program, written in Macro-10 (the Tops-10 assembly language). It was
never "ported" to native Tops-20 JSYS I/O while the Tops-20 product
line was alive and commercially viable. There was a version released
after the cancellation of the "jupitor" project, but by then it was too
little too late.
--------------------------------------------------------------------------
It was definately ported and was native by 83(?)(version 115) I guess.
Dev went atleast to version 121.
We had many customers who needed more performance from their 10/20's.
They were quite unhappy waiting for DEC. We found we could help them.
They were overjoyed. We sometimes devoted half of our devel resources
to performance vs features (eg this case: native mode TOPS-20). We
probably were TOO good. In the big picture we probably hurt our
company and DEC. Each version ran 15 to 25% faster! This was unheard
of. Usually more features meant bigger and slower software (look at
Windows!). Our customers got spoiled. (one large customer with many
KL's once equated a new version of 1022 with 1 1/2 KL's not bought)
When DEC pushed them towards VAX, they looked at our companion product
1032. It was elegant, but not a speed demon. When they benckmarked,
they found they would need many 8600's to replace 1022 from the 10/20.
Some of these were BIG customers, such as yourself. They decided to
look elsewhere. One decided to buy US. Compuserve found that by
improving 1022, they could delay buying DEC hardware until they had
built their own SC clones. 1022's performance created expectations that
could'nt be met via VAX or 1032.
Pat Farrell
> Hi Pat(one of my biggest customers back then) - Dave Carr here.
Hi!
I think we were the number two revenue source for Chas and Andy, below
CompuServ of course.
> BTW on previous posts on the origin of the name 1022: I was perhaps one
> the closest to Andy and 1022, and he NEVER told me the origin other
> than he just thought it sounded good. It seemed to be a sensitive
> subject with him, so I never asked again.
Well ol' Mumblefratz Gazingus told the best story, some yarn about
running on tens and 20s too, but that was many years after it was called
1022. I always believed it was the date that Charlie got it to do the first
FIND command. It was well established folklore at FDC that it was built on
night cycles left over from timesharing at ol 400 Toten Pond Road. And
stories that Charlie spent so long waiting for compiles that he would crack
the combination locks into the computer room by raw presistance.
I first ran into it in 75, maybe 74, running on the pre-fire KAs and KLs at
First Data.
> I had no idea that the 20 had been revived until last week. I never
> thought I'd be able to show my son what I done with a big chunk of my
> life.
So you get an account on Allen's KL yet to show him the real thing?
Of course, one of Crispen's PCs is way faster than any KL could dream of.
> This may sound really sick, but I still have dreams in which I'm doing
> half word manipulation, JSYS's, and variable sized bytes(oxymoron-try
> explaining that to a UNIX) in MACRO-10.
I still can't think of a byte as being anything not pointed to by a byte
pointer.
> We had customer contracts that provided that the customer had access to
> an escrow tape in the event if we ever stopped supporting 1022. So does
> anyone know where one of the escrow tapes is?
I think we (AMS) had such a contract, but I never saw the tapes.
> (BTW: The only time the
> source was given out: I had to quietly provide a source tape to NSA.
> Now what do you think THEY wanted it for????
Doing rapid FIND commands?
--
Pat
Pat Farrell
> Pat Farrell wrote:
>> I started collecting some folklore on it on
>> http://www.pfarrell.com/technotes/sys1022.html
>> corrections, extensions, etc. welcome
>
> On your page:
> --------------------------------------------------------------------------
> While System 1022 ran on Tops-20, it was fundamentally a TOPS-10
> program, written in Macro-10 (the Tops-10 assembly language). It was
> never "ported" to native Tops-20 JSYS I/O while the Tops-20 product
> line was alive and commercially viable. There was a version released
> after the cancellation of the "jupitor" project, but by then it was too
> little too late.
> --------------------------------------------------------------------------
>
> It was definately ported and was native by 83(?)(version 115) I guess.
> Dev went at least to version 121.
It was too long ago, but sometime in that period it was ported.
And they finally fixed the half word record number limit.
We even had a copy of it. But it was not stable enough for production usage
while I was involved. We had customers like the US General Accounting
Office (now Accountablility) using it for accounting, and they did not
tolerate much downtime.
The simple business fact was that by 1982 or so, we were dead without
massively faster/bigger systems. That is when the KL stopped being
commercially viable. We bought more, and used them, but the bloom was off
the rose.
And sometime in 80/81 we had stopped using 1022's IO for updates. It was
just too slow. Once Tops-20 had extended addressing, we turned it on for
users and used separate sections to PMAP in files. We might have been doing
that as early as 1980. Then we had batch processes that took the PMAP'd
files and did updates to the databases, so users could do queries and
reporting. They loved the retrival and reporting code.
