🚀 go-pugleaf

KXH...@psuvm.psu.edu (Ken Hornstein) writes:
>Here a question for all people who know much more than me: On IBM VM/CMS
>systems we have what are known as "spool" files ... aka reader files.
>And you can "spool" the virtual console.  What I wanted to know, where did
>"spool" come from?  I understand about reader being short for virtual card
>reader, PUN files being punch files, etc ... but "spool" mystifies me.  Anyone
>got any folklore on the significance of "spool"?

It's one of the more clever acronyms in the biz (as opposed to
"kermit"): Simultaneous Peripheral Operation OnLine. Back in the days
before multitasking, being able to have your operating system read
input jobs from the card reader AND print output to the printer AND
execute a job was a very big deal.

Back in the Old Days when we had our 7094 here (before my time, alas),
the operating procedure was to gather card decks for an hour, then read
them into an offline card reader, which would just copy the 80-column
records to tape. The tape was then taken across the street to where the
7094 CPU was located, and it would

	while not end-of-tape
	   read in a job
	   run the job
	   print the output to another tape

The output tape was then brought back across the street where another
offline device, a printer, would read the 133-column records off the
tape and print them on an impact printer. We had a great big table
which was our conference room table in my time which back then was the
"bursting table"--where operators would separate the program outputs
from each other and disseminate them, rubberbanded with the input card
deck, to the users.

We have it so lucky these days.

_____
Charley Kline, University of Illinois Computing Services
c-k...@uiuc.edu

In article <90080.235...@psuvm.psu.edu> KXH...@psuvm.psu.edu (Ken Hornstein) writes:
>Here a question for all people who know much more than me: On IBM VM/CMS
>systems we have what are known as "spool" files ... aka reader files.
>And you can "spool" the virtual console.  What I wanted to know, where did
>"spool" come from?  I understand about reader being short for virtual card
>reader, PUN files being punch files, etc ... but "spool" mystifies me.  Anyone
>got any folklore on the significance of "spool"?

"Spool" stands for "Simultaneous Peripheral Operation On-Line."  I've heard
it "Off-Line" as well, but the book I just looked it up in (Tannenbaum's
_Operating Systems_) has "On-Line."  I suppose this'll cause another
mega-discussion about which it actually is...

~mark

>+----------------------------------------------------------------------+
>| Ken Hornstein (kxh105@psuvm)  | "When in doubt, I whip it out ...".  |
>+-------------------------------+                        - Ted Nugent  |
>| Sorry, no cheap ASCII         +--------------------------------------+
>| graphics here ....            | Disclaimer: It was Fiend!!           |
>+----------------------------------------------------------------------+


      o o     o   o             o o   o     o        o   o     o o          
              o                       o	o   o  	   o       o   	 o     	     
  o   o     o     o         o     o     o   o      o   	 o     o       	   
Mark Shoulson:  shou...@cunixf.cc.columbia.edu   shoulson@cunixc.bitnet

>From article <90080.235...@psuvm.psu.edu>,
by KXH...@psuvm.psu.edu (Ken Hornstein):
>
> ...  What I wanted to know, where did "spool" come from?
> 
I've had the privilege of seeing a (second generation) spooler that used
real paper tape, but no spools.

It was in a quonset hut south of Urbana Illinois, and the hut smelled of
rotten soybeans, having been used to store beans before a computer
landed in it (high protien beans rot the same way meat does, and it was
very hot when I visited).

The guy who owned the hut liked to buy old computers, and he'd just gotten
a used vacuum tube machine from the FAA for a dollar and trucking costs.
He wanted it so he could hack around with a transistorized I/O processor
that had been added later, but he got the entire system, including drum
impact printers with specialized weather reporting symbols and all of the
communications interfaces.

All incoming communications lines were spooled so that if a message
arrived while the machine was doing something else, it wouldn't be lost.
Spoolers operated as follows:  The incoming line ran to a paper-tape
punch, the output of the punch ran into a bin, and from the bin, the
tape ran through a reader.  When the "spool file" was empty, there was
no tape in the bin and the tension of the tape between the punch and the
reader closed a microswitch to stop the reader.  To guarantee that a
message was fully received, you had to make sure it was followed by
enough nulls to span the gap between punch and reader.

The name for the thing came from the spools of paper tape involved, and
I gather that the first generation of spoolers stored the punched tape
on real (or reel?) spools between the punch and reader.  Some bright
person later found that letting the tape fall into a bin that was exactly
the same width as the tape would allow it to be pulled out without risk
of tangles, so physical spools weren't used.

By the way, the same man also had a working CDC 1620 (if memory serves
me right) in his basement.  It was the machine that once ran the Plato III
system, and he said he only used it in the winter.

					Doug Jones
					jo...@herky.cs.uiowa.edu

In article <90080.235...@psuvm.psu.edu> KXH...@psuvm.psu.edu (Ken Hornstein) writes:
>Anyone got any folklore on the significance of "spool"?

spool = "simultaneous peripheral operation on line"

It meant that the computer could drive the card readers, card punches,
and line printers simultaneously with the compute load.  Formerly, the
peripherals were attached to small computers, such as an IBM 1401, which
transferred card or line images to/from magnetic tape.  The tape was
hand-carried to/from the "mainframe" which then didn't burden itself with
running the slow peripherals while it had more important computing to do.

----------------------------------------------------------------------

Dave Dodson		                             dod...@convex.COM
Convex Computer Corporation      Richardson, Texas      (214) 497-4234

People have been describing the "spooling" process of
>    card reader --> mag tape
>    magtape --> computer --> magtape
>    magtape --> card punch or printer

Interestingly, I seem to have discovered this independently, about 30
years after the fact.  :-)

I had an undergraduate assembler language class that used an IBM 360.
(Yes, a real 360, and this was the early 80's!)  Rather than punch
cards on one side of campus, take them diagonally across campus to the
room with the 360, then go back for each iteration, I managed to
figure out enough JCL to transfer control to the tape drive and kept
those two cards with me.  I then made up my "decks", JCL and all, on a
PDP-11 running UNIX and used "dd" to write them onto tape.  The end of
the "decks" would contain JCL to transfer control back to the card
reader.  Luckily, the lowest density of the PDP-11 tape drive (800
bpi) matched the highest density of the 360 tape drive (which went as
low at 200-odd bpi).

