Conference 7.286::digital

    
    I can add my 2cents about the early 70s era at DEC.
    
    o Training was held in building 11.  free coffee and donuts!
    o used to have a 1 weeks class on ASR33 repair.  you had to 
      take one apart and put it back together to pass.
    o A lot of rumors at that time from peple who *said* that they
      threw computers, modules, and tool boxes out the door.
    o later had an office on 3-5.  the floor was so uneven, you had
      to position your desk so that you would roll into it and not away!
    is there still access to all those tunels? I can't remember where
    they are all located, but it used to break up  a boring day.
    Newcomers having to navigate from one fire exit number to another,
    is there a map now?
    And will VMS ever be as good as tops-10 or TOPS-20? How could
    a company go backwards in its software offerings?
    ed badger
    

    Tool boxes in the mill pond, definitely.
    
    Machines?  Maybe.  One of the real dog (to repair) machines on the
    first floor of building 1 did disappear late one night in 1970.
    Did it go in the mill pond?
    
    They drained the millpond once while I was in Maynard to repair
    some of the gates and there was a lot of mud covered "somethings"
    laying within throwing (heaving?) distance of the buildings.
    
    I'd still like to know what happened to "Popsicle Mary", a PDP-8
    that was so poor that it wouldn't run without tons of popsicle
    sticks jammed between the modules to keep them tight.

    One night, some folks were debating the depth of the mill pond.
    So Richie, I think it was, took a DECtape and unrolled it out his
    MLO5-5 (oops it was ML5-5 back then) window and measured the distance
    to the water. Then again, measuring the distance to the bottom.
    The difference, was presumed to be the depth.
    
    Of course, DECtape being the reliable medium that it was, the tape
    was then put back in service the data used!

    >> Are there really old prototypes swimming around in the Mill pond?
 
    Don't know on this one.   
    
    >> Why did the PDP-6 use 36 bits? 
    
    (Taken from Computer Engineering, a DEC View of Hardware Systems
    Design, by C. Gordon Bell, J. Craig Mudge, and John E. McNamara,
    Digital Press, 1978.)
    
    Initially, the PDP-6 was designed to extend DEC's line of 18-bit
    computers. (The PDP-5 was designed at the same time--it was the
    predecessor of the PDP-8 and hence a 12-bit computer.) "Earlier
    DEC designs and the then-current six-bit character standard forced
    a word length that was a multiple of 6, 12, and 18 bits. Thus, a
    36-bit word was selected." (p. 492)
    
    >> Which operating systems were written by DEC, and which came from
    >> outside? 
    
    (I won't try to remember the dozen early -11 systems or all the
    -8 systems.)
    
    PDP-8: OS/8, TSS/8 were written by DEC.
    
    PDP-10: TOPS-10 written by DEC. TENEX written by BBN. TOPS-20 evolved
    from TENEX by DEC.
    
    PDP-11: PTS, DOS-11, RT-11, RSTS, RSX-11A, RSX-11B, RSX-11C, MUMPS-11,
    RSX-11D, RSX-11M, IAS, RSTS/E, RSX-11M-PLUS, P/OS, Micro/RSX,
    Micro/RSTS, MicroPower/Pascal all written by DEC. UNIX and TSX-PLUS 
    written outside. (Actually, RSTS was written by outsiders on contract 
    to DEC.)                                        
                                                    
    VAX: VMS, VAX/ELN written by DEC. UNIX written by Bell Labs.
    
    PC: CP/M, MS-DOS written outside.
    
    >> How did the PDP-8 come about? 
    
    (ibid)
    "After visiting Chalk River in the winter of 1962, DEC engineers
    decided that a 12-bit design based on the DC-12 would be excellent
    for such a front end in PDP-4 process control applications. The
    instruction set for the new machine was specified in detail by Alan
    Kotok and Gordon Bell, and the logic design was carried out by Edson
    DeCastro...."
                 
    >> How did the PDP-11 become so popular? 
                 
    I am sure there are many views. So I will quote again:
    "The PDP-11 has evolved quite differently from the other computers
    discussed in this book and, as a result, provides an independent
    and interesting story. Like the other computers, the factors that
    have created the various PDP-11 machines have been market and
    technology based, but they have generated a large number of
    implementations (ten) over a relatively short ([its first] eight-year)
    lifetime. ...there are multiple implementations spanning a performance
    range at the same time.... The PDP-11 designs cover a range of 500:1
    in system price ($500 to $250,000) and 500:1 in memory size....
    
    >> Anyone know anything about the 11/780 development project? 
    
    Yes. The project came from a business task force in Feb 1975 to
    evaluate growing the -10 down vs the -11 up. Members: Peter Conklin,
    Bruce Delagi, Robin Frith, Larry Wade. An initial architecture team
    was called together on April 1, 1975, (truly!) "to investigate a
    minor extension to the PDP-11 to run larger programs." The team
    was called VAXA; initial members: Gordon Bell, Tom Hastings, Richie 
    Lary, Dave Rodgers, Steve Rothman, Bill Strecker. The 780 project was 
    initiated at the end of June, 1975. It shipped December, 1977.
    
    >> Who wrote the VAX architecture? 
    
    In addition to the VAXA members above, major contributers were also
    Ron Brender, Peter Conklin, Dave Cutler, Rich Grove, Dick Hustvedt, 
    Marty Jack, Jud Leonard, Peter Lipman, Mary Payne,
    Bob Stewart.
    