Especially after we forced Software House to rewrite their DPL to allow
reasonable if/then/else structuring with our SDPL hack, stolen directly
from FLECS and Ratfor.
I will grant you that one can argue with my page's terms, but
"while the Tops-20 product line was alive and commercially viable"
was over before the Jupitor death annoucement.
> to performance vs features (eg this case: native mode TOPS-20). We
> probably were TOO good. In the big picture we probably hurt our
> company and DEC. Each version ran 15 to 25% faster!
It was nice, but I don't think it was important in the big picture.
We needed five times more power, not 20%. A five times faster system
(end-to-end) for three times the money would have been attractive. Ten
times would have been better.
> When DEC pushed them towards VAX, they looked at our companion product
> 1032. It was elegant, but not a speed demon. When they benckmarked,
> they found they would need many 8600's to replace 1022 from the 10/20.
> Some of these were BIG customers, such as yourself. They decided to
> look elsewhere. One decided to buy US. Compuserve found that by
> improving 1022, they could delay buying DEC hardware until they had
> built their own SC clones. 1022's performance created expectations that
> could'nt be met via VAX or 1032.
Now some of that was because of a decent engineering decision that Marty and
others made (maybe even Dave), I assume with Charlie and Andy's approval.
They assumed that CPUs would get faster over time, but that disk drives and
memory would not get faster, cheaper and better in proportion. So they put
in a lot of bit packing to make the DMS files be small, which saved disk
and saved virtual memory, which VMS needed since Cutler was such a bozo.
No one would have expected that disks would follow their own equivalent to
Moore's law. For a while, with the 780/750/730, and even 8600, CPUs did get
faster more quickly. But the user's demanded more.
By the time DEC figured it out, it was too late. Timesharing was dead except
for the likes of Compuserv and AOL. At least until 1990 or so when
client-server became the rage.
--
Pat
Mark Crispin
> So you get an account on Allen's KL yet to show him the real thing?
FWIW, Paul Allen's KL runs TOPS-10. The TOPS-20 machine is an XKL
machine.
> Of course, one of Crispen's PCs is way faster than any KL could dream of.
fsCrispen$Crispin$$
Lingling.panda.com is based on an Athlon 1700+, which at the time that I
built it (2001) was quite the speed demon at 1467 MHz. It's about 75
times faster than a KS, 16 times faster than a KL, and 10 times faster
than an XKL machine.
In modern CPU terms, Lingling is about the equivalent of 50MHz; that's
faster than my first UNIX desktop and first Mac laptops. klh10 is quite
remarkable in that it has such a low emulation cost factor; most emulators
are considered good if the cost faster is 100x.
Today, an Athlon 1700+ is old technology. It shouldn't be difficult to
build a modernized Lingling that is at least twice as fast.
However, even as matters stand, Lingling is a very fast TOPS-20 system.
The TOPS-20 monitor builds from sources in about 12.5 minutes. As many of
us remember, building the monitor from sources was a very time-consuming
task, even on a KL.
It would be really nice to get a copy of System 1022; and even nicer to
get permission to include it in the Panda distribution. I got permission
for to include Scribe in the distribution, and am trying to locate the
legal owners of NCPCalc and TOPS-20 Macsyma (not as easy as you think).
-- Mark --
http://panda.com/mrc
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.
System 1022 Guy
Pat Farrell wrote:
>
> Especially after we forced Software House to rewrite their DPL to allow
> reasonable if/then/else structuring with our SDPL hack, stolen directly
> from FLECS and Ratfor.
FLECS - I was the one who did the intial port to DecSystem. The
creator (i can't remember name) got in touch with me thru Syracuse U
(circa 1975).
At BBN modified COMPIL to seamlessly PreProcess with ratfor and autorun
Fortran.
System 1022 Guy
Pat Farrell wrote:
> So you get an account on Allen's KL yet to show him the real thing?
> Of course, one of Crispen's PCs is way faster than any KL could dream of.
>
Paul Allen? When I last checked, links seemed very old and didn't get
me anywhere.
Any help. Thanks.
System 1022 Guy
sho...@trailing-edge.com
> Does anyone know if it still runs anywhere? Does CCA have the rights
> to it now? Does CompuServe still run any 36bit hardware? Thanks.
Some executables from a System 1022 installation and other random files
seem to live on the DECUS LIB10-01 tape:
http://pdp-10.trailing-edge.com/decuslib10-01/index.html
in the 43,50126 directory.
Almost certainly not enough to get it running again, but what do I
know, cuz I certainly never installed it before in my life!
Seems to be release 115 or so of System 1022 circa early 1983.
Some of the most interesting stuff on the DECUS tapes seems to be the
stuff that "accidentally" made it on there. Contrast with the DEC
software distribution tapes which seem to be minty-fresh clean in terms
of not having random detritus on them!
Tim.
