> Jim Stewart wrote:
>> I disagree. I installed a fully functional 8-user timeshare system
>> on a PDP 8 with 12k words of memory back in 1972. The early
>> Unix and Decsystem 10's were amazingly efficient for the resources
>> available then.
> I don't know the relative performance of a PDP 8 or GE system, but I
> can't help but wonder that a computer of 1972 was much more powerful
> than of 1962 given the massive progress in electronics.
The issue isn't power. The PDP-8 is not a powerful computer. It was
what high schools bought to drop kids on.
The issue is memory management. The DAT box was a device that did
memory mapping, so that the addresses the _application_ saw weren't
necessarily the same as the hardware addresses.
You can, indeed, do real timesharing without memory management, but
it's hard to do, and you invariably have applications stepping on one
another when users write bad code (as high school users probably
will). But TSS/8, the time sharing system for the PDP-8, did amazing
things considering the hardware it had to work with was very stripped
> Also, I'm not sure if the PDP 8 could do batch processing -- handling
> high volumes of punched cards, mag tapes, high speed printing, at the
> same time it was handling time sharing. Handling both batch and on-
> line was a desired quality but not always possible.
No. The PDP-8 was a toy computer by the standards of the day. The point
is that if such a toy can do real timesharing without memory management,
it shows that memory management is not impossible to live without.
> Also, I believe the simultaneous GE users numbered about 30 or more
> (judging by the size of the rooms they showed) as opposed to 8.
By the early seventies, I was seeing a couple hundred people logged
onto a single machine running GECOS. Of course, at the time a lot of
the actual front end processing of the timesharing system was being
done by the terminal or channel controller.
>> I've always been amazed that not a single science fiction writer *got*
>> the internet. All of the SF saw the future as monolithic central
> The "Internet" is a network of networks, far more sophisticated than
> mere timesharing.
I'd dispute that too. It's _different_ and not necessarily more
sophisticated. In some ways, the loose coupling of the internet is a
big step down from a centrally-administered system.
If you had told me in 1975 that everybody would have a computer on
their own desks in 20 years, I might have believed it. But if you had
told me that each one of them would have to do their own backups and
constantly repatch their flaky operating system, I would have laughed
> In the early 1960s, it was considered quite amazing to dial into a
> computer and have it run BASIC programs or look up stuff from a data
> file. The interface in those days was quite slow--10-15 characters
> per second on a simple typewriter, and of course it was all by very
> curt command prompts. If you were accessing a data base computer, you
> would type in a single command followed by coded search arguments
> carefully coded, e.g. SEATBL1JDSX4355. If any of the arguments were
> wrong you got back a bland "INVALID COMMAND" message and had to
> figure it out for yourself. That's all computers in the early 1960s
> had the storage and speed to support, very bare bones.
That's all computers STILL do. It's just that today you are seeing a glossy
front end on top of all that stuff. In some cases, the glossy front end is
a great thing because it allows people to use computers without really
understanding how they work. In other cases, it is a terrible thing
because it means most people using computers have no understanding how they
I do not consider the development of glossy front ends and cheesy GUIs to
necessarily be a technological advancement.
"C'est un Nagra. C'est suisse, et tres, tres precis."