Sorta reminds me of when Coleco stopped making the Adam and all my CS friends
were amazed that it was still being made.
Atari stopped making the 8-bit a coupla years back.
I don't remember when Tandy/Radio Shack stopped making the Color Computer or
the TRS-80 line.
Does Commodore still make any 8-biters (C64/128) ???
<< Michael Rogero Brown | Disclaimer: I speak >>
<< CS Graduate Student-Florida Atlantic Univ | only for myself. >>
<< Internet: mich...@sol.cse.fau.edu | All my opinions are >>
<< BitNet: m_brown@fauvax | that: mine & opinions. >>
What is so great about the PDP 10? It seems most hackers who have used them
would kill to get one...why? What made them so...cool?...great?...gods gift to
hackers? Was it architecture? performance? cost? implementation? 36 bit
Be specific, why was it a great machine (or why not...)
A more important question may be, what made the PDP 10's operating systems
(TOPS-10/20, ITS,...) so wonderful? Why do people claim that U*IX, VMS, [and
DOS :-)] are a step backwards?
Although it may sound like I am PDP-10 hostile, I am not, I just want sound
reasons to support all the claims I have heard (so when I design my own
PS: Please don't direct me to alt.sys.pdp8, as we don't have it here yet
PPS: I don't want to know why/how the PDP 10 died/was killed, there has been
enough on that recently.
John W.F. McClain
Disclaimer: I don't speak for DEC and I have never used a PDP-10 (do like
I think it would take (me, anyway) a lot more than eight bites to get
through a C64...
Garrett A. Wollman | Shashish is simple, it's discreet, it's brief. ...
wol...@emba.uvm.edu | Shashish is the bonding of hearts in spite of distance.
uvm-gen!wollman | It is a bond more powerful than absence. We like people
UVM disagrees. | who like Shashish. - Claude McKenzie + Florent Vollant
>Be specific, why was it a great machine (or why not...)
>A more important question may be, what made the PDP 10's operating systems
>(TOPS-10/20, ITS,...) so wonderful? Why do people claim that U*IX, VMS, [and
>DOS :-)] are a step backwards?
As a former TOPS-10 developer (one of the last) I'll speak for that side and
let the -20 people speak for themselves, although some of what I say applies
to the -20 as well.
I think the thing that most made the -10 such a cult is that it was easy to
play around ('hack') and do something constructive. Some of this is due to
the fact that the -10 and -20 existed before security became as much an issue
as it is now (although Un*x would also fit in this category), but some was
just due to the way things were developed.
Perhaps one of the central things to this was DDT, the debugger. Much more
powerful than ODT or anything since, although it was not source level. You
could use DDT.EXE (.SAV in the old days) as a program to write anything from
a one-instruction program to reasonably complicated programs, using it either
interpretively (instruction<ESC>X executes the instruction) or writing it into
the patch area. This means you didn't have to go through a bunch of syntax
just to get something to compile and load. Quite useful if you, say, just want
to know what kind of results a particular system service (UUO) gives you or
exactly what the parameters are (and you can't figure out or don't have the
doc). DDT was also the same as the executable file patcher and the system
debugger. Since it was assembled from the same source, there was a unity of
syntax. Oh, and it was also your object file debugger.
Along the same lines, you didn't have to get everything perfect to
be able to debug a program. You could have a program with, say, undefined
symbols and still debug it. VMS (at least as I last used it) won't let you
link a .OBJ with errors (as opposed to warnings). Great for incrementally
On large-system lines, TOPS-10 always 'thought big'. It wasn't
crippled by FILES-11 (where you have to figure out how many files you're
going to allow on the volume when you initialize it). Accounting and control
information was readily available to write programs so that you could
do accounting whatever way you wanted (in the end, for example, you could
include or exclude monitor overhead time in what you charged a user for).
The schedular on T-10 was more realistic about what people wanted to see.
Little things like the control-T status line which told you really useful
things (like the reason your program wasn't running was because it was waiting
for terminal input and you missed the prompt), or, if it was running, just
how fast it was running.
I'll think of more later.-kby
>Does Commodore still make any 8-biters (C64/128) ???
They're still churning 'em out. The C64 is selling quite well in Eastern
Europe actually; on the order of hundreds of thousands of units per year.
I'm not really sure what to make of this :-) The C128 was discontinued
a while ago, though.
"Thank you, Dr. Science."
[Your blood pressure just went up.] Mark Sachs IS: mbs...@psuvm.psu.edu
DISCLAIMER: If PSU knew I had opinions they'd probably try to charge me for it.
>I don't know whether I should be amazed that the IIGS was still being made or
>that the IIe is still being made.
I'm not. The //e is a lovely machine. Just beautiful. I could kiss it.
>Does Commodore still make any 8-biters (C64/128) ???
They sell about 700,000 64's a year. They seemed to be quite the rage in
Poznan, Poland last summer.
"And in the heartbreak years that lie ahead, |++| ad...@rice.edu |++| Cthulhu
Be true to yourself and the Grateful Dead." --Joan Baez | 64,928 | fthagn!
"Very often, a common stone, thrown away and despised, is worth more than
a cow." -- Paracelsus | If these were Rice's opinions I'd shoot myself.
That's a function of the compilers -- they tend to take the object file
away from you.
Undefined externals aren't a problem -- the linker will complain bitterly
about them but still produce an executable with 0 in each of the undefined
references. The executable will run, but if you hit one of the references
you'll go piling into the protected zero page and ACCVIO real fast.
Of course one trick is to be compiling kernel mode stuff, leave out a
module, run it by mistake (because the RUN command was in the typeahead;
it was *going* to work, after all 8-) and watch the crash dump get
written.... (privileged users only, obviously.)
Plus if you really wanted a tight loop to wail on the KA-10 or KI-10 you
would just put the loop into the registers. (Not so much true onthe KL-10
as the Cache was as fast as or faster then register access)
The views expressed are my own.
Why?? As the proud owner of an Apple IIgs I can tell you that there is nothing
particularly out-of-date about the GS!!! The operating system is quite nice...
loosely based on the ideals of the Mac... but with some VERY nice extra
features like the concept of FST (file system translators) that allow foreign
disk formats to be mounted on the system transparently. The ONLY other system
I have seen that works in the same fashion is the NeXT.