John Everett
wrote:
>Well ol' Mumblefratz Gazingus told the best story, some yarn about
>running on tens and 20s too, but that was many years after it was called
>1022. I always believed it was the date that Charlie got it to do the first
>FIND command. It was well established folklore at FDC that it was built on
>night cycles left over from timesharing at ol 400 Toten Pond Road. And
>stories that Charlie spent so long waiting for compiles that he would crack
>the combination locks into the computer room by raw presistance.
>
>I first ran into it in 75, maybe 74, running on the pre-fire KAs and KLs at
>First Data.
Pat:
There were no pre-fire KLs at 400 Toten Pond Road. We took delivery of
our first KLs at 40 Second Avenue. When Swithers and I ran our first
monitor build on a KL we just laughed and laughed over the
unbelievable speed. ;-)
BTW, I heard that same story from Mumblefratz "GS" Gazingus. Perhaps
we were together at the time.
jeverett3<AT>earthlink<DOT>net http://home.earthlink.net/~jeverett3
Pat Farrell
> On Mon, 25 Sep 2006 22:20:12 -0400, Pat Farrell <no...@nospam.info>
> wrote:
>
>>Well ol' Mumblefratz Gazingus told the best story, some yarn about
>>running on tens and 20s too, but that was many years after it was called
>>1022. I always believed it was the date that Charlie got it to do the
>>first FIND command. It was well established folklore at FDC that it was
>>built on night cycles left over from timesharing at ol 400 Toten Pond
>>Road. And stories that Charlie spent so long waiting for compiles that he
>>would crack the combination locks into the computer room by raw
>>presistance.
>>
>>I first ran into it in 75, maybe 74, running on the pre-fire KAs and KLs
>>at First Data.
>
I knew that, dumb keyboard.... :-)
It was a bunch of KAs and KIs that burned. Boy, did they stink.
> We took delivery of
> our first KLs at 40 Second Avenue. When Swithers and I ran our first
> monitor build on a KL we just laughed and laughed over the
> unbelievable speed. ;-)
Actually, I ran the first KL that you and Swithers hacked our (FDC's)
monitor on out in the Fishbowl in Marlboro. Probably for a month before DEC
could deliver one to 40 Second Avenue with the big power sags....
> BTW, I heard that same story from Mumblefratz "GS" Gazingus. Perhaps
> we were together at the time.
I think so, and he told it often in his job selling System 10?2 before he
moved to Oracle and made a fortune.
--
Pat
Pat Farrell
> System 1022 Guy wrote:
>> Does anyone know if it still runs anywhere? Does CCA have the rights
>> to it now? Does CompuServe still run any 36bit hardware? Thanks.
>
> Some executables from a System 1022 installation and other random files
> seem to live on the DECUS LIB10-01 tape:
>
> Seems to be release 115 or so of System 1022 circa early 1983.
Would be interesting to know if the pieces contain the usual Software House
death code. In the early days, Charlie would sell it by taking the software
on a tape, installing it, and letting people play with it for a while. They
would usually get hooked, and he'd then come back and say "pay me or it
will stop working." Their were several layers of date detection logic. The
first level was trivial to find with DDT, it got the system time with a UUO
and checked it with a simple SKIP/CAIx instruction (or maybe a skipret from
the checking routine). So anyone could patch it out and it would keep
running.
Or look like it was running.
There were several time bombs sprinkled like land mines in the code, which
checked to see if the initial first gate was patched out. Depending on
which land mine went off, bad things would happen, some without notice.
Sometimes they even went off when you'd paid the legal license fee.
That was ugly.
--
Pat
System 1022 Guy
sho...@trailing-edge.com wrote:
> Some executables from a System 1022 installation and other random files
> seem to live on the DECUS LIB10-01 tape:
>
> http://pdp-10.trailing-edge.com/decuslib10-01/index.html
>
> in the 43,50126 directory.
logged in as anonymous
but couldn't get any further
dir didn't return anything and i didn't know the format of the dir i
should use in cd command
help?
sho...@trailing-edge.com
Didn't mean to skip a few generations there :-).
That's a URL you type into your web browser.
Some of us prefer to telnet to port 80 instead of using a browser,
sometimes... better user interface!
Tim.
System 1022 Guy
Yes, I see the index using the browser link from my windows desktop.
But I wanted to move a 36bit file(eg EXE) from there to twenex. I
telneted to twenex and ftped from there to pdp-10.trailing-edge.com,
but got no further. How else do I get a binary file from the archive
to a 10/20?
Thanks
sho...@trailing-edge.com
Best way is to move the tape image to the Twenex system and unpack it
there. If it's an emulated Twenex system then the system manager ought
to be able to help you out.
Most FTP servers these days don't really support the 36-bit filetypes,
unfortunately.