    >> Who spec'ed out VMS? 
    
    In addition to the above people Patti Anklam, Bill Brown, Scott Davis, 
    Lois Frampton, Sue Gault, Andy Goldstein, Roger Heinen, Herb Jacobs, 
    Leo Laverdure, Hank Levy, Carol Peters, Trev Porter   

< Note 90.4 by HUMAN::CONKLIN "Peter Conklin" >
>   (Actually, RSTS was written by outsiders on contract to DEC.) 
    
    If you're thinking of EG&H, they only did BASIC-PLUS (which was
    sort of synonymous with RSTS in those days) but the operating system
    was done in-house.
    
  --Simon
    

Speculation--

  I didn't think that DEC did TSS-8 (Time Sharing System-8).

  I also thought that TSS-8 grew into Rsts Version <4A.  
  Simon, You'd know better than I but I thought Rsts was an 
  outside product until Version 4A.

  Mumps was originally designed at MGH (Massachusetts 
  General Hospital).  What was the original host hardware?  
  When did it become a Digital product?

                                   Atlant

Mark Bramhall confirms my previous reply:

EG&H did (the first version of) BASIC-PLUS.
RSTS was done totally in-house.  Originally, by Nathan Teicholtz and
someone else I cannot remember.  And, yes, I was there starting in
RSTS V2A-19...

    re TSS-8:
    
    Yes, RSTS is culturally in the same family tree as TSS-8, DECSYSTEM-10,
    and even (remotely) MIT's CTSS (which didn't run on a DEC machine).
    But RSTS didn't inherit any code from TSS-8.
    
  --Simon
    

    	And you have missed RTS-8 (done by Richie Lary, reputedly in
    a spare weekend). That is a multi-tasking, priority scheduling system,
    similar to RSX.
    
    	And CAPS-8 and CAPS-11. I am not too sure who wrote them, but
    CAPS-8 was written to take advantage of the new advanced features
    of the VT05, so it was almost certainly written in-house.
    
    	And further back, there was the 4k Disk Monitor system for the
    PDP-8. As a customer I modified that to run on TD8-E DECtapes as
    the system device. The command language was closely modeled on the
    PDP-6 operating system, so that was probably in-house too.
    
    	We have had a mention of OS/8, but of course that was really
    just PS/8 with code in it to detect a PDP-8 look-alike and crash.
    (Incidentally, there was a bug in that bit of code, which I SPRed.
    After clearing the bootstrap blocks on your system disks, it failed
    to correctly clear memory. I am not sure if the bug was ever fixed
    :-)).
    
    		Dave

    Does anyone remember GLC-8 ? A Gaschromatograph turnkey application.
    Was it ever successfull ? I used to service a customer who after
    a painful startup-phase finished to like it quite much. I believe
    that he is still using it.
    Did DEC write this software ?
    And how about FOCAL ? Is it DEC's ? What made it loose versus BASIC
    
    Fred

    The first project I worked on at DEC was GLC-8/II.  (GLC stood
    for Gas and Liquid Chromatography.)  I believe it had only one
    big customer, but I never really learned how well it did.
    
    						-- Ann

    FOCAL was designed and built in DEC by Rick Merrill.  Rick is still
    around the mill somewhere.  For a while we pushed it as our in-house
    developed alternative to BASIC for the 8.
    
    TSS-8 was (mostly) done by George Berry (who now is somewhere in
    CXO).

        Rick Merrill of Focal Fame, is now a Product Manager for the
        LN03.
        

Martin Minow did an implementation of FOCAL in C.  Since it runs under DECUS C
(11's with all current OS's except MUMPS) and VAX C, FOCAL should easily win the
contest for "language implemented on the widest variety of DEC processors".

I'm pretty sure that the original MUMPS ran on an 11.  It's been a DEC product
for at least 8 years, probably longer.
							-- Jerry

    	I did an implementation of FOCAL in MACRO-32. It was submitted
    to DECUS in August 1978, and had to be given a PDP-11 number, since
    they did not have a category for VAX programmes then. It was later
    renumbered as VAX-1.  The latest version has a number of modules
    in BLISS.
    
    	FOCAL was originally written for the PDP-5, and the earliest
    version that I used still had hooks in so that it would run on the
    PDP-5 as well as the PDP-8.  It lost out against BASIC because it
    was designed and pushed by this tiny startup computer company that
    no-one had heard of, while BASIC was designed and pushed by this
    large, influential educational establishment.
    
    	There are implementations of FOCAL for almost every operating
    system that DEC has ever produced. I have seen it on DOS-11, DOS-15
    and IAS, just to quote examples. At one time it was the only high
    level language that had good RT-11 GT graphics support built in.
    
    	I think the original MUMPS was written by a Massachussets hospital
    for the PDP-15.  I used the PDP-11 DEC product 10 years ago, but
    I am not sure when DEC first took it in house.  Since the early
    versions could have been considered as not much more than FOCAL
    with added support for string indexes and virtual memory, maybe
    we could claim it had never really been out of house?
    
    		Dave

    Rick Merrill is NOT the Product Manager for the LN03!!!  Rick mainly
    works with fonts and software applications w/the LN03.
    