Also, maybe the original hardware on the GS mayy be a little out-dated, but it
can easily be upgraded. With the newest accelerator cards, the GS can support
a full-color GUI that is almost identical to the Macs a LOT faster than most of
the colors Macs currently on the market. There are even programs that allow
the GS to use "TrueType" fonts that are totally scaleable. We can take a Mac
disk with TT fonts on them, shove it in the GS, and copy them to our hard drive
and use them directly... no mess no fuss...
It's too bad that Apple doesn't recognize the still-capable power off the GS.
GS' provide a little less power than a Mac for a LOT less money. It's the
perfect solution for schools who don't need a lot of computing power and who
don't have a lot of money... and that goes for mainstream "novice" computer
The GS is a little known gem that has provided a lot of power for the buck...
you haven't seen anything until you've seen a GS running GNO/ME, a multitasking
environment that turn the GS into a little UNIX-like box!!!!
Words from a true GS lover...
P.S. I'm even currently typing this on my GS while logged into our school's
You missunderstood. I don;t think the IIgs is out of date. It's just that
Apple for the last couple of years seems to do nothing but Mac stuff. I had
assumed that the II line was dead. Especially when they came out with that
II-emulator card for the LC.
>It's too bad that Apple doesn't recognize the still-capable power off the GS.
>GS' provide a little less power than a Mac for a LOT less money. It's the
>perfect solution for schools who don't need a lot of computing power and who
>don't have a lot of money... and that goes for mainstream "novice" computer
Yeah. If schools have to get some inexpensive computer, I'd rather it be the
IIgs then the older II (or PC-clone). I was never impressed by the original
>The GS is a little known gem that has provided a lot of power for the buck...
>you haven't seen anything until you've seen a GS running GNO/ME, a multitasking
>environment that turn the GS into a little UNIX-like box!!!!
>In article <1992Dec9.0...@cybernet.cse.fau.edu>, mich...@cse.fau.edu (Michael Rogero Brown) writes:
>>I don't know whether I should be amazed that the IIGS was still being made or
>>that the IIe is still being made.
>Words from a true GS lover...
> Ryan Moore
>P.S. I'm even currently typing this on my GS while logged into our school's
A GS VAX combination .... need i say more :-)
"In order to know the material of which an idea is made, one needs only to let
fall upon it a drop of strong acid." - Eugene Zamiatin
A basic concept is being expressed here:
The PDP-10 is the best ever *machine* to do something for. Most of today's
people are attempting to make an argument about the fact that (likely some
variant of unix) some software system is the one they most like to write code
Besides the argument about assembly language versus high-level abstracted
languages, something else has been lost: the *identity* of the underlying
machine. Perhaps some machine will come along that will be "worthy" of
being compared to a PDP-10 in some ways, but we'll never know it, because
it will be merely considered by most people as to whether or not it seems
to be a somewhat better or somewhat worse (or faster or slower, or other
less than subtle yardstick) "platform" to be become to what extent their
favorite (or not) place to run **ix on.
PDP-10's could run unix. Virtually all of the people who like the machine
wouldn't want to.
Many Elementary schools in the Southern US have computer educational
programs, and they all run on the Apple ][+ and ][e. So, there is a sound
reason for keeping the Apple ][ family of computers going; money...
Jim Pierce Bach. of Sci. in Applied Computer Science USM - Gulf Park Campus
jmpi...@whale.st.usm.edu Disclaimer: Standard.
There is *nothing* like MAGSPY.
Nothing like installing a hot patch in running monitor. Or figuring out how
to install a patch that wasn't meant to be installed as a hot patch in the
running monitor in such an order for each piece so the system wouldn't go BOOM
while you were patching it.
Sort of like an expert Edsel mechanic looking back on his days of glory now
James Harvey IUPUI OIT Technical Support/Networks
har...@iupui.edu har...@indiana.edu uucp:iugate!harvey bitnet:harvey@indyvax
In some ways this doesn't surprise me. The IIe must cost them next to
nothing to make. I think they are built in Singapore. I would guess that
they still have a life as lab instrument controllers, and they can be
networked in the classroom. There are hundreds of addon cards for the IIe.
The IIGS, which has more expensive custom chips, has been left by the
wayside by developers.
Rick Kelly r...@rmkhome.UUCP unixland!rmkhome!rmk r...@frog.UUCP
Commodore still makes the C64, and sells them like hotcakes in Eastern
The Apple IIe is unique in that it probably costs Apple pocket change to
build them, and they have all those slots.
Apple IIe processor cards:
Stepper motor control
Security system control
Well, from a purely amateur user's viewpoint (not a hacker or any
kind of "real programmer"), command and filename completion,
reasonably meaningful command names, and online help were all
wonderful...especially the first of these.
I still recall my disbelief, after being forced, kicking and
screaming, to move from Stanford's TOPS-20 systems to Unix systems
when the last DEC and Foonly machines were shut down, when I learned
that the allegedly much better Unix OS did NOT have command or
John>> What is so great about the PDP 10? It seems most hackers who
John>> have used them would kill to get one...why? What made them
John>> so...cool?...great?...gods gift to hackers? Was it architecture?
John>> performance? cost? implementation? 36 bit words?...
Hmmm. Makes me think a lot about those days. There are several answers
from my standpoint. In no particular order...
1. I first came across PDP-10s after working on IBM 360 OS/MVT and with
TSO. Not exactly the worlds most interactive systems. I was very good
at them, but...
Seeing a true timesharing system that handled many users easily was
really eye opening.
2. TECO, its variants and eventually emacs.
3. A real macro assembler. The stuff IBM had at the time claimed to be a
macro assembler, but MACRO-10 was great fun and flexible. We were
writing very structured assembly language in my days at the old
CompuServ (before they became famous).
4. The architecture. Now how many different no-ops were there?
+ General registers were locations 0 - 17 (octal) in memory.so any
memory reference instructionc also referenced the registers.