The PDP-11 was only half-way across campus, so I didn't have to walk
too far, and could use a screen editor on my programs....  I didn't go
as far as to write the output back onto another tape to print on a
nicer printer; I just used the old 360 chain-drive....
--
John Owens		jo...@jetson.UPMA.MD.US		uunet!jetson!john
+1 301 249 6000		john%jet...@uunet.uu.net

Reply-To: zb...@umd5.umd.edu (Ben Cranston)
Distribution:
Organization: University of Maryland, College Park

KXH...@psuvm.psu.edu (Ken Hornstein) writes:

: Where did [the term] "spool" come from?

In article <1990Mar22....@ux1.cso.uiuc.edu> kl...@ux1.cso.uiuc.edu
(Charley Kline) writes:

: One of the more clever acronyms: Simultaneous Peripheral Operation OnLine.
: Back in the Old Days when we had our 7094 here (before my time, alas),
: ...gather card decks for an hour, then read the 80-column records to tape.
: The tape was then taken to the 7094 CPU... A printer would read the
: 133-column records off the tape and print them on an impact printer.

Charley, you forgot the most interesting bit of the folklore.  The
"customary and usual" 7094 configuration had the input tape on tape drive
number 5 and the output tape on tape drive number 6.  These numbers might
be familiar to those who have worked with Fortran, and coded:

       READ (5,100,END=15O) MUMBLE,FROTZ
       WRITE (6,110) SMUTZ,FROZZL

(Of course, Fortran is traditionally written in upper case.)  We still
have a bursting table, though since we don't do cards anymore we don't
wrap a rubber band around the printout...

On our 7094 and the Univac 1100s that replaced it used unit 1 for the punch
output unit, and 0 and 30 as "reread units" that would allow you to
reread the last card read (usually with a different format statement).
I don't think these numbers were quite as well standardized as 5 and 6.

-- 

"It's all about Power, it's all about Control
 All the rest is lies for the credulous"
-- Man-in-the-street interview in Romania one week after Ceaucescu execution.

In article <90080.235...@psuvm.psu.edu> KXH...@psuvm.psu.edu (Ken Hornstein) writes:
>Anyone
>got any folklore on the significance of "spool"?

I understand that spool is some acronym, but I forget what it stands for. I
would guess that what ever it is it's an acronym after the fact.
Unit record equipment (card readers, punches) and printers were, even back 
then, much slower than the computers. So in order to get maximum utilization
out of that expensive computer, one wanted to access this equipment in the
"background". This was done with intermediary mag tape units.  So:
    card reader --> mag tape
    magtape --> computer --> magtape
    magtape --> card punch or printer

This process was called "spooling" because the data would be spooled on the
magnetic tape.

FORTRAN I/O should be a little more obvious after this discussion -- you read
and write to tape units representing the card reader/punch (now "console")
and printer.

Tom Almy
to...@tekgvs.labs.tek.com
Standard Disclaimers Apply

Trivia eh?
Well, then you know what HASP is!

Hey I have time to read news!.... my supercomputer is fast, that's why I'm
waiting.

Another gross generalization from

--eugene miya, NASA Ames Research Center, eug...@aurora.arc.nasa.gov
  resident cynic at the Rock of Ages Home for Retired Hackers:

  "You trust the `reply' command with all those different mailers out there?"
  "If my mail does not reach you, please accept my apology."
  {ncar,decwrl,hplabs,uunet}!ames!eugene
  This signature will self destruct in 9 days.
  [If it ain't source, it ain't software -- D. Tweten]

In article <6571@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes:
>Trivia eh?
>Well, then you know what HASP is!

Houston Automatic Spooling and Priority system.  Developed by NASA for the
360 model 91 which put us on the moon, I believe.

gkn
San Diego Supercomputer Center

In article <12...@ucsd.Edu>, g...@ucsd.Edu (Gerard K. Newman) says:
>
>In article <6571@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes:
>>Trivia eh?
>>Well, then you know what HASP is!
>
>Houston Automatic Spooling and Priority system.  Developed by NASA for the

Is this the same as $HASP messages issued by JES2?!?  Just wondering.
Does anybody out there know where the $HASP message from MVS comes from?
Are the an acronym for something?

Yes, that's right, I am Ack!, Bill the Cat's fur ball...   NOT the JOKER!

 _   /|              ESH101@PSUVM                Wot?  These PSU's ideas?  Ha!
 \ o.O`              ESH...@PSUVM.PSU.EDU        They don't even know how to
 =(___)=  -- Ack!    ESH101%PSUVM@PSUVAX1.UUCP   *THINK*!!  If I would drink, I
    U                HVO...@MATH.PSU.EDU         would be normal.  We couldn't
                     HVOZDA%MATH.PSU.EDU@PSUVM   have that, now could we??

In article <12...@ucsd.Edu> g...@ucsd.edu (Gerard K. Newman) writes:
>In article <6571@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes:
>>Trivia eh?
>>Well, then you know what HASP is!
>
>Houston Automatic Spooling and Priority system.  Developed by NASA for the
>360 model 91 which put us on the moon, I believe.