    

    I believe MUMPS began at Mass. General Hospital (MGH Utility (?)
    Multi Programming System).  I first read about it in the early
    '70s in, if I recall, a blurb from (low rumble in background)
    Data General.  It was on a Nova.  Maybe they also had it on a DEC
    machine too.

    I claim that TECO is the most widely-implemented programming
    language on DEC machines.  In addition to the machines quoted
    for FOCAL, it was on the PDP-1.
        John Sauter

    MUMPS was first implemented on a PDP-1 by BBN for MGH, as I recall.
    I saw it running in the mid 60s.

    FOCAL was the first language other than assembler actually running
    on the VAX. As I recall, it beat FORTRAN by a couple of days.

< Note 90.17 by DSSDEV::SAUTER "John Sauter" >
                             -< Prog. Lang. TECO >-


	TECO wins in TWO categories then, Language AND Editor!

    Re: PDP-6 word length...
    
    To approximate an Alan Kotok quote:
    	"Well, we wanted to do LISP, and storing two addresses in a
    	 single machine word would benefit a LISP implamentation.
    	 Since 18 bits of address was more than ANYONE would ever need..."
    
    Danny

    re .9:
    
    FOCAL lost out to BASIC (in Digital) for interesting reasons.  One
    of the chief of them was the then startup PDP-11 product line didn't
    want it (originally) and opted for BASIC instead.  Rick Merrill
    had written a PDP-11 version that performed spectacularly, but nobody
    was interested during the early period.
    
    Re the GLC-8:
    
    Originally, it was the CasChrom-8, but we were hit with a Cease
    & Desist order by some company who already had a trademark
    "Gas-Chrome," or some such, that was just too close.
    
    The original software was developed outside Digital under contract,
    as I recall, but it had problems.  Gary Cole (now gone from Digital)
    did some of the first fixes on that package.
    
    Steve Kallis, Jr.
    

    Re:  Early Success of the PDP-ll
    
    There are several reasons why the PDP-11 did so well, but one of
    them was some good marketing.  There were lots of charts - in
    all the literature and sales aids (such as flip charts, and slide
    presentations) showing the UNIBUS with all the system components
    hanging of it:
    
           <--------------------------------------->
             |     |       |        |        |
          CPU     MEM.    DISK   TERM.     PERIF.   ETC.
    
    
    This simple architectural concept was the starting point for
    all presentations and lead to messages like:  ease of
    interfacing, ease of expansion, peripheral device registers
    simply locations in memory, etc., etc.
    
    It slaughtered the competition - they just couldn't prove
    in simple terms that their architectures were better or provided
    other benefits.
    
    Some competitors tried to claim that the Unibus was a bottleneck,
    but the simplicity of the charts, and presentations, and messages,
    was overwhelming successful in sales situations.
    
    dd
    
    
           CPU   
             

Dave Conroy wrote a version of Focal in C -- I had nothing to do with
it. 

Focal appears to me to be a descendent of JOSS (Jule's [Schwartz] Own
Software System) that was developed in the fairly early 1960's. They
seem to have many of the same quirks (such as Do a statment or
paragraph, for example). 

I would suspect that Fortran and Cobol predate Teco/Focal and both
are certainly more widespread.

Martin.

    re .24, et al.:
    
    When I forst saw a FOCAL demo, it was called "FORGE," and was being
    run on a PDP-8/S.  It was wonderful!  You'd input a problem, then
    go to lunch while it chugged its way through the 8/S....
    
    Steve Kallis, Jr.
    

    We have had PDP's ranging from -1 to -16. I have never seen any
    reference to a PDP-13. Is superstition present even at DEC ?
    
    Will the PDP-16 remain forever the last numbered PDP ?
    
    Fred_who_loved_the_PDP-15
    

    I have heard it said that the PDP-13 developers left Digital and
    produced the machine at another company.  The other company couldn't
    call it the PDP-13, so that name was never used.  When the machine
    was released it was called the Nova, or something like that.
        John Sauter

    I'm relaying this message from Jack Bowie (Engineering Manager of MSG),
    who was at MGH when MUMPS was originally  developed. 


    Bah Humbug to you all.  MUMPS was developed at MGH Laboratory of
    Computer Science (not to be confused with the Laboratory for Computer
    Science (MIT)), or was it the other way around??.  Anyway the first
    implementation in the mid-60's was on a PDP-7.  It was rapidly migrated
    to the PDP-9.  DEC picked it up in the late 60's and put it on the
    PDP-15.  This was the first Digital MUMPS product.  Next came MUMPS-11
    developed in conjunction with Meditech in Cambridge.  In the mid-70's
    following the standardization effort, we brought out the first version
    of DSM (Digital Standard MUMPS).  This was DSM-11.  Then in the early
    80's we developed VAX-DSM.  It could have been ported to a PDP-1 at BBN
    but I am not aware of that implementation.  I do not believe there has
    ever been a PDP-8 version, although there was a very early attempt at a
    layered product on the 10. 

    (The 10 version was not a Digital effort needless to say) 

    Jack 


    So now you all have the straight scoop on how/when MUMPS came about. 
    (BTW, this year is the 20th anniversary of MUMPS.)

                                                     Bob 

re: < Note 90.27 by DSSDEV::SAUTER "John Sauter" >

As I recall (and that takes a lot of recalling) when I joined DEC in February
1968, that project was called the PDP-X.  It and its engineers eventually
went on to become DG.  One of the reasons I heard was that DeCastro wanted
to make the system out of quad and hex boards.  The powers to be at that time 
said no.  