There was a ``neat hack'' in the KA-10 version of teco that I
dearly love. Since the -10 was a word addressable machine, byte
manipulations where a bit tricky. Teco kept a hole in the middle of
your buffer in memory where you were currently inserting text. When
this hole got full it needed to move a bunch of words to make room
for more text. It used a little loop of note more than 5 or 7
instructions to do this. But since the ACs (the registers) were
much faster than real memory, it loaded the sequence of
instructions into the ACs and did a JUMP 1 to transfer control to
the loop. The last instruction in the ACs did a JUMP xxx back to
the mainline code. Very cute.
+ Orthogonal instruction set. Any combination of things seemed to
work and work properly.
+ It was a very comfortable machine to program in assembly language.
+ Lots of software was available across the ARPA net for this beast.
John>> A more important question may be, what made the PDP 10's
John>> operating systems (TOPS-10/20, ITS,...) so wonderful? Why do
John>> people claim that U*IX, VMS, [and DOS :-)] are a step backwards?
I never really thought the OS was that great, in truth. It was the
obvious progenitor to VMS. Lots of stuff from TOPS-10 and -20 made it
into VMS (as did stuff from the -11 line)
Unix does a lot right. It is so easy to create subprocesses, do device
independent I/O, file redirection. Some of this was very tricky in
TOPS-10, others downright impossible.
+ File I/O is much cleaner in Unix than it was in TOPS-10. There was a
difference between devices so readang from a tty: was different than
+ The compile command was useful but make is definitely superior.
+ The command line interpreter was integral to the OS and could not be
+ But many programming languages were supported. I used:
+ Bliss-10 (my favorite)
+ Bliss-36 (this too)
at one time or another on DEC-10s.
+ DDT was really revolutionary. Break points, stepping through
instructions, examining memory. Very useful. I came from an
environment where post-mortem dump analysis was the most common way to
debug assembly language programs (ABEND).
+ And who can forget the great phone-book version of the manuals. I
foolishly pitched mine some time ago. Sigh.
Well, this ought to keep things going for a while.
Ex vitio alterius |Ed Hirgelt |e...@tss.com
sapiens emendat suum. |Teknekron Software Systems, Inc|
(Publilius Syrus) |Palo Alto, Ca. |
The first-level reaction can be taken as offensive because the Edsel was looked
down in its day; the -10 was looked down upon by PDP-11 Marketing because it
was the only part of DEC they couldn't pull snow jobs on. When PDP-11
marketroids overquoted some overbloated -11 systems, LCG pulled rank and
insisted that they had the right to cancel any -11 system order priced over
$250,000, etc. In all probability the customer was being cheated, getting
mini-computer performance for a mainframe price.
But all other parts of DEC looked *up* to the -10, not down, so the analogy
to the lowly Edsel isn't deserved.
On second look, Edsels are now high-priced collector's items, demanding far
more today than they were ever sold for originally. Presumably an Edsel
mechanic is a well-respected craftsman, so perhaps this actually *is* a
cj "Is an EDSEL like an EDVAC?" l
I like them because they had such an anachronistic instruction set (for 1987
when I last hacked one). You had 36-bit words, half-word instructions,
variable byte lengths, addressable registers, skip instructions and
operating system errors indicated by skipping. It was so very bizarre but
also remarkably easy to program (the entire instruction set fit on two
pages). I wrote some games and did numeric analysis homework in macro.
>A more important question may be, what made the PDP 10's operating systems
>(TOPS-10/20, ITS,...) so wonderful?
I only semi-liked the operating system. It had cool features: ESC
completion, '?' help, magical make, automatic memory allocation, and
replaceable shell. But just about everything else about it sucked. UNIX is
One thing that really made the it likeable was the enthusiastic
documentation available for it. People would take the MACRO-20 programmer's
manual (written by someone from Stamford who's name escapes me) for months
past the return date. Those without this book had to suffer with the crummy
DEC documentation. If you want to see real enthusiasm for this machine,
look at any classic LISP text.
Also real emacs was neat. The TECO source for it looked like garbage.
/* jha...@world.std.com (126.96.36.199) */ /* Joseph H. Allen */
What was great about the PDP-10 architecture?
1. 36 bits... we're talking _usable_ single-precision floating point here.
2. The instruction set architecture was incredibly powerful for human
programmers... I've heard many people who think that it is the best ISA
they've ever programmed for... I sure do. The combination of powerful
comparisons, masking instructions, skips, jumps and mapping the accumulators
(_please_ don't call them registers) as relocated (or virtual) memory
locations 0-17 was unmatched... _no_ ISA I've seen since let you do more
things in just a few instructions. Also, before cache memory, tight loops
could be copied into the ACs and executed from there (they could afterwards
too, but it hurt instead of helping).
3. Both operating systems (TOPS-10 and TENEX/TOPS-20... I can't speak for
ITS) had very efficient (if baroque) timesharing (as I recall, TOPS-10 was
more efficient but TENEX/TOPS-20 had more function, such as "copy on write"
Back as early as the late 1960's, long before virtual memory, the KA10 had
extensive provisions for sharing _one_ copy of reentrant code (loaded as the
"high" segment, starting at relocated location 400000) in memory among _all_
users. The first KA10 I used in the mid-1970's (when it was almost obsolete)
was about as powerful CPU-wise as a large Sun workstation (if that), yet it
would support thirty or forty users banging away at BASIC with only a few
antique 2314-clone (RP03?) disks. This late-1960's machine was this usable
even though it had refrigerator box sized 16K word _core_ banks and _no_
integrated circuits in the processor (only SSI in the memory controller)... it
was closer to an IBM 7094 (also 36 bits, by the way) than an IBM 360. Also
(for you architecture buffs), the KA10 was self-timed logic with only 50/60 Hz
power for an external clock. Unlike later PDP-10 processors (but like the
Cray Research Y-MP), the KA10 had to "shuffle" users around in memory to make
holes to put other users in... it did have relocation registers (so users
could move, unlike that silly "roll-in/roll-out" for IBM's MVT), but each of a
user's segments (one or two if using reentrant programming) had to be
contiguous in core.
What killed off the PDP-10 architecture?
1. 36 bits... not a power of 2! Seriously... the original addressing limit
of 2 ** 18 36-bit words was a hardship... though the DEC-10 extended this and
the DEC-20 could have. Multiprocessing wasn't implemented in TOPS-20 either.