In the past I've attempted to bring up the subject of 36-bit MIME
encoding formats on this newsgroup (after all, one of the MIME experts
posts here regularly) but didn't make much progress. In the perfect
world the .EXE's and other binary files could be sent over the web
uncorrupted to/from a real PDP-10 via proper MIME encoding.
Tim.
Rich Alderson
> Yes, I see the index using the browser link from my windows desktop.
> But I wanted to move a 36bit file(eg EXE) from there to twenex. I
> telneted to twenex and ftped from there to pdp-10.trailing-edge.com,
> but got no further. How else do I get a binary file from the archive
> to a 10/20?
Get the tape image. I've written some utilities to deal with tape images on
the PDPplanet Toad-1; if you're interested, let me know.
--
Rich Alderson | /"\ ASCII ribbon |
ne...@alderson.users.panix.com | \ / campaign against |
"You get what anybody gets. You get a lifetime." | x HTML mail and |
--Death, of the Endless | / \ postings |
System 1022 Guy
the net lose the functionallity or did we lose the FTP software, or
maybe it was the bit serial 1822 hardware interface to the ARPANET?
Anyway: I wanted to look at those 1022 images on a decus tape
mentioned above.
BTW: I've heard CCA is considering my request for 1022. I'll let you
know how it goes.
Rich Alderson
> I thought I remember FTPing 36 bit files when I was at BBN ~1980. Did
> the net lose the functionallity or did we lose the FTP software, or
> maybe it was the bit serial 1822 hardware interface to the ARPANET?
The net has indeed lost a lot of the ability to FTP 36-bit files: The
8-bitness of most of the world has in conjunction with the general demise of
transfer protocols other than HTTP led us to this point. You can still do
36-bit transfers from and to Tops-20 systems: On U*x systems, the command is
"tenex", and on Windows "type tenex".
> Anyway: I wanted to look at those 1022 images on a decus tape
> mentioned above.
Oh. On the PDPplanet Toad-1 system, look in PS:<DECUS.LIB10.0126> and you'll
find them. I don't remember whether Tim got the DECUS tapes from me while I
was at XKL, or from someone else.
> BTW: I've heard CCA is considering my request for 1022. I'll let you
> know how it goes.
That would be cool. Wish it had happened 10 years ago when a customer needed
it.
sho...@trailing-edge.com
> "System 1022 Guy" <qqdc...@gmail.com> writes:
>
> > I thought I remember FTPing 36 bit files when I was at BBN ~1980. Did
> > the net lose the functionallity or did we lose the FTP software, or
> > maybe it was the bit serial 1822 hardware interface to the ARPANET?
>
> The net has indeed lost a lot of the ability to FTP 36-bit files: The
> 8-bitness of most of the world has in conjunction with the general demise of
> transfer protocols other than HTTP led us to this point. You can still do
> 36-bit transfers from and to Tops-20 systems: On U*x systems, the command is
> "tenex", and on Windows "type tenex".
>
> > Anyway: I wanted to look at those 1022 images on a decus tape
> > mentioned above.
>
> Oh. On the PDPplanet Toad-1 system, look in PS:<DECUS.LIB10.0126> and you'll
> find them. I don't remember whether Tim got the DECUS tapes from me while I
> was at XKL, or from someone else.
I already had the 10LIB tapes from somewhere else, but I'm 99% sure I
got the 20LIB tapes from you.
What's strange about 10-0126 is that it doesn't seem to have any DECUS
material at all but to be the ASCII-betical beginning of a dump of
somebody's SYS:. Oh well, some of the most interesting stuff ended up
on the DECUS tapes "by accident"!
Tim.
System 1022 Guy
System 1022 Guy
of CCA. I'll let you know if I hear anything.
Pat is right about the date expiration code. It is scattered. If you
get past the first one, you may get bitten later and not know it
(corrrupted data). If a customer reported the bad data as a bug, we
knew he was trying to hack an unauthorized copy. This strategy was
hotly debated. It stayed until one of our authorization patches didn't
unlock all of them for a customer. The possibility and liability of
corrupting a valid customer's data was too much. I think that the
"introduce bug" code was removed in later versions although I don't
remember when.
Hacking as suggested by BAH wasn't so simple. Lock code was dynamicly
built in AC's and was not FILDDT searchable. Also critical code
sections were checksummed and lock code would detect any changes,
breakpoints, tracing. The newer SET ADDR BREAK commands would simplfy
hacking though.
hea...@aracnet.com
> What's strange about 10-0126 is that it doesn't seem to have any DECUS
> material at all but to be the ASCII-betical beginning of a dump of
> somebody's SYS:. Oh well, some of the most interesting stuff ended up
> on the DECUS tapes "by accident"!
It's been a long time since I looked, but ISTR that tape is missing some
DECUS stuff that was supposed to be there and has that instead.
Zane