Another thing I heard was that the machine was designed for the specific
purpose of never getting it past the Engineering Committee (or what 
ever it was called back then).  The engineers (DeCastro, Sogge, Burkhardt)
wanted to go to form another company but needed a product so they developed
it here fully intending to go somewhere else and build/sell it.

Maybe there is someone still around who can fill us in on what really 
happened.

    re .29:
    
    I was here just _after_ the events.  My only regret is that Digital
    didn't make deCastro, te al. live up to their employee's agreement;
    since what became the Nova was developed here, it was a pity they
    were able to take it with them.
    
    re .26:
    
    The VAX-11/780 was originally slated to be called the "PDP-11/770"; It
    _might_ have become the PDP-770, but wiser heads prevailed.
    
    Steve kallis, Jr.
    

    re .30:
    
    Actually, when we went to name the VAX, the sequence at Operations
    Committee was:
    
    We knew that it had to be a PDP, because everyone knew Ken wanted
    all systems called PDPs. We wanted to emphasize the -11ness of the
    product. It was much more than the 11/70--at that time the top of
    the -11 line.
    
    So we proposed PDP-11/170. But Andy Knowles had reserved the 100s
    for a very secret project. Turned out to be the PDT!
    
    So we proposed PDP-11/270. But Andy's reservation included planning
    ahead for the future. (Incidently, DECmate is a 278, but...)
    
    We purposely skipped the 300s because they were already used for
    DATAsystems. We didn't bother explaining that.
    
    Operations Committee therefore decided to call the product the
    PDP-11/470. It took us several weeks to convince them that the 470
    name was already widely known in the industry!
    
    Since the 500s were in use by the DATAsystem family, we decided
    to learn something from Andy's taking out a series plus room to
    grow.
    
    We had fond dreams of the VAX family growing down and up. So we
    "reserved" 600s, 700s, and 800s. We therefore proposed and got approval
    to call the product the PDP-11/770.
    
    Then an expert on names and promotion asked why use PDP. We said
    because Ken wanted it. Ken said "who, me?" And the expert said,
    a much better choice would be pronounceable and have the letter
    X because short words with X stand out. We had such a name in the
    project, VAX. So the decision was VAX-11/770.
    
    When the final logo was done by an artist, it looked like
    
    		VAX-11
    		    770
    
    and was fine, until we noticed that the 1 slanted down and ran into
    the left end of the 7's top. (I can't draw it here, sorry.) But
    Bernie LaCroute, being French, noticed that two of these next to
    each other were exactly the lapel sign of the Nazi SS troupers.
    He decided this was very error prone and would be bad promotion.
    So we changed to VAX-11/780.
    
    Eight years later, we introduced the 610 and 630 processors (MicroVAX).
    And the 8000 series, which is just the 800s with an extra zero thrown
    in for emphasis. So the reservation plan actually worked!

    	The 500 series must have been considered for a bit longer than
    .31 implies, because I still have a draft copy of the RMS-500 manual.
    
    		Dave

    The DeCastro machine was indeed the PDP-X.  The PDP-13 was skipped
    for the usual hokey reasons.
    
    The PDP-X was similar in design to the NOVA, but not the same. 
    DeCastro wouldn't have been that dumb to give us a legal grip on
    him.

    Either my mind is going or I remember a talk by KO some number of years
    ago in which he indicated they did not take legal action against DG at
    the request of one of the people who originally went to start DG but
    then blew the whistle on them.  The story I remember is that this
    person came and told KO the exact story of what DeCastro was up to
    because his lawyer had warned him it was illegal, but expressed great
    fear at what would happen (to him) if Digital went after DeCastro.  KO
    decided it wasn't worth the law suit.  Sounded like a movie script to
    me, but that's the way I heard it. 

    Re .33, .34:
    
    There's a raft of stories involving deCastro & associates that old
    timers could possibly tell, but for the sake of avoiding contentions,
    it's probably better that these are left unsaid.
    
    Steve Kallis, Jr.
    

    RE: .31
    
.                                            We had such a name in the
.    project, VAX. So the decision was VAX-11/770.
    
    What was the origination of the name VAX ?
         
    -phil

    
    Vitrual
    Address
    Extended (or Extension)
    
    Brian
    

    (footnote to .7:)
    
    RSTS was first called BTSS.  Dave Knight and (the person whose name
    Dave nor Mark Bramhall could remember) first worked on BTSS before
    it was renamed RSTS and before Nate Teicholtz joined the project.
    ('BTSS' stood for BASIC Timesharing System, I think.)
    
  --Simon
    

    I remember spending an hour or so with Julius Marcus (PDP-11 Marketing
    Manager at the time) trying to come up with a catchy name that
    conveyed the idea of a combined time-sharing and resource-sharing
    system as opposed to a plain old disk operating system.  We just
    couldn't latch on to anything good, so he finally said "Well, I'll
    settle it now, it will be Resource Sharing - Time Sharing System
    (RSTS).  Little did we realize that this name would catch on
    quite easily and be pronounced as "RISTUS".
    
    dd

    re .37
    
    I'v heard many times that VAX stands for:
    
    		Virtual  Array  Extension 
    
    
    

    Re: .40:   Nope, it's Virtual Address Extension.