Most important (IMHO) is #2.
2. The architecture was incredibly difficult to pipeline... mapping
accumulators as memory creates enormous potential structural and data hazards.
Also, I read an article once claiming that it was the hardest architecture for
a compiler to generate code for.
3. The focus of the operating systems was on timesharing... not batch, not
tapes, not data base, not client/server... just timesharing... as much or more
Even though the PDP-10 architecture couldn't handle the age of RISC and
pipelining (it's full of "N-th implementation artifacts", for many different
N), I still miss it.
John R. Grout | INTERNET: j-g...@uiuc.edu
The above 2 sentences describe the situation perfectly. If you don't
understand the statement, you probably deserve Unix.
Frank R. Borger - Physicist __ Internet: Fr...@rover.uchicago.edu
Michael Reese - Univ. of Chicago |___ Phone : 312-791-8075 fax : 567-7455
Center for Radiation Therapy | |_) _ Toots Shor's restaurant is so
| \|_) crowded nobody goes there any more.
"Birthplace of Softball" |_) Yogi Berra
> In article <1g5hqm...@master.cs.rose-hulman.edu> moo...@HYDRA.ROSE-HULMAN
> >>I don't know whether I should be amazed that the IIGS was still being made
> >>that the IIe is still being made.
> >Why?? As the proud owner of an Apple IIgs I can tell you that there is
> > nothing particularly out-of-date about the GS!!! The operating system
> >is quite nice...
> Many Elementary schools in the Southern US have computer educational
> programs, and they all run on the Apple ][+ and ][e. So, there is a sound
> reason for keeping the Apple ][ family of computers going; money...
While that is true to an extent, it is pretty obvious that Macint, er,
Apple, Inc. would like to "move on" to the Macintosh and get the whole
world running on 030 or better Macs. If Apple really wanted to keep the II
around for money, they might actually spend a red cent on marketing the
machine. I find it amazing that the Apple II continues to exist even
though it has been years since Apple has done anything to promote the
Even authorized Apple dealers are a joke. If they even have a GS on
display, and it's turned on, its usually running some cheesey 8-bit
software from 1983. My local authorized Apple dealer told me that the GS
had been dropped in 1990. They told me that I couldn't get a video overlay
card, there was going to be no ethernet card made, etc, etc, etc...
The support that the users are able to give the machine is incredible.
I'm not worried that all of the thousands of dollars I have wrapped up in
my IIs has gone to waste, because of that support. New software is still
being developed, support is still being given by user groups, the net,
etc, new hardware is being developed. And even a new third-party CPU is
under development. When the Avatar GS clone arrives, I'll probably drop
the money right there to get it. I love my IIs, I doubt that I'll ever
give them up.I don't think it's all to do with the machine, it's the
spirit behind it, the spirit that created it, and that which keeps it
going. The spirit that Apple has forgotten all about.
While I think that technically the Macs are, for the most part, neat
machines, I don't think I'd ever get one for several reasons. a) they're
closed boxes. The Apple II was the original open system, and everything
that I've used and liked has been an "open system."
b) in 2002, will Mac users be squealing that Apple has dropped
them in favor of the PowerPC, having the PowerPC run a mediocre emulation
that covers "older" Mac stuff only?
c) "What do you get when you cross Apple and IBM? IBM." I HATE
| C. Matthew Curtin ! "But I am the enlightened one, they are |
| P.O. Box 27081 ! but mere sheep, following each other in |
| Columbus, OH 43227-0081 ! the name of compatibility." -B. Heineman |
| 614/365-3272 (voice mail) ! Apple II Forever! |
So you had a very powerful machine that was as user-friendly (to the
techno-nerd sort of user) sitting at the Teletype as if it were a mini;
it was designed from the ground up for the interactive terminal user.
DEC could have advertised it as "not the kind of computer your banker
would buy" and the people who would buy it would understand perfectly.
"Ya can argue all ya wanna, but it's dif'rent than it was!"
"No it aint! But ya gotta know the territory!"
Meredith Willson: "The Music Man"
>While that is true to an extent, it is pretty obvious that Macint, er,
>Apple, Inc. would like to "move on" to the Macintosh and get the whole
>world running on 030 or better Macs. If Apple really wanted to keep the II
>around for money, they might actually spend a red cent on marketing the
>machine. I find it amazing that the Apple II continues to exist even
>though it has been years since Apple has done anything to promote the
Another good reason to keep the Apple II series alive is due to all the
specialized software/hardware that is available for the handicapped. Recently
I have been involved with helping a local school for handicapped children
(the Tammy Lynn Center) to get a few Apple II systems set up for use in their
classes, and eventhough I am an avid Apple supporter, I tried to find some
other system for TLC to use. Since I saw the 'light at the end of the tunnel'
for the Apple II line (didn't everyone), I hated for them to put money in a
'dead' platform. However, we could not find any other system (including
Apple's Mac) that had the breadth of specialized hardware/software needed
for this specialized market.
There are several vendors that sell large membrane style keyboards, special
breath-activated input devices, speech synthesizers, touch sensitive screens,
etc, that work with the Apple II line. When I called Apple and spoke to
their DISABILIES group, they suggested that we purchase the MAC/LC so we can
take advantage of all the Apple II peripherals/software UNTIL THE EQUIVALENT
SOFTWARE/HARDWARE IS AVAILABLE ON THE MAC LINE...!
So, for the sake of the handicapped, I hate to see the Apple IIgs being
shelved (the IIe can't be far behind...). As a former Apple II developer
(spent several years developing a 100K+ UCSD Pascal program for the IIe),
I have some very fond moments sitting in front of an Apple IIe (usually in
the wee-hours of the morning). But enough nostalga...
PS: If anyone is looking to get rid of a IIe or IIgs and would like to donate
it to TLC, contact me. I'll pay for shipping, you will get a nice tax
break, and you will be helping a very worth cause! Thanx.
Rodney Radford || Computer Graphics/Imaging
sas...@unx.sas.com || SAS Institute, Inc.