It ran on more than the -91. I ran it on a -50 in the late '60s. The name
is right (or close), although at the time the joke was that it stood for
Half-ASP (ASP being ancestor of what is now JES3, as HASP was the ancestor
of JES2). As I recall ASP stood for something like Attached Spool Processor...
anybody remember for sure?
-- 
UUCP:  litt...@amdahl.amdahl.com
  or:  {sun,decwrl,hplabs,pyramid,ames,uunet}!amdahl!littauer
DDD:   (408) 737-5056
USPS:  Amdahl Corp.  M/S 278,  1250 E. Arques Av,  Sunnyvale, CA 94086

I'll tell you when I'm giving you the party line. The rest of the time
it's my very own ravings (accept no substitutes).

dod...@convex.com (Dave Dodson) writes:

>In article <90080.235...@psuvm.psu.edu> KXH...@psuvm.psu.edu (Ken Hornstein) writes:
>>Anyone got any folklore on the significance of "spool"?

>spool = "simultaneous peripheral operation on line"

Ok, that's right.

>It meant that the computer could drive the card readers, card punches,
>and line printers simultaneously with the compute load.  Formerly, the
>peripherals were attached to small computers, such as an IBM 1401, which
>transferred card or line images to/from magnetic tape.  The tape was
>hand-carried to/from the "mainframe" which then didn't burden itself with
>running the slow peripherals while it had more important computing to do.

"Simultaneous" meant in addition that more than one user(-programm) could
use the card reader, card punches and line printers at the same time,
because the records where "SPOOL"-ed.

Werner
-- 
Werner Icking
Gesellschaft fuer Mathematik und Datenverarbeitung mbH (GMD)
Schloss Birlinghoven		    email: icking@gmdzi.uucp
D-5205 Sankt Augustin 1, FRG 	    phone: (+49 2241) 14-2443

In article <63...@umd5.umd.edu>, zb...@umd5.umd.edu (Ben Cranston) writes:

> Charley, you forgot the most interesting bit of the folklore.  The
> "customary and usual" 7094 configuration had the input tape on tape drive
> number 5 and the output tape on tape drive number 6.  These numbers might
> be familiar to those who have worked with Fortran, and coded:
>
>        READ (5,100,END=15O) MUMBLE,FROTZ
>        WRITE (6,110) SMUTZ,FROZZL

You are too modern - that is Fortran IV! In Fortran II, it was

         READ INPUT TAPE 5,100,MUMBLE,FROTZ

and there was no fancy feature like END=150.

On IBM's VM/CMS one method of file transfer between virtual machines is
connecting your virtual punch to the other guy's virtual reader:
 SPOOL PUN TO FRED RDR           ( approximately)

I can visualize that, a mechical connection between a card punch and reader,
but you can also connect your printer to a reader! The mind boggles.
--
Ian W Moor
  ARPA: i...@doc.ic.ac.uk
  JANET: iwm@uk.ac.ic.doc
           
 Department of Computing,         There are two kinds of fool,
 Imperial College.                one who says, this is good because it is old,
 180 Queensgate                   the other, this new so it's better
 London SW7 UK.                    W.R. Inge

Going for old AND obscure, here's one (and I'll include the answer
since no one will care):

   PEST.


That's the Pittsburgh Executive System Tape for the IBM 7070/72/74
series.  I guess it was a lot faster and simpler than the IBSYS or
whatever official OS was supposed to be used with the beast.

	/JBL
=
Nets: le...@bbn.com  |  "There were sweetheart roses on Yancey Wilmerding's
 or {...}!bbn!levin  |  bureau that morning.  Wide-eyed and distraught, she
POTS: (617)873-3463  |  stood with all her faculties rooted to the floor."

In article <21ps02X...@amdahl.uts.amdahl.com>, litt...@uts.amdahl.com (Tom Littauer) writes:
> is right (or close), although at the time the joke was that it stood for
> Half-ASP (ASP being ancestor of what is now JES3, as HASP was the ancestor
> of JES2). As I recall ASP stood for something like Attached Spool Processor...

ASP was ``Attached Support Processor''.  A smaller machine would serve
several larger machines (and possibly itself), handling all unit record
I/O via channel-to-channel adapters.  From the existence of ASP and HASP,
one can easily -- and correctly -- conclude that the standard reader and
printer spoolers (and the batch job scheduler) were quite inadequate.

ASP did a bit more, though; it also acted as the operator for the
CPUs it controlled.  This was most important for its tape setup
capability, but it also was necessary for its CTC support -- the
operating system thought the channel adapters were tape drives.  Of
course, sometimes ASP would get confused; I remember any number
of occasions when the operator would be told to mount a tape on
a piece of electronics somewhere...

lme...@eds.ericsson.se (Bengt G{llmo, L M Ericsson, Stockholm, Sweden) writes:
>In article <63...@umd5.umd.edu>, zb...@umd5.umd.edu (Ben Cranston) writes:
>> Charley, you forgot the most interesting bit of the folklore.  The
>>
>>        READ (5,100,END=15O) MUMBLE,FROTZ
>>        WRITE (6,110) SMUTZ,FROZZL

>You are too modern - that is Fortran IV! In Fortran II, it was
>         READ INPUT TAPE 5,100,MUMBLE,FROTZ

Actually I believe the Lawrence Livermore dialect of fortran, called
LRLTRAN, still has that old feature, along with its output equivalent,
WRITE OUTPUT TAPE, in the form of RIT and WOT statements. Took me a 
while to figure out what they stood for. Me, I never had to use them;
the only LRLTRAN I touched was the kernel for the Cray Timesharing System.

Yeech.


_____
Charley Kline, University of Illinois Computing Services
c-k...@uiuc.edu

"Everyone out of the genetic pool!"