    "what does VAX stand for" is beginning to sound like "what does LASER
    stand for".
    
    What it originally stood for is less important than what it is now
    assumed to stand for.
               
    
    /. Ian .\
    
    PS LASER originally stood for "Light Amplification by Stimulated
    Emission Resonance", but the "R" is now assumed to stand for "Radiation"
                                       
    

    
    	Well, in England they refer to vacuum tubes as "valves".
    
    It is generally assumed that this came about as a way of describing the
    fact that they can be turned on and off (athough we don't regard then
    now as current devices).   What (documented) history has neglected
    is the fact that the first commercial manufacturer in Europe was
    a french company by the name of   V.A. l'vie.   In the same way that
    "Honda" has become the generic name for small motorcycle (and later
    LARGE motorcycle) their company name was adopted by the folks across
    the channel (who contracted the pronunciation to "valve") to describe
    ALL vacuum tube devices. 
    
	Reg
    
    		(History gets itself confused TOO often)
    

Way back when, there was VAX (Virtual Address[ing?] Extension, in contrast to
PAX (Physical A X).  The two acronyms refered to different approaches to dealing
with the PDP-11's 16-bit address limitation.  PAX referred to the use of 22-bit
physical addresses (on the "Q-Prime" bus - the original 18-bit Q-bus is now gone
and mainly forgotten, except by people who have the pain of supporting old
18-bit-address only peripherals, like the RX01, on 22-bit machines) and the PMI
on Unibus machines like the 11/70.  Of course, "VAX" ended up implying a lot
more "physical" extension than PAX ever did!

re:  LASER:  LASER is a "secondary acronym".  The original device was the MASER,
Microwave Amplification by Stimulated Emission of Radiation.  MASER's pre-date
LASER's by several years.  When the first "LASER's" were built, they were often
referred to by the nonsensical appelation "optical MASER".  Now, LASER's are
so much more visible thatn MASER's occaissionally get called "microwave LASER's"
and of course we have the proposed "X-ray LASER's" as part of SDI.

							-- Jerry

    aside: what MASER stands for.
    
    The early research papers defined MASER as Microwave Amplification
    by Stimulated Emission Resonance, which is more accurate than the
    tautological, but commonly stated definition.
    
    /. Ian .\

TSS-8 was originally written by Don Witcraft (sp), and then made to work by
Nat Teicholtz after Witcraft left DEC. Mark Bramhall was also one of major
contributors to TSS-8.

Nat Teicholtz later went on to do RSTS, which was originally called BSTS or BTSS
(I can recall which) until that name was discovered to already be a trademark.

Jeff Scott, long since departed, was the other DEC guy on the original RSTS
team. If I recall correctly, Jeff was responsible for parts of the Basic-Plus
runtime. EG&H did the compiler and the rest of the runtime code.

EG&H were hard up for work back in those days, and bid something like $12k for
the contract as a "loss leader", I guess.

The various EDUsystems for the PDP-8 were omitted from the list. There was
EDUsystem-10 (single user BASIC), -20 (a batch BASIC using OCR cards) and
-30 (TSS-8 BASIC), I think.

As has been related elsewhere, PS-8 was designated by Richie the "First Upward
Compatible Keyboard Monitor" to denote it's similarity to the PDP-10. The
acronym FUCKM survived in some of the program listings until someone told
the SDC about it.

Then there was INDAC-8 and TYPESET-8, both developed at DEC.

And others, I'm sure, that I've forgotten.

			Regards,
			   George

Can Anyone tell me story of PDP-11/74 ?   Which OS is running on ?

BTW,
(OFFICIAL) VAX ARCHITECTURE HANDBOOK says, "The Letter VAX suggest the premier
future of VAX computers -- Virtual Address eXtension" 

} if

        I think there was a discussion of 74's and RSX in the RSX notes-11
        notesfile. RSX Developement still has a 74 at ZK Castor/Pollux::
        
        q
        

    	.47 missed EDU-25, which was a time sharing BASIC system.
    I used to have the sources of it at one time.

        In fact, checking my copy of the handbook, there were :-
    Starter Systems --      EduSystems 5, 10, 15 and 20
    Intermediate Systems -- EduSystems 25, 30, 40 and 40/EDP
    Total Systems --        EduSystems 50 (and 80)
    
    	A footnote on page 10-13 explains "EduSystem 80 is only available
    on the PDP-11. Contact the Educational Products Group, or refer
    to the PDP-11 Resource Time-sharing System Users Guide
    (PL-11-71-01-01-AD) for more information.".
    
    	Other careful reading reveals that "A Total System means
    time-sharing power to handle the whole school's needs simultaneously,
    and resource sharing power that keeps the whole computer at work
    all day long.". To do this, an EduSystem 50 required a minimum of
    16k 12 bit words of memory. Another feature was that it supported
    file sizes (on disk or DECtape) of up to 350,000 characters. Supported
    languages were Fortran-D, Focal, Algol, Basic and PDP-8 assembler.
    
        Dave

re:  11/74

As I (mis)remember it, the 11/74 was essentially an 11/70, but had two plusses:

	dedicated cpu backplane slots for a "commercial instruction
	set" (just like the way the floating-point option works), and

	quad-ported MOS memory, so you could hook up to four of them
	together in a multi-processing configuration.

I remember some of the RSX-11M folks being pulled to work on hacking 
together a version of -M that would run this beast.  I heard they even got 
a 2-CPU version up and running.