(919) 677-8000 x7703 || Cary, NC 27513
Gee, Macs have had slots since, oh the Mac II in 1986. You can get a 'closed'
Mac if you don't need expansion, but you have many other choices.
> b) in 2002, will Mac users be squealing that Apple has dropped
>them in favor of the PowerPC, having the PowerPC run a mediocre emulation
>that covers "older" Mac stuff only?
Nope. Because PowerPC IS Macintosh.
> c) "What do you get when you cross Apple and IBM? IBM."
Cheryl Lins Oberon Paladin li...@apple.com | Real miracles, reasonable prices.
>3. A real macro assembler. The stuff IBM had at the time claimed to be a
> macro assembler, but MACRO-10 was great fun and flexible. We were
> writing very structured assembly language in my days at the old
> CompuServ (before they became famous).
Yeah, I always though MACRO-32 was a pain to use.
>4. The architecture. Now how many different no-ops were there?
> + Orthogonal instruction set. Any combination of things seemed to
> work and work properly.
> + It was a very comfortable machine to program in assembly language.
Which reminds me. I always thought the instruction set was a good basic set.
It was always fairly easy to do what you wanted to do, and the result was
usually fairly logical code (unless you deliberately did something for
speed or whatever; recall HACK.MEM). The example I remember most is LSH
(logical shift) LSH AC,positive shifts the AC to the left and LSH AC,negative
shifts to the right. You could compute how many bits you wanted to shift
and not care the sign of the result; the same instruction accomplishes
the "desired" result. On VAX, you need to do an EXTZV to go right and
an ASH to go left.
>I never really thought the OS was that great, in truth. It was the
>obvious progenitor to VMS. Lots of stuff from TOPS-10 and -20 made it
>into VMS (as did stuff from the -11 line)
...Not much of the good stuff.
>Unix does a lot right. It is so easy to create subprocesses, do device
>independent I/O, file redirection. Some of this was very tricky in
>TOPS-10, others downright impossible.
TOPS-20 was a lot cleaner in subprocesses. File redirection is much nicer
in Un*x (one of the few things I'll think is nicer), but the -10 was more
"independent" than VMS if you pull out of RMS. For example, a program
which specifically uses terminal-specific instructions works under batch.
>+ File I/O is much cleaner in Unix than it was in TOPS-10. There was a
> difference between devices so readang from a tty: was different than
> from dsk:[1,2]foo.bar
Only sort of true: you can read from tty: similar to dsk:foo.bar[1,2] but
not the reverse. If you treated everything as file-structured, things
worked right. If you made assumptions about not being file-structured (which
was true in most cases since people were lazy (myself included)), it wasn't
>+ The command line interpreter was integral to the OS and could not be
But it did make things easier for the system mugger ("my delete command
doesn't work..."). We did allow foreign commands later, and that does
allow good abbreviations, which doesn't work as well in systems
that'll run anything there's a file for around (although the -20 did
that better as well).
And, speaking of DDT, I remember that I liked having the monitor rather
than the linker usually write the .EXE file. Two examples of where this
was useful follow. One was an encryption module for .EXEs which could use
the linker's expertise at relocating code, but then run briefly to encrypt
the linked-in program and save. In the VMS case, it's difficult to figure
where to put the self-decrpytion code since you can't easily figure out where
any given random program doesn't have address space (I cheated and used
page 0 when I did this). A second was a diagnostic patch wherein I generated
an undefined symbol list which I followed in DDT so I could patch in something
to track references to the locations involved (I can't remember why, but
address break would work there, probably because it was a range of addresses
or something like that and we didn't have the Jupiter, obviously).
Oh well, enough for now.-kby
(Ten cool points to whoever figures out what that code does.
Twenty to whoever knows where I found it :-) )
Dave "Budding TECO hacker" Brown
Who needs command recognition with two-character command names:-)?-kby
Actually, TOPS-10 has a wonderous tape management system. It sure beats
"dd" :-). It also has a rather nifty database available (1022) that is
very pleasant, that was a third party product.
Also remember that, before networks were common, the only way to co-operate
on computers in a modern manner was to share a machine, so concentrating
on timesharing wasn't necessarily a bad idea. If you had people all over the
place who had to share in realtime nice big PDP-10(s) were a very good choice.
In fact that's pretty much how Compuserve operates to this day.
-dave fetrow- INTERNET: fet...@biostat.washington.edu
< 9 bits> < 4 >1 < 4> < 18 bits >
| Op code | reg|@| Xr | Displacement |
At a time when most real programmers used assembly language, the PDP-10 instruction
set was a joy to program. It is still the cleanest architecture I know.
The address space was 256K words or over a megabyte (Huge for the 1970s). Every
instruction could directly address any location in memory. Every instruction evaluated
its effective address before looking at the opcode.
The PDP-10 was invented at a time when it was not clear if the 6-bit code used by IBM
in the 7090 and 7094 would be replace by 7-bit ASCII or maybe 8-bit EBCDIC. The PDP-10
included the ability to easily process strings of bytes with an arbitrary byte size.
This abily may sound silly in 1992, when everyone knows that a character is 8-bits,
however, it was very useful in the 1970s when 7 and 9 track tapes came in all sorts
PDP-10 COBOL also supported 6-bit characters in data files. For applications with a
large data base (tens of megabytes), that saved hundreds of thousands of dollars over
I have been an assembly language programmer for 32 years. I have programmed everything
from 7090s and 1401s to the latest RISCs. I have not found a machine that is as easy
to program at the assembly langauge level as the PDP-10. [[The 1401 comes close but
the address space is limited to 16000 characters. Yes, 16000 not 16384 because the
1401 was a decimal machine.]]
When the 18-bit address space became a problem, the architecture was messed up in
various ways. It was still one of the cleanest extended addressing scheme ever
invented. The loss of the ability to pack two pointers into a register or a word
is memory was a big loss. LISP people loved the PDP-10 because it was just so
natural to build linked lists as:
| Back pointer | Forward Pointer |
| A more important question may be, what made the PDP 10's operating systems
| (TOPS-10/20, ITS,...) so wonderful? Why do people claim that U*IX, VMS, [and
| DOS :-)] are a step backwards?