In article <IWM.90Ma...@gould.doc.ic.ac.uk>, i...@doc.ic.ac.uk (Ian Moor)
says:
>
>On IBM's VM/CMS one method of file transfer between virtual machines is
>connecting your virtual punch to the other guy's virtual reader:
> SPOOL PUN TO FRED RDR           ( approximately)

TRANSFER is a great command.  On XA systems (at least, I don't think this
will work on HPO) you can create a file, transfer it to someone who is not
logged on, and THEN transfer it to a friend.  But on his screen it will show
up as him getting a file from the person you transferred it to first!  And
then to add to the confusion, you can purge it from his reader without him
getting a message ... REALLY messes people up when you do this to them 30-40
times.  Although, I *DID* lose my account the last time I did this :-)

Ken

In article <100...@convex.convex.com> dod...@convex.com (Dave Dodson) writes:
>In article <90080.235...@psuvm.psu.edu> KXH...@psuvm.psu.edu (Ken Hornstein) writes:
>>Anyone got any folklore on the significance of "spool"?
>
>spool = "simultaneous peripheral operation on line"
>
>It meant that the computer could drive the card readers, card punches,
>and line printers simultaneously with the compute load.  Formerly, the
>peripherals were attached to small computers, such as an IBM 1401, which
>transferred card or line images to/from magnetic tape.  The tape was
>hand-carried to/from the "mainframe" which then didn't burden itself with
>running the slow peripherals while it had more important computing to do.

It wasn't the burdening with running slow peripherals. It was the problem
of waiting for slow operators :-). The systems of which we speak, say the
IBM 7094, or the IBM 7090, were single (count it, 1) job at a time systems.
After a job finished, one could wait for an operator to set up another job
get the cards, read 'em in, find the output,..., or, one could set up a
stream of jobs (now where did job stream come from) on magtape using a cheap
computer (relatively) so that the 709x could just read the next job from tape
with negligible delay. Multi-processing in action. Then, think how slick it
would be if you hooked a cheaper box to an expensive box directly, hence the
7044/7094 DCS where the 7044 handled the unit record I/O. My favorite was of
course the IBM 5081 unit record device.

When folks figured that computers were capable of concurrent operation and
could be interrupted from one task to perform another, then it was reasonable
to have one box manage its own I/O and its own compute load.

Enough of these grey boxes that blink blue, green, red, and white. I wanna
talk about blue boxes and salmon-pink boxes and little boxes covered with
wood-grain paper and how the little ticky-tacky boxes of today have no good
flashing lights at all :-(.
--
Mike Murphy  Sceard Systems, Inc.  544 South Pacific St. San Marcos, CA  92069
m...@Sceard.COM        {hp-sdd,nosc,ucsd,uunet}!sceard!mrm      +1 619 471 0655

In article <21ps02X...@amdahl.uts.amdahl.com> litt...@amdahl.uts.amdahl.com (Tom Littauer) writes:
>In article <12...@ucsd.Edu> g...@ucsd.edu (Gerard K. Newman) writes:
>>In article <6571@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes:
>>>Trivia eh?
>>>Well, then you know what HASP is!
>>
>>Houston Automatic Spooling and Priority system.  Developed by NASA for the
>>360 model 91 which put us on the moon, I believe.
>
>It ran on more than the -91. I ran it on a -50 in the late '60s. The name
>is right (or close), although at the time the joke was that it stood for
>Half-ASP (ASP being ancestor of what is now JES3, as HASP was the ancestor
>of JES2). As I recall ASP stood for something like Attached Spool Processor...
>anybody remember for sure?
>-- 

We had ASP at Princeton in the late '60s. The acronym was
"Attach(ed) Support Processor" (spooling was a side-effect).
The attached processor functioned as a programmed operator
for the main computer; doing things like adjusting priorities
of jobs, submitting jobs in the "best" order and so on,
and generally overcoming deficiencies in "Big OZ" (MVT).

HASP was routinely used at many sites; Princeton was one
of very few sites to have ASP. ASP was installed more or
less in conjunction with our acquiring the 360/91. ASP
usurped control of MVT through a hack using a user
appendage to a device driver routine.

The 360/91 was one of the first computers with caching
and pipelining (to my knowledge). Any other Princeton
folks out there?

Michael Stimac

In article <3384@tymix.UUCP>, romeo@hobbes.uucp (Michael Stimac) writes:
> The 360/91 was one of the first computers with caching
> and pipelining (to my knowledge). Any other Princeton
> folks out there?
> 
> Michael Stimac

Actually, I don't believe that the /91 had a cache; I think that the
360/85 was the first model IBM introduced with one.  It definitely
had pipelining (*love* those imprecise interrupts), but it was hardly
the first machine to do that; that honor, I belive, goes to the IBM 7030,
a.k.a. Stretch.  And I'm fairly certain that caching goes back a few
years before the 360/91; wasn't it Wilkes's idea?

In article <54...@bbn.COM> le...@BBN.COM (Joel B Levin) writes:
>   PEST.
>That's the Pittsburgh Executive System Tape for the IBM 7070/72/74
>series.  I guess it was a lot faster and simpler than the IBSYS or
>whatever official OS was supposed to be used with the beast.

Wasn't IBSYS originally written by some folks associated with SHARE?  The
folklore I heard was that the IBJOB operation system was such a piece of trash
that the users of the system rose up and wrote their own O/S.  Any old 70xx
hackers out there who remember?

By the way, speaking of folklore and names of operating systems, what was the
name of the operating system that the National Security Agency used on its
Crays (and maybe other machines??) before they converted to UNICOS?  (Sort of
a trick question)
--
Steve Lamont, sciViGuy	(919) 248-1120		EMail:	s...@ncsc.org
NCSC, Box 12732, Research Triangle Park, NC 27709
"...though you may have the falcon yet we certainly have you."
				Dashiell Hammett, _The Maltese Falcon_

lme...@eds.ericsson.se (Bengt G{llmo, L M Ericsson, Stockholm, Sweden) writes:

>In article <63...@umd5.umd.edu>, zb...@umd5.umd.edu (Ben Cranston) writes:
[...]
>You are too modern - that is Fortran IV! In Fortran II, it was
 
>         READ INPUT TAPE 5,100,MUMBLE,FROTZ
 
>and there was no fancy feature like END=150.
 