It never got into volume production, though.  It was never sold to 
customers (although I seem to remember us having a hard time getting one 
back from a field-test site).

After it was withdrawn, all the units built were sold to in-house groups 
who wanted -70's and couldn't get them.

All this was in 1973.

Ray

re: the 11/74

A quad processor still lives in the raised floor computer room in ZK1.
RSX-11M-PLUS was originally designed to support the 74 and the code is still
there.

	- pb (co-author of M-PLUS)

>A quad processor still lives in the raised floor computer room in ZK1.
>RSX-11M-PLUS was originally designed to support the 74 and the code is still
>there.

In fact, the RSX Group still uses Castor for RSX-11M-PLUS and Micro/RSX 
development. It is the LAST remaining (I believe) multi-processor 11/74 
still in use in the Corporation...

        And I understand it is on a raised floor because of the numerous
        cables involved.
        
        q
        

    ELROND the DECnet/RSX group's development system used to be a dual
    11/74. It got dismantled and never reincarnated in the move from
    Tewksbury to Littleton. I think DECnet/RSX still has multi-processing
    code conditional assembled out.
    
    Dave

    The 11/74 project was not as early as 1973. In 1976, there was advanced
    development activity to determine how to modify RSX-11M to work
    in a multi-processor environment. As I recall, the AD work was on
    a modified 11/34. Then the project was implemented on a modified
    11/70--hence the 11/74 name. This project had two parts: a Commercial
    Instruction Set processor, and the quad processor. The CIS part
    was cancelled around 1979. The quad processor was approved for
    announcement by the predecessor to PAC around 1980. It was withdrawn
    from announcement the next week because marketing decided it had
    no plan! At the time of the cancellation, approximately 100 processors
    were in production in Burlington, Vermont. They were used for shipment
    to DEC facilities in the New England area.

re .56

Pretty much right on the nose.  The first more or less usable base levels of
M-PLUS were in early 1977.  BL2A which was the first "releasable" base level
was late in 1977 and was the basis for TRAX.  BL2A had many short comings and
probably was not the right one to use for TRAX.  Looking back I think we 
would have done it differently. 

The software was initially (up to BL2A) done by a small team of 5 players 
working for Bill Munson and Frank Hassett.  Only two remain with DEC, Frank
and myself.  The remainder of the developers were Kim Kinnear, Peter 
Wanheden, Eric Baatz, and Tom Miller.

    I thought there was a newspaper publishing company in Toronto that
    had a 11/74 being used in production.   I remember them being one
    of the field test sites that refused to let it go because it was
    so nice.  Is it a myth?   (I remmeber this from about 7 years ago
    when I was working in the DDC, but I have to admit I never worked
    on the machine myself.)

    No, the only 11/74 configuration brought into Canada was field tested
    by Alberta Government Telephone (AGT).  And they, in fact, let it
    go for VAX processors.  (I'm not sure of the "why" details).  The
    '74 was taken back by Field Service, which used it for their call
    handling and contracts system as two '70's.  New backplanes were
    ordered for both of them, but only one was ever installed.  The
    two systems now sit in the Calgary (CGO) office where they're used
    for Field Service diagnostic training and testing.
    
    b.

    re: .54 The RSX development system is in a raised floor room not
        because of cable complexity, but more for protection. When the
        project was cancelled, the engineering of the above the floor
        cable troughs was not yet complete, so we opted for raised floor.
        Due to memory cable length, the four CPUs have to be arranged
        in a square.
    
        As was pointed out earlier, a significant amount of info on
        the 11/74 can be found (I believe) in the RSX support notes
        file on VAXWRK.
    						-Brian
    

>        As was pointed out earlier, a significant amount of info on
>        the 11/74 can be found (I believe) in the RSX support notes
>        file on VAXWRK.

I think it was in SYSENG""::SYS$NOTES:PDP11.NOT.
(A NOTES-11 notesfile, the last time I looked.)

Much of the detail that's there was entered by the author of .-1.

Bob

    By the way, what's the history of the PDT series?  The PDT-11/110,130,
    and /150 got released, but I understand there was also something
    called the PDT-11/250 in the works at one point.
    
    john

	RE: .56

>				This project had two parts: a Commercial
>   Instruction Set processor, and the quad processor. The CIS part
>   was cancelled around 1979. The quad processor was approved for

	There were two proto units built.  We had one 11/74 CIS system
	running as a single (fast) CPU up until last year but it was
	scraped (we still have the spares though).

	There was a third part to the project.  A little module called
	the IIST ("eye_squared_S_T" stood for Interprocessor Interrupt and
	Sanity Timer).  This provided the path to synchronize the separate
	CPUs and allowed for some high availability hooks where one processor
	could be re-booted by the others.

		Bob L.

    
    I hear tell Y0.00 of VMS ran on an 11/70. True?

>    I hear tell Y0.00 of VMS ran on an 11/70. True?

    Don't know about Y0.00, but the VAX hardware was emulated in the Mill in
    '75-'76 on a pair of PDPs (probably 11/70s, can't remember now)
    so this may have started the rumor.
    