PDP-10s were designed to be BIG machines which competed with mainframes but were
engineered for real people to use them. TOPS-20 was designed with virtual memory
as a key component of the operating system from day 1.
While PDP-10 had a small amount of main memory and disk by today's standards (for
example, the "terminal" I am using at the moment has 32 Mb of memory and 400 Mb
of disk and most of the work is done by a server someplace else in the network) PDP-10
were huge for their day. Things like on-line help files, command recognition, and so
on were possible. A great deal of the machines resources were spent on the user
UN*X and other PDP-11 operating systems were designed to fit into much smaller machines.
DOS was designed to work in a very constrained environment. There are practical and
business reasons why one wants to be compatible with existing practice and the
installed base. Compatibility comes at a cost.
Most of the great leaps forward in computing were compatible with nothing and
broke new ground. For example, the Macintosh. The PDP-10 (the PDP-6 really) was
a great leap forward 20 years before the Mac.
| Disclaimer: I don't speak for DEC and I have never used a PDP-10 (do like
| PDP-8s though....)
Well, you have to use a PDP-10 to love one. And unless you are the type of person
who would love to drive around town in a red 1968 Firebird or GTO, you probably had
to use one when they were the new machine on the block.
Donald A. Lewine (508) 870-9008 Voice
Data General Corporation (508) 366-0750 FAX
4400 Computer Drive. MS D112A
Westboro, MA 01580 U.S.A.
Since nobody else has said this I will and will endure the flanes. In
the early 80s DEC had 3 choices:
1) kill the 10/20
2) kill the VAX
3) Continue to develop and market both, acknowledging that they were
competing for exactly the same customers; that every sale of one was a
lost sale for the other.
Which would you choose? Which would you choose if you were a
Not true. ASH with a negative shift count will go right. I do it all the time.
Roger Ivie "My God! That computer is full of Pentium!
iv...@cc.usu.edu It's a wonder that you haven't been turned
>Gee, Macs have had slots since, oh the Mac II in 1986. You can get a 'closed'
>Mac if you don't need expansion, but you have many other choices.
Gee, the Apple ][ had slots from day 1, the Mac got them *years* later.
You can open an Apple ][ with your bare hands. To open an early Mac (any
model before the Mac SE), you need a special screwdriver that's hard to
obtain and the screws are carefully hidden.
With your Apple ][ you got the schematics, and a commented assembly
listing of the monitor. When the Mac appeared, no documentation of MacOS
was available. Initially, software development on a Mac was impossible, as
there were no tools. No compilers, no assemblers, no debuggers, no nothing.
(Apple used a cross-development kit running on a Lisa).
Wonder why the Mac had such a hard time competing against the ][ ?
>Cheryl Lins Oberon Paladin li...@apple.com | Real miracles, reasonable prices.
Ah, that explains a lot. ^^^^^
Hans Mulder ha...@cs.kun.nl
Yes, it goes right, but it doesn't do a LSH (it shifts the sign bit in).-kby
The simple, little Apple IIe, a general purpose computer with a lot of slots,
has many good things going for it. There are thousands of specialized boards
available for it that help automate jobs in the laboratory, the classroom,
and in industry. It is easy to write control programs from Applesoft BASIC,
a language known by many non-computer industry professionals.
The Macintosh has little of this hardware available, and requires the casual
programmer to climb a steep learning curve before they can produce anything
meaningful. It is a canned application machine.
The Apple IIe is a programmable tool that can do an amazing number of things,
considering it's age.
>> b) in 2002, will Mac users be squealing that Apple has dropped
>>them in favor of the PowerPC, having the PowerPC run a mediocre emulation
>>that covers "older" Mac stuff only?
>Nope. Because PowerPC IS Macintosh.
PowerPC is UNIX with a Mac emulation on top.
Taligent is not compatible with MacOS.
>> c) "What do you get when you cross Apple and IBM? IBM."
That remains to be seen.
But this is the problem of "all your eggs in one basket" since there were plans
to make the TT-10 and the Jupiter, and all in-between. Moreover, the plan
to make the VAX largely turns its back on a sizeable software investment in
the -10/20 stuff. So, what do you tell the stockholders:
We want to market this new machine that largely overlaps our existing product.
Moreover, we'll have to write new software mostly from the ground up, yet
this software will not be distinctively different from where we are today with
the existing software for the existing machine. And as the design for both
the existing machine and the new design migrate upward and downware, they
will both find applications in the same market and will likely be a duplication
of effort, redundant products, etc.
In fact what DEC always does, is not tell the tale in the simplistic way
you presented the choices. They simply lied about how much better the VAX
would be, and proceeded to backstab the -10/20 the same way the -11 did in
I find it gratifying in recent years to find many of the -11 supporters seeing
their favorite models disappearing, such as the 11/70, etc., as they see the
-11 relegated to an also-ran product, destined to be scrapped entirely shortly.
In fact, I find it *more* gratifying to see the VAX being edged out in much the
same way as the Alpha takes over.
The point is that DEC often uses "shoot-yourself-in-the-foot" as part of its
marketing process, as the short-term goals and hidden agendas of individuals
guide the marketing strategy instead of someone high up enough to guide the
company's overall best interests, etc. (The DEC pyramid management and
middle-management conspiracy has been discussed elsewhere, etc.)
Will I come back here in 6 years to hear about how the Alphans are being
edged out of the company as the Betans take over?
cj "New often does not mean better" l
> >While I think that technically the Macs are, for the most part, neat
> >machines, I don't think I'd ever get one for several reasons. a) they're
> >closed boxes. The Apple II was the original open system, and everything
> >that I've used and liked has been an "open system."
> Gee, Macs have had slots since, oh the Mac II in 1986. You can get a 'closed'
> Mac if you don't need expansion, but you have many other choices.
Gee, uh, there's more to an "open system" than just slots. You still can't
buy all the pieces for a Macintosh and put it together yourself. Why do
think that in order to clone a Mac, you need to have the ROMs from a real
Macintosh? They just don't come much more proprietary than the Mac...
How about the JFFO instruction? I miss the byte pointer stuff.
so - it was fun. more fun than vms, unix, ms-dos, windows, x, nt, or even
os/2! haw haw haw.