But once I mistyped
 
          WRITE INPUT TAPE 5, ...    or may be
          WRITE OUTPUT TAPE 5, ...
 
and it was not rejected neither by the compiler nor by the system (?) or by
the tape unit. I think the tape units did'nt have write protection rings; but
I'm not sure. 

The operation succeeded, the input tape was overwritten i.e. all the job
SPOOLed (:-() after mine were not processed.

I was a student at that time and (therefore) the system administrator wanted
to sever my account. But I could convince him that the system really made the
error and not me. (~1966)
-- 
Werner Icking
Gesellschaft fuer Mathematik und Datenverarbeitung mbH (GMD)
Schloss Birlinghoven		    email: icking@gmdzi.uucp
D-5205 Sankt Augustin 1, FRG 	    phone: (+49 2241) 14-2443

In article <18...@speedy.mcnc.org> s...@duck.ncsc.org (Steve Lamont) writes:
>What was the name of the operating system that the National Security
>Agency used on its Crays (and maybe other machines??) before they
>converted to UNICOS?

Surely NSOS (No Such Operating System). :)

In article <3384@tymix.UUCP> romeo@hobbes.UUCP (Michael Stimac) writes:
>The 360/91 was one of the first computers with caching
>and pipelining (to my knowledge)...

The 91 had no cache, but it was pipelined to an extent seldom seen before
or since.  The 85, a rather simpler machine with a cache, performed nearly
as well and sold a lot better, a lesson that was not lost on IBM...
-- 
"Apparently IBM is not IBM    |     Henry Spencer at U of Toronto Zoology
compatible." -Andy Tanenbaum  | uunet!attcan!utzoo!henry he...@zoo.toronto.edu

In article <18...@speedy.mcnc.org> s...@duck.ncsc.org (Steve Lamont) writes:
>
>By the way, speaking of folklore and names of operating systems, what was the
>name of the operating system that the National Security Agency used on its
>Crays (and maybe other machines??) before they converted to UNICOS?  (Sort of
>a trick question)

Why, Folklore, of course.  I think there were others.

gkn

ps - DEC hasn't apologized for the Lynx;  I still have 3.  Want 'em ;-)  Make
you a deal -- I'll throw in (up?) two DRB32-C VAX Supercomputer gateway boards...

In article <12...@ucsd.Edu>, g...@ucsd.Edu (Gerard K. Newman) says:
>
>In article <6571@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes:
>>Trivia eh?
>>Well, then you know what HASP is!
>
>Houston Automatic Spooling and Priority system.  Developed by NASA for the

Harvard, wasn't it ?
--
Henry Troup - BNR owns but does not share my opinions
..uunet!bnrgate!hwt%bwdlh490 or  H...@BNR.CA

This is only a rumor as far as I'm concerned; for all I know it is
still classified.  But I think the CRAYs were running AMOK.

In article <1990Mar26.165323.27009@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes:
>
>The 91 had no cache, but it was pipelined to an extent seldom seen before
>or since.  The 85, a rather simpler machine with a cache, performed nearly
>as well and sold a lot better, a lesson that was not lost on IBM...

There were other issues, too. The 91 was heavily oriented to floating point,
unruly in its handling of interrupts, and really, really, expensive.
The 85 was more in line with the slower "compatible" series.

[The interrupt unruliness was caused by the pipelining -- an interrupt could be
three or four instructions too late on FP errors.]

The first IBM supercomputer, the Stretch, was sort of a prototype for the Big
Blue top-of-the-line-real-expensive-doesn't-work-too-well-only-a-dozen-made
series.  See Bell and Newell for a description of this beast, down to the
weird gates it used. It was pipelined.

The Stretch was the first computer I know of to have an optional tunnel-diode
memory.  It was installed on one and only one machine, and I know of no further
use of the technology.

				phil

In article <2132@gmdzi.UUCP>, icking@gmdzi.UUCP (Werner Icking) writes:

> But once I mistyped
>  
>           WRITE INPUT TAPE 5, ...    or may be
>           WRITE OUTPUT TAPE 5, ...
>  
> and it was not rejected neither by the compiler nor by the system (?) or by
> the tape unit. I think the tape units did'nt have write protection rings; but
> I'm not sure. 
> 
> The operation succeeded, the input tape was overwritten i.e. all the job
> SPOOLed (:-() after mine were not processed.
> 
> I was a student at that time and (therefore) the system administrator wanted
> to sever my account. But I could convince him that the system really made the
> error and not me. (~1966)

The tapes had file protection rings all right, and it was careless of some
operator not to pull it out on that input tape. But the compiler couldn't 
care less. It was perfectly legal from a language point of view to call the 
output tape 5 and write to it. 5 for input and 6 for output was just a
convention.

In article <10...@zorch.SF-Bay.ORG>, ph...@zorch.SF-Bay.ORG (Phil Gustafson) writes:
> [The interrupt unruliness was caused by the pipelining -- an interrupt could be
> three or four instructions too late on FP errors.]

``Three or four''?  You must have hit a best-case...  I recall one
program I wrote (for Columbia University's 360/91) where I got a
zero-divide interrupt 14 instructions later.  I'm not exagerating;
I counted.  There were only two divide instructions in the entire
program, and the other was by a constant that the SYSUDUMP (remember
those?) showed to be intact.

There were other oddities in that machine.  For example, it had
16-way interleaved memory -- which made some operations on 257x257
matrices faster than 256x256, since the latter tended to beat on
the same memory box all the time.

But it did have *great* blinkenlites.  My SPARCstation is much faster,
and has 8 times the memory the old clunker had, and it's all mine --
but the only light is a stupid little ``power on'' thingy.

In article <12...@ulysses.att.com>, s...@ulysses.att.com (Steven M. Bellovin) writes:
> There were other oddities in that machine.  For example, it had
> 16-way interleaved memory -- which made some operations on 257x257
> matrices faster than 256x256, since the latter tended to beat on
> the same memory box all the time.