    The story at the time was that a couple of 16-bit machines were being
    run in parallel to make one 32-bit CPU ...   :-)

        Word that I had heard, andy, was that the Original  780's  were 
        proto-ed in drastically modified 11/70 backplanr/cabinets.
        
        q
        

    I remember seeing a couple of PDP-11/70's highly modified so that
    they would run VAX code.  They had a very large WCS and ran at about
    1/10 real time of the "Star" prototype.  I think there were 2 of
    these funny 11/70's and one proto here in the Mill on ML3-5.  In
    those days they even frosted the windows so that no one would see
    them from the hall.
    
    	/s/	Roger
    

I remember those frosted windows all too well.  At the time, security (as far
as locked labs go) was virtually unheard of.  The first time anyone came to 
this odd looking room the normal thing was to go in and see why it was frosted.
Then, upon finding out the door was locked you'd go and ask someone why and were
told something like, "that's where they're building the VAX, but don't tell 
anyone 'cause it's a secret", even though everyone knew what the VAX was.  You
just couldn't see one.

-mark

    re .66: (From one of the hardware designers of the 780)
    
    The "hardware simulator" was not made from an 11/70, although they may
    have used some 11/70 I/O equipment.  The box was a microprogrammed data
    path (like the 780 and 750) but was very flexible without much
    dedicated hardware for address modes, etc. so that it could change when
    the architecture changed. I can't recall any more how it was
    constructed but I'd guess that it was wire-wrapped on specially built
    collage modules.  Physically, I think it was mounted in two 72"
    cabinets. 

The Star processor was also software simulated. The CAD group (ML1-1)
(actually Phil Corman) coded a MIMIC-based Register Transfer Level (RTL)
simulator of the -780 processor. It was used for debugging the microcode. 

Also the same approach was used for the 11/74 Commercial Instruction Set
processor. That project was on again/off again a few times and was finally 
canceled. I know, i coded the simulator and never got to completely debug it!
The CIS was an AMD2901 bit sliced architecture design with about a 60 bit wide
microword. 

paul sears

    Haven't seen anything about the 36-bit Jupiter project that suddenly
    went "south". Anyone know anything about what it looked like, what
    it's performance was, why the project died, etc? 
    
    							Warren
    

    From all accounts I heard, JUPITER was 'out of control'.
    
    
    It was way over budget, and manufacturing could not consistently
    manufacture a machine that worked. 
    
    
    I spoke with one of the engineering managers one day when I was
    on a customer visit.  He's now a salesperson in the NYC office,
    handling Citicorp.
    
    
    
    -bill

According to rumor, they had a few problems ECO'ing the speed of light.

M.

It's interesting that the CUSTOMERS have the perception that it was mainly due
to a decision to focus on the VAX architecture.

/john

    A story I heard at the time from a reputable source was that the
    project team based their work on a preliminary estimate of CPU speed 
    using a simulation program, into which was fed a (Gibson-like) "mix" 
    of the supposedly most frequently-executed instructions.
    
    Unfortunately, the mix they used was old, out-of-date, and consisted
    of non-privileged user instructions. In particular it omitted one
    specific instruction which was used frequently by the TOPS system;
    the consequence was that when the prototype began to run and TOPS-20
    was booted, the average instruction rate was nowhere near what the
    simulation had predicted.

    The impression I had at the time was that if it was shipable when
    it was cancelled that they may have shipped it.  But since it was
    not ready that it would be too little too late.
    
    KO came down and gave us a pep talk that didn't quite pep us...

    Re:.71
    
      The following is a memo regarding the JUPITER project a  friend forwarded
to me a couple of years ago.  I don't know when it went out (I started at DEC
2.5 yrs ago) and I don't remember who wrote/sent it out but believe it was 
one of the managers of the project.

-Jim



As I said, the problems with Jupiter, which ultimately resulted in
its demise, can be summarized as:

1) The architects didn't understand the PDP-10 architecture to the
point where they could predict which instructions would be executed
most frequently.  Consequently, the machine they designed took too
long to do frequent operations, like effective address calculation, or
extend operations.

2) Relating to 1), and compounded by a lack of money to fund engineers,
the design was "hand waved" in certain critical areas; thus, when they
built the breadboards, they found out that the Ibox could not be made
to work.  They redesigned the Ibox twice, but the problem was subtle
faults in the machine architecture.

3) Because of 1), and some by 2), the microcode store turned out to be
inadequate to implement the PDP-10.  The current design had 4K words,
and they needed more like 16K.

4) The design was too complex for the CAD/CAM approach that was taken.
Because all the timing analysis, (for example) was done by hand, there
were many mistakes.  Timing skews were a real problem.  The simulation
done was minimal, with the result that with a design of this
complexity (60K+ gates) the machine could not press the technology
hard enough (signals take a whole nanosecond to go less than a foot of
foil, etc.) to get adequate performance.  The design goal was a cycle
time (i.e., microclock rate) of 22 nanoseconds. This is enough time
for a beam of light to go from one end of the cab to the other side,
and back.  Not much time for logic.

When they tried to fix all these problems, they ran into others.  They
ran out of power in the CPU bay, the memory didn't have the bandwidth
to support the CPU performance, there were cooling problems, etc.  It was a
size 0000 can of worms.

The design did work, and it did implement the PDP-10, including extend
instructions (although you needed different microcodes if you wanted
to do floating point, for example), but it would have been just a
little slower than a KL10 for the same money.  Fixing it would take
three (lets be realistic, five) years.  It was decided that it wasn't
worth it.