"This is not your father's Oldsmobile"
Peter da Silva. <pe...@sugar.neosoft.com>.
`-_-' Oletko halannut suttasi t än ä än?
Tarjoilija, t äm ä ateria el ä ä viel ä.
Argh! Why does EVERYONE ignore the Amiga. :)
>| C. Matthew Curtin ! "But I am the enlightened one, they are |
And you were doing so well until this point.
The Mac was not a great leap forward. The Xerox 1100 and Star were a great leap
forward... the Mac was a shoddy knock off, and its advantage is social (cheap,
and Apple kept the developers sticking to their interface model) not technical.
You seem to think the VAX was a completely standalone effort. It wasn't.
The 11/780 was basically a replacement for the pdp11/70, which had taken
the pdp11 line to about as large as it could practically get (as the 11/74
effort had largely proven). It was always intended that users of large
pdp11 systems would migrate to VAX, not users of the -10 line. (This is
1977 I'm talking about; not the lead-up to the cancelling of Jupiter.)
I don't think the VAX would have been as successful if the migration from
the pdp11 was there. RSX code had been known to just run on the VAX.
RSTS code took a little more persuasion, but it wasn't hard (and the VAX
BASIC compiler has a few bits of compatibility stuff for RSTS syscalls).
Other languages under RSTS were generally ports of RSX code and just worked.
Having ported quite a lot of code from RSTS/E to VMS, I can report that it's
not too difficult, and most of the problems were to do with differences in
the third party "database" software libraries on the two platforms.
I know of one outfit who were running an 11/70 with RSX, and were intending
to migrate to a VAX. Then the system disk on the 11/70 died. They worked
over a weekend to recompile the system onto the VAX (very few changes
required), mounted the 11/70's data disks on the VAX and continued as if
nothing had happened. There's a fair chance that just renaming the .TSKs
to .EXE and running under VAX RSX would have worked too, but they didn't
for some reason (the target VAX may have been one without compatibility
mode -- this was only a few years ago).
While the VAX was intended as a follow on from the pdp11 line, the
architecture allowed for a potentially bigger machine than the -10
architecture did, and the only limiting factor in building bigger and
bigger VAXes was the technology available. Compare this to the fact the
20xx line were already requiring tweaks to the architecture to support
bigger address spaces. The pdp11 line had already been through these kinds
of tweaks to get the physical address space up (16, 18 and 22 bit physical
addresses, each requiring different memory management code) and a couple of
half-hearted attempts to increase virtual memory (split I&D, supervisor
mode), and I can quite understand why DEC didn't really want to carry this
on into the 80s with the -10 line.
Remember that the -10 was a bunch of architectural and engineering tradeoffs
that were valid in the 60s, whereas the VAX's tradeoffs were valid in the
70s. Neither are valid now.
I suspect that when there was demand for a bigger VAX, DEC found themselves
facing the prospect of supporting two "big" machines, and had to make the
choice between them.
> I find it gratifying in recent years to find many of the -11 supporters seeing
> their favorite models disappearing, such as the 11/70, etc., as they see the
> -11 relegated to an also-ran product, destined to be scrapped entirely shortly.
> In fact, I find it *more* gratifying to see the VAX being edged out in much the
> same way as the Alpha takes over.
The 11/70 is 17 years old for chrissakes. Sure we get nostalgic about 'em.
It reigned as the most powerful genuine pdp11 until the 11/84, and even then
could outrun an 11/84 quite comfortably on real applications after you
factored in the I/O; it could be argued that it wasn't completely obsoleted
until 1990 with the release of the 11/94 (which *still* only has Unibus for
I/O). 15 years is a bloody long time for one model, and there's still more
of 'em chugging along just fine than any kind of -10.
I'm not sure when DEC stopped supplying 11/70s, but I'd guess that it was
after the 11/84 was released in '84/'85.
As for VAXes: it's been known for a while that the VAX is going to lose the
MIPS/$ race, but current VAXes are in the bunch with the not-quite-bleeding-
edge RISCs at this time even so. DEC has learned a thing or two from the
previous "upgrade" debacles; Alphas will run VMS just fine (or even unix if
you really must 8-). No-one's being left out in the cold like they were in
the -10 to VAX "migration".
> Will I come back here in 6 years to hear about how the Alphans are being
> edged out of the company as the Betans take over?
I doubt it. Apart from obvious losers, major architectures last a long
time at DEC, eg the pdp11 at 22 years (and still going strong), the PDP-8
at 25? 26? years, although now discontinued. The -6/-10 lasted nearly 20
years, and the VAX will still be kicking on its 20th birthday in 1997.
The Alpha is *not* an obvious loser. It'll be around well into the next
century, unless DEC goes down the tubes, in which case there won't be any
Betans. There's also no competition within DEC at this stage -- the VAX
crew have been co-operating closely with the Alpha lot -- the NVAX+ chip
in the VAX 7000 & 10000 is pin compatible with the 21064 for example.
>While the VAX was intended as a follow on from the pdp11 line, the
>architecture allowed for a potentially bigger machine than the -10
>architecture did, and the only limiting factor in building bigger and
>bigger VAXes was the technology available. Compare this to the fact the
>20xx line were already requiring tweaks to the architecture to support
>bigger address spaces. The pdp11 line had already been through these kinds
>of tweaks to get the physical address space up (16, 18 and 22 bit physical
>addresses, each requiring different memory management code) and a couple of
>half-hearted attempts to increase virtual memory (split I&D, supervisor
>mode), and I can quite understand why DEC didn't really want to carry this
>on into the 80s with the -10 line.
While I'll admit that going from KA/KI architecture to KL architecture was
pretty crocky in terms of extended addressing (although I still feel they
did a lot better than 8086/8 to 80286/80386 transition), going beyond the
actual KL architecture should not have been necessary for a while. Although
the KL itself was 32 sections (512 pages of 1024 36-bit words), the KL
architecture spec was 30-bit addressing (4096 sections). I don't know
about the -20, but I know when I did the -10 stuff I made the monitor such
that going to 4096 sections was at least theoretically possible from the
23-bit 32-section addressing. The big change was getting out of section 0;
after that it was pretty much the same. I think there were "only a couple"
of places where 23-bit maximum size to addresses (or 14-bit page numbers or
5-bit section numbers, whichever you prefer) were taken advantage of :-).-kby
Probably because there is no one here from Commodore claiming that the
Amiga is an open system.