The Burroughs Scientific Processor (BSP) had some interesting
characteristics like that.  BSP had a scalar unit and a parallel
unit (I don't remember the acronyms any more).  The parallel unit
was 16 lockstep processors.  To avoid problems with the processors
waiting for the memory, they used 17 way interleaved memory.  Thus,
accesses would NEVER hit the same memory bank unless you used multiples
of 17, which doesn't happen naturally in too many parallel programs :-)

I was told that Burroughs wanted to patent the use of (2**n)+1 way
memory for all "n".

For those who never heard of it, the BSP was a late-70s/early-80s
attempt by Burroughs to compete with Cray.  It used a B6700 or B7700
as a front end, and was water cooled.  I think a few were sold...
-- 
Jeremy Epstein
TRW Systems Division
703-876-4202
j...@virtech.uu.net

In article <1990Mar30.144209.4749@virtech.uucp>, jje@virtech.uucp
(Jeremy J. Epstein) writes:
|> 
|> The Burroughs Scientific Processor (BSP) had some interesting
|> characteristics like that.  BSP had a scalar unit and a parallel
|> unit (I don't remember the acronyms any more).  The parallel unit
|> was 16 lockstep processors.  To avoid problems with the processors
|> waiting for the memory, they used 17 way interleaved memory.  Thus,
|> accesses would NEVER hit the same memory bank unless you used multiples
|> of 17, which doesn't happen naturally in too many parallel programs :-)
|> 

The interesting thing about the 17-way interleaving is that it required
a modulo-17 calculation in the addressing hardware.  Normal power-of-two
interleaving schemes make for much easier bank calculation, since you
essentially just take a few bits out of word to address the bank, but in
the BSP it was necessary to do arithmetic instead.  They had a special
piece of hardware to do this divide so that it would be fast enough.

The BSP could also do two-dimensional vector operations (complete with
a skip between columns).

I have heard that they built one BSP, and had one order in hand.
The BSP project was scrapped because they were using a hard-to-get chip
technology, and their supplier (which was within Burroughs)
discontinued the manufacture of their chips.

In article <12...@ulysses.att.com> s...@ulysses.att.com (Steven M. Bellovin) writes:
>In article <10...@zorch.SF-Bay.ORG>, ph...@zorch.SF-Bay.ORG (Phil Gustafson) writes:
>> [The interrupt unruliness was caused by the pipelining -- an interrupt could be
>> three or four instructions too late on FP errors.]
>
>[eleven lines about program generated errors]
>
>But it did have *great* blinkenlites.  My SPARCstation is much faster,
>and has 8 times the memory the old clunker had, and it's all mine --
>but the only light is a stupid little ``power on'' thingy.

I have fond memories of machines with blinking lights.  The High Energy
Physics people at Purdue (circa. 1977) had a 360 model 40 which had
a habit of hanging - you'd look at the lights after the card reader had
stopped for 3 or 4 seconds to see if it was dead or just slow.

My favorite blinken light machine was a NanoData QM-1 which was for
microcoding, and had an LED on every bit of every register and every
bit of the 144-bit wide instruction register, over 2000 LEDs in all.
The best use for this machine (besides room heater, it used 74Sxx chips)
was a program called TSQ, short for Times Square, which put a running
message in the lights.  Pretty cool, huh?

I wish computer designers would recognize the importance of blinking
lights and put more of them on machines.  They'd be much more informative
than the silly bit-mapped display on my Mac. :-)

James Thiele

In article <12...@csli.Stanford.EDU> po...@csli.stanford.edu (Bill Poser) writes:
>In article <18...@speedy.mcnc.org> s...@duck.ncsc.org (Steve Lamont) writes:
>>What was the name of the operating system that the National Security
>>Agency used on its Crays (and maybe other machines??) before they
>>converted to UNICOS?
>
>Surely NSOS (No Such Operating System). :)

Actually, it's Folklore.

I guess that this must be the appropriate newsgroup to discuss it.

If my sources are right, it was written in a language called IMP,
and utilities names were full of puns on mythical creatures.
-- 

Mark Nelson                   ...!uunet!udel!nelson or nel...@udel.edu
This function is occasionally useful as an argument to other functions
that require functions as arguments. -- Guy Steele, Jr.

nel...@ee.udel.edu (Mark Nelson) wrote:
>> s...@duck.ncsc.org (Steve Lamont) writes:
>>> What was the name of the operating system that the National Security
>>> Agency used on its Crays (and maybe other machines??) before they
>>> converted to UNICOS?

> Actually, it's Folklore.

> If my sources are right, it was written in a language called IMP,
> and utilities names were full of puns on mythical creatures.

Is that any relation to the Edinburgh IMP?  That was a systems programming
language used by the Edinburgh University Computing Centre to implement
their EMAS (Edinburgh Multi-Access System) OS on ICL mainframes (1900 and
2900 series; I think the 3900 caching and pipelining schemes make EMAS
unusable).  I believe the parser was written by an AI hacker using
natural-language parsing techniques; you could have embedded spaces in
identifier names.

I've never even seen any IMP source, so all this is hearsay.

-- 
--  Jack Campin   Computing Science Department, Glasgow University, 17 Lilybank
Gardens, Glasgow G12 8QQ, Scotland   041 339 8855 x6044 work  041 556 1878 home
JANET: ja...@cs.glasgow.ac.uk    BANG!net: via mcvax and ukc   FAX: 041 330 4913
INTERNET: via nsfnet-relay.ac.uk   BITNET: via UKACRL   UUCP: jack@glasgow.uucp

In article <49...@vanuata.cs.glasgow.ac.uk> ja...@cs.glasgow.ac.uk (Jack Campin) writes:
>
>>> s...@duck.ncsc.org (Steve Lamont) writes:
>>>> What was the name of the operating system that the National Security
>>>> Agency used ....
>
>> If my sources are right, it was written in a language called IMP,
>> and utilities names were full of puns on mythical creatures.
>
>Is that any relation to the Edinburgh IMP?  That was a systems programming
>language used by the Edinburgh University Computing Centre to implement
>their EMAS (Edinburgh Multi-Access System) OS on ICL mainframes (1900 and
>2900 series; I think the 3900 caching and pipelining schemes make EMAS
>unusable).  I believe the parser was written by an AI hacker using
>natural-language parsing techniques; you could have embedded spaces in
>identifier names.
>  -- Jack Campin ??