IBM, too, found with its 3081 series that hand design methods just
didn't do the job.  INTENSE cad/cam is what they finally did.  It took
them fourteen (seriously) years to go from start development of the
TCM technology, to the first announcement of the first product (3081.)

Do forward this to the appropriate people..

    re .77
    There was also a problem with the two individuals that were designing
    the machine. They hated each others guts. It seems to me that that
    problem should have been resolved by their boss. But it was not.
    This was a major cause of the machine failing. There is a lot of
    information that is not known about the real reasons that this project
    failed.
    
    

    A few DEC products end up failing--or nearly so--because of personality
    conflicts.
    
    MINC I--the laboratory machine--nearly got derailed because of major
    conflicts at the grass-roots (peon) level, and the middle managers
    admitted that they didn't know how to handle it.  A complete shift
    of responsibilitiy to different managers finally sorted things out
    (one of the rare cases where reorganization was both meaningful
    and effective).
    
    IVIS was years late in development and delivery, in my opinion,
    because middle and upper management were not effective.  The project
    kept getting whipped around, and technical issues were either ignored
    or surpressed.  IVIS is an example of a good idea gone wrong because
    of poor planning and leadership.

RE.: .74
    
    When I started with DEC in '79, installing and maintaining KS10's
(DECSYSTEM-2020's) it was told to me that they would not catch-on due to
them supposedly being direct competition to the 780's. I would say that
it has not been just a customer perception that we have become the VAX
Corporation by paring down architechtures and operating systems. Many
people who had worked in LCG before July '83's announcement felt that
20's and TOPS-20 would be around for a long time to come. Obviously
not so. Now VAX/VMS people feel that VAX will be around for a long time
to come. I wonder.

lee

    For obvious reasons, any discussion of a "new machine" technology
    wouldn't be posted in a NOTES Conference.  I would guess, however,
    that there are post-VAX machines being designed and built within
    DEC somewhere.
    
    PR1ME announced a new computer with parallel-processing architecture,
    in an obvious move to overtake companies like DEC and IBM.  One
    joke about the VAX development project was that it was the worst-kept
    secret in the industry.  Perhaps DEC learned from this, and is keeping
    things under wraps.

    Talk about secrets leaking. Check out the "Charlie Matco" column
    in the latest issue of "Digital Review". It mentions some pretty specific
    code names for DEC's next generation of processors. Obviously there
    are some within DEC that can't keep secrets. Either that or "Charlie"
    has a direct line "bug" in some VIP's office.
    
    						Warren
    

I really don't know how Charlie does it, but his references to a hunting lodge
for rich weirdos (or whatever he said) make me think that he does some variation
on the following scenario:

	Charlie has contacts with a number of headhunters.  Once a year,
	with the help of the headhunters, he invites all the headhunters
	and a select list of people who have just left DEC that the head-
	hunters have provided to the lodge, all expenses paid.  There he
	pumps everyone for info.

/john

    re: "secrets"  It is no secret that customers who sign "Non-Disclosure"
    agreements *TALK*.  It's hard to catch 'em, and I doubt we'd procecute
    them if caught.  Sooo, always remind people that '___ is company
    confidential information' otherwise they might think that because
    you mentioned enjoying the lobster and the FizVax project in the
    same breath that they were both casual mentions. [ oops, forget
    I said "FixVax Project" back there! ]
    
    	Rick
    	Merrill
    

    One possibility is the following:
    
    1: Go to DECUS and make note of the DECies that seem to know what
       is going on.
    
    2: Hang out at the pizza places, shusi houses, and Chinese
       restaurants local to important DEC facilities or even DECUS.
    
    3: Change tables a lot.
    
    This is an over simplified view, however, when I worked for a third
    party controller company this worked with some degree of success.
    
    Now, you may ask how one man could do all this? In Charlie's last
    article he said he could be a group of people and no one person.
    
    Even if there is a Charlie Matco, I'm sure he is not personally
    responsible for obtaining all the information he reports.
    
    							--- Neal  
       
    
    

    Interesting discussion, but what does this have to do with "Digital
    History"? 

    BTW...
    
    Charlie Matco aka Terry C. Shannon
    
    
    -yuval brash

    Any histories on the pro series.
    
    Before joing Digital 2 years ago I though that the pro was sombody's
    idea of a joke that got sent to production by accident.
    
    -Yuval Brash
    
    PS. I apologize to anyone who takes offense at my statement.

    re: .89
    
>    Charlie Matco aka Terry C. Shannon

    Shannon turned in his trench coat and coffee mugs at the end of
    last year.

    Charlie Matco is now Mike Lazor (spelling?)
    
    bill

    Please be advised there is a notes file dedicated to Digital's history:
    LDP::DEC_HISTORY. 
    
    Jamie
    

          <<< HUMAN::DISK$HUMAN_WRKD:[NOTES$LIBRARY]DIGITAL.NOTE;1 >>>
                          -< The DEC way of working >-
================================================================================
Note XXX.0*                 DEC_HISTORY NOTES FILE                    No replies
PYRITE::PEARSON                                       9 lines   7-OCT-1988 11:55
--------------------------------------------------------------------------------

    A notes file dedicated solely to Digital's history resides on LDP::
    DEC_HISTORY. It is moderated by the Digital Historical Collection
    Program.
    
    Having seen the interest in history/culture in this notes file, I
    thought I'd pass on the info.
                     
    Jamie