Also, this is a group where many people would consider the Mac to be
an open cistern.
>The 11/70 is 17 years old for chrissakes. Sure we get nostalgic about 'em.
>It reigned as the most powerful genuine pdp11 until the 11/84, and even then
>could outrun an 11/84 quite comfortably on real applications after you
>factored in the I/O; it could be argued that it wasn't completely obsoleted
>until 1990 with the release of the 11/94 (which *still* only has Unibus for
>I/O). 15 years is a bloody long time for one model, and there's still more
>of 'em chugging along just fine than any kind of -10.
>I'm not sure when DEC stopped supplying 11/70s, but I'd guess that it was
>after the 11/84 was released in '84/'85.
The death of the 11/70, in the US, was an FCC ruling that deemed it to
Of course, there is still the 11/94 and the MicroPDP-11.
>As for VAXes: it's been known for a while that the VAX is going to lose the
>MIPS/$ race, but current VAXes are in the bunch with the not-quite-bleeding-
>edge RISCs at this time even so. DEC has learned a thing or two from the
>previous "upgrade" debacles; Alphas will run VMS just fine (or even unix if
>you really must 8-). No-one's being left out in the cold like they were in
>the -10 to VAX "migration".
>I doubt it. Apart from obvious losers, major architectures last a long
>time at DEC, eg the pdp11 at 22 years (and still going strong), the PDP-8
>at 25? 26? years, although now discontinued. The -6/-10 lasted nearly 20
>years, and the VAX will still be kicking on its 20th birthday in 1997.
>The Alpha is *not* an obvious loser. It'll be around well into the next
>century, unless DEC goes down the tubes, in which case there won't be any
>Betans. There's also no competition within DEC at this stage -- the VAX
>crew have been co-operating closely with the Alpha lot -- the NVAX+ chip
>in the VAX 7000 & 10000 is pin compatible with the 21064 for example.
A friend of mine, who works for DEC ( in "The Mill" and other sites ), tells
me that they used to move from 10's to Vaxs interdepartmently by pulling the
boards out of 10's and stuffing in a set of Vax boards.
You're talking about PowerOpen. PowerPC is the name of a series of computer
chips. The 601 is the first (reported in EE Times, MacWeek, etc). PowerOpen
is a Unix implementation running on the PowerPC chip. The PowerPC based
Macintoshes will not be based on Unix. There is 68020 emulation at the
bottom. This was all publicly announced at the WWDC in May, '92.
Cheryl Lins Oberon Paladin and Witch-in-training li...@apple.com
"Capitalism has failed." -- Consolidated
I'm amazed DEC put the split-I&D mode in the -11, because so far as I can
tell they didn't provide any O/S support for it until a pretty late release
of RSX-11. When I was using the -11 heavily, the only O/S that let you USE
split-I&D seemed to be 2BSD and derivitives.
> The 11/70 is 17 years old for chrissakes. [...]
> 15 years is a bloody long time for one model, and there's still more
> of 'em chugging along just fine than any kind of -10.
Yes, remember DEC is planning on bringing the Alpha replacement onto the
stage 15 years from now, and that's a product line... not just a model.
I mean, in terms of execution environment a PDP-11 is just a little less
desirable than an 80286. If you want a perfect sensation of what hacking
on an -11 is like, run an old version of Coherent or Minix that only
supports small model.
And I can understand Charles Lasner's devotion to the systems he used when
he was just getting started, but 4K of 12 bit words on the -8 was painful
back in the late '70s when I was using one, unless you were doing some
really trivial stuff. By the end my pocket calculator had nearly as much
Crowing that the -11 users are finally getting their comeuppance is sad:
I'm an -11 fan and even I think DEC has kept it alive (or been forced to
keep it alive) way longer than I ever expected.
> I don't think it's all to do with the machine, it's the
> spirit behind it, the spirit that created it, and that which keeps it
> going. The spirit that Apple has forgotten all about.
The "spirit" behind brain-dead Prodos... Having to use PREFIX..
not having a CLI... Incompatibility with almost every ported program
because there's no command line ability... [Although, I haven't
been keeping up with the II world, and I think there are some shells
out there, now.]
Ug. The best thing that could happen to the II line, really,
is an entirely new operating system. (What was that about unix on a GS?)
Forget Prodos. And why does the Finder have to look exactly like a
Mac's finder? Don't we all realize that GUI's are usele--err...
> c) "What do you get when you cross Apple and IBM? IBM." I HATE
Uh, you really mean 80x386 architechure. I'd have to say I hate
it too, but it's cheap and it runs Linux. :)
j...@vector0.SAC.CA.US Life is like a kiwi.
>Which would you choose?
>Which would you choose if you were a stockholder?
Rich Alderson 'I wish life was not so short,' he thought. 'Languages take
such a time, and so do all the things one wants to know about.'
--J. R. R. Tolkien,
alde...@leland.stanford.edu _The Lost Road_
>Crowing that the -11 users are finally getting their comeuppance is sad:
>I'm an -11 fan and even I think DEC has kept it alive (or been forced to
>keep it alive) way longer than I ever expected.
We were checking used prices the other day and noticed that the used
price of the vax8650 had fallen below that of the used high end pdp-11.
I'm waiting for my favorite PDP-11 groupie to call me up and give me
some crap about that (I committed the foul deed of converting to, learning,
managing, and advocating Vaxes). Oh well - I'll just ask him when he thinks
he can port WNT to the 11!
are we going to be able to get a KL emulator before 1999?
"you can hack anything you want - with TECO and DDT!"
There seems to be two "rumors" floating around.
1. PowerOpen, which will run on the PowerPC, will be a base OS consisting
of a mixture of AIX and A/UX, with a MacOS emulation and Multifinder
running on top.
2. Taligent, which will run on the PowerPC, will be released without any
application compatibility to any other OS.
Why would Apple base an OS on a processor chip that they no longer support
in their product line?