Is IMP <<still>> around at Edinburgh ??  It was developed in the 60's
and was a descendant of Atlas Autocode (as used on the Manchester Atlas
machine). I used the language 20 years ago on various machines:  the
KDF9, English Electric System 4, and an IBM 360/50. I'm surprised that
it (and EMAS, which was very impressive for its time, IMHO) have
survived. And it's worthy of note that the implementation of EMAS in
IMP pre-dates UNIX implemented in C by a few years ..

-----------------------------
Simon Rosenthal:
Concurrent Computer Corporation
Westford, MA 01886
Internet: si...@westford.ccur.com
uucp:	{uunet,harvard}!masscomp!simon

In article <53891@microsoft.UUCP> jamesth@microsoft.UUCP (James THIELE) writes:
>My favorite blinken light machine was a NanoData QM-1 which was for
>microcoding, and had an LED on every bit of every register and every
>bit of the 144-bit wide instruction register, over 2000 LEDs in all.

I remember seeing one at the 1979 National Computer Conference in New York.
In its idle mode, the LEDs would spell out QM-1 in large letters.  I just
ran across a brochure for it yesterday, but unfortunately they had the
doors closed on the cabinet when they took the picture.

I *LIKE* blinkenlights.  On our old PDP-11/45, the Unix OS had a feature
whereby it would read the console switches, peek into that location in
kernel data space, then show the value in the console display register.
I think it was 2.9bsd that allowed this.

>I wish computer designers would recognize the importance of blinking
>lights and put more of them on machines.  They'd be much more informative
>than the silly bit-mapped display on my Mac. :-)

Ditto!

	---Bob.

--
Bob Hoffman, N3CVL       {allegra, bellcore, cadre, idis, psuvax1}!pitt!hoffman
Pitt Computer Science    hof...@cs.pitt.edu

> In article <53891@microsoft.UUCP> jamesth@microsoft.UUCP (James THIELE) writes:
>My favorite blinken light machine was a NanoData QM-1 which was for
>microcoding, and had an LED on every bit of every register and every
>bit of the 144-bit wide instruction register, over 2000 LEDs in all.

  I once worked on an HP 3000. This one had these 2 large panels with lots
of LED's on them on either side of the operator's CRT. I later met another
3000 that looked nothing like the one I saw; perhaps the one I used was a
prototype? [This was in the early 70's.]

	Terry Kennedy		Operations Manager, Academic Computing
	terry@spcvxa.bitnet	St. Peter's College, US
	te...@spcvxa.spc.edu	(201) 915-9381

si...@westford.ccur.com (Simon Rosenthal) wrote:
> ja...@cs.glasgow.ac.uk (Jack Campin) writes:

> Is IMP <<still>> around at Edinburgh??  It was developed in the 60's
> and was a descendant of Atlas Autocode (as used on the Manchester Atlas
> machine).  I used the language 20 years ago on various machines:  the
> KDF9, English Electric System 4, and an IBM 360/50.  I'm surprised that
> it (and EMAS, which was very impressive for its time, IMHO) have
> survived.  And it's worthy of note that the implementation of EMAS in
> IMP pre-dates UNIX implemented in C by a few years ..

Well, there is still a site emas-a.ed.ac.uk, which suggests that it still
runs EMAS, and if it runs EMAS there must be an IMP compiler to support it.

Some while ago, my email to a friend who uses our computing centre's ICL
mainframe (and has a desk 100 yards away) had to go through the EMAS node
in Edinburgh, that being the easiest way to get mail from a Unix-based site
into the twilight zone of VME; it may still go that way.

Come on, there must be *someone* at Edinburgh reading this!
-- 
--  Jack Campin   Computing Science Department, Glasgow University, 17 Lilybank
Gardens, Glasgow G12 8QQ, Scotland   041 339 8855 x6044 work  041 556 1878 home
JANET: ja...@cs.glasgow.ac.uk    BANG!net: via mcvax and ukc   FAX: 041 330 4913
INTERNET: via nsfnet-relay.ac.uk   BITNET: via UKACRL   UUCP: jack@glasgow.uucp

Does anyone have a copy of the Blinkenlights poster?
This was a pseudo-German funny thing that ended with the words

     "Relaxen und vatch das blinken lights"

or something to that effect.
--
--------------------------- Signature Version 1.2 ---------------------------
| be...@bu-pub.bu.edu                                  eng...@buacca.bu.edu |
|									    |
|        "Love is like oxygen.  You get too much, you get too high.         |
|          Not enough and you're gonna die.  Love gets you high."           |
-----------------------------------------------------------------------------

Yeah.. got a stack of them for when a machine goes bad, in fact. here goes.


		Achtensun!
Das machinen ist nodt fur gerfingerpoken un mittengrabben.
Ist ezy fur brekken das springenwerks. Blowen fusen mit loudsch popen
und spitzensparken.Ist nix geverken by dumkoffs! Das rubberneken
sighdtseeren und stupidisch goofoffers bast relaxen. Kipp hands in
pokkets, Vatch der blinkenlights, und das stoffa vill kum oudt.

				Das Manichmindt

Origin of the word "snarf".  I've only seen it in net postings.  Anyone know
where it comes from?  Thanx

          -- Ed Krauss

In article <1990Apr10.0...@agate.berkeley.edu> ph11...@violet.berkeley.edu () writes:
>Origin of the word "snarf".  I've only seen it in net postings.  Anyone know
>where it comes from?  Thanx
>
>          -- Ed Krauss


  What?  I've ALWAYS heard "snarf."  Usually in the context of "snarfing up
food."  "Snarf" is the sound one makes when one ingests a plate of food all
at once, and so on.
  Other usages are probably just a result of the word sounding neat.

-- Mark Wilkins

The only computer-related (this IS a.f.c) use of snarf I can
think of is in MIT's Incompatible Timesharing System (from
which BSD stole job control).  If you had a job that got
horribly sick, but had subjobs you wanted to rescue, you could
:SNARF them.  :SNARF would write a small program into the dead
job's address space to do the actual work, then start the job.

					-=EPS=-

In article <1990Apr10.0...@agate.berkeley.edu> ph11...@violet.berkeley.edu () writes:
>Origin of the word "snarf".  I've only seen it in net postings.  Anyone know
>where it comes from?  Thanx
>
>          -- Ed Krauss

>From _Structure_and_Interpretation_of_Computer_Programs_ (pseudo-underline),
by Harold Abelson and Gerald Jay Sussman with Julie Sussman (The MIT Press,
Cambridge, Mass: 1985), Pg.316, footnote 13:

"Snarf: "To grab, especially a large document or file for the purpose of 
 using it either with or without the owner's permission." Snarf down: "To
 snarf, sometimes with the connotation of absorbing, processing, or
 understanding." (These definitions were snarfed from Steele et al. 1983.)"

This is the textbook for my scheme class, and I found this while reading the
chapter on the meta-circular evaluator. (I thought it was quite humorous!)

-- 
+----------------------------------------------------+------------------------+
|"You only live once so take hold of the chance,     | prof...@wpi.wpi.edu   |
| Don't end up like others the same song and dance!" | profesor@wpi.bitnet    |
|   -"Motorbreath", by Metallica                     | Really: - Matt Cross - |

In article <1990Apr10.0...@agate.berkeley.edu> ph11...@violet.berkeley.edu () writes:
>Origin of the word "snarf".  I've only seen it in net postings.  Anyone know
>where it comes from?  Thanx
>          -- Ed Krauss

Local folklore says that it's a composite of 'snag' and 'scarf'
-- one snags a document from another machine and scarfs down the
information in it immediately. Say 'snag and scarf' enough times
and it collapses to 'snarf'.



-- 
David Boyes       |  "Where's the ka-boom? There's supposed to be an
dbo...@rice.edu   |  Earth-shattering ka-boom!...Heavens! Someone has
                  |  stolen the Illudium Q-38 Explosive Space Modulator!
"Delays, delays!" |  The Earth creature has *stolen* the Space Modulator!"

"Snarf" is indeed one of the words found in the delightful book "the
Hackers Dictionary", a rehash of the file HACKER.JARGON wich have been
circulated around the nets since Time Began. It's full of folklore,
highly recommended.

(Is it still around here somewhere? Maybe co-author Mark Crispin is
still with us on this channel and could give us a pointer? Or perhaps
he prefers that we go out and buy the book? :-)

Also, "snarf" is aviator jargon. Probably came from the time when you
rescued a downed fellow aviator up into your Tiger Moth by yanking (or
snarfing) him by his pilot's scarf.

-- 
Per Lindberg (The Mad Programmer)  !      __!__
Front Capital Systems              !  _____(_)_____  Ceci n'est pas une Piper
Linneg 5, 11447 Stockholm, Sweden  !     !  !  !

In article <10...@front.se> p...@front.se (Per Lindberg) writes:
>rkeley.edu>
>"Snarf" is indeed one of the words found in the delightful book "the
>Hackers Dictionary", a rehash of the file HACKER.JARGON wich have been
>circulated around the nets since Time Began. It's full of folklore,
>highly recommended.
>
>(Is it still around here somewhere? Maybe co-author Mark Crispin is
>still with us on this channel and could give us a pointer? Or perhaps
>he prefers that we go out and buy the book? :-)

The US edition of "The Hacker's Dictionary" by Steele, Woods, Finkel,
Crispin, Stallman, and Goodfellow, is ISBN 0-06-091082-8, published by
Harper & Row.  It has been out of print for several years.  There is
an effort underway to develop a revised edition and have MIT Press
publish it (and hopefully do better in marketing it!).

There are still foreign editions in print.  I have a copy of the
Swedish edition (called "Uppslagsbok for Computer-Freaker" or
something similar) and of the recently-published Japanese edition,
"HAKKAA Eiga Jiten".  I also heard of a German edition?

 _____     ____ ---+---   /-\   Mark Crispin           Atheist & Proud
 _|_|_  _|_ ||  ___|__   /  /   6158 Lariat Loop NE    R90/6 pilot
|_|_|_| /|\-++- |=====| /  /    Bainbridge Island, WA  "Gaijin! Gaijin!"
 --|--   | |||| |_____|   / \   USA  98110-2098        "Gaijin ha doko ka?"
  /|\    | |/\| _______  /   \  +1 (206) 842-2385      "Niichan ha gaijin."
 / | \   | |__| /     \ /     \ m...@CAC.Washington.EDU "Chigau. Gaijin ja nai.
kisha no kisha ga kisha de kisha-shita                  Omae ha gaijin darou."
sumomo mo momo, momo mo momo, momo ni mo iroiro aru    "Iie, boku ha nihonjin."
uraniwa ni wa niwa, niwa ni wa niwa niwatori ga iru    "Souka. Yappari gaijin!"

> The US edition of "The Hacker's Dictionary" by Steele, Woods, Finkel,
> Crispin, Stallman, and Goodfellow, is ISBN 0-06-091082-8, published by

	Does anyone know Finkel's first name.  We have a CS professor here
at Whoopie Tech called Finkel.  Is he the same ?

TIA

`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'`'
David Ritchie     | "If anyone doubts my veracity, I can only say I pity his 
da...@wpi.wpi.edu | lack of faith." - Baron Munchasen
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~