Imlac PDS-1, was Re: terminal inventiveness (was Re: an odd question

From: Bob Shannon <bshannon_at_tiac.net>
Date: Sat Aug 18 11:09:58 2001

Paul Williams wrote:

> [jeopardectomy performed]
>
> Bob Shannon wrote:
> >
> > Chuck McManis wrote:
> > >
> > > Lets define a TERMINAL to be a cathode ray tube that displays
> > > characters and a keyboard input unit that communicates through
> > > a serial channel (can be coax, RS-232, current loop, etc.)
> > >
> > > 1) What was the first example of a TERMINAL (See above)?
> > >
> > > 2) When was a processor (micro, bit-slice, whatever) first
> > > DEDICATED to operating a terminal?
> > >
> > > 3) What is the first terminal that could manipulate its own
> > > buffer memory in response to a "control" code?
> > >
> > > 4) What is the first terminal that could display graphics and
> > > text? (either alternately or simultaneously)
> >
> > I'll bet on the Imlac PDS-1.
>
> What year? Which of the four questions does this answer?

All of them to some degree!

Imlac PDS-1's were the first-ever 'smart terminal', and the first
networkable graphics terminals, developed during the mid-1960's, and
first introduced circa 1968. So the PDS-1 was commercially available
around the same time as the first glass teletypes (question #1).

For question #2 through #4...

The PDS-1 is a dual processor 16 bit minicomputers implemented with core
memory that was shared with a display processor. The display processor
updated a CRT display using vector (stroke generated) characters. The
characters themselves were totally defined by software so even the
'font' was programmable.

In Imlac documentation, the minicomputer is sometimes refered to as the
display list processor. This 'main' CPU also took care communication
with a host system. This processor did not (normally) have general
purpose I/O capability, and was 'computationally challanged' as compared
to say a Nova or a HP 21xx of that same era so this is clearly a
dedicated 'terminal control' processor, although it can run modest
application programs as well.

Some application programs were written that emulated other types of
displays from simple glass TTY's to other graphical displays. This is a
clear example of the functionality of question #3.

Lastly, the PDS-1 generated all its text characters by short vectors, so
mixed graphics and text was what the Imlac design was all about. In
fact some of the first (the first) networked graphics protocalls were
developed for the PDS-1.

Imlacs are very interesting and historically signifigant machines, I'd
expect them to be better known and more widely collected than they seem
to be. I can only assume that this is due to their being somewhat
rare. While I don't have any information on the numbers produced, its
clear from looking at the hardware that the PDS-1's were essentially
hand-made machines, targeted to a niche application in their day.

Also, the PDS-1 is probably THE ONLY 'terminal' that came with an
optional switches and lights programmers console. Early short-vector
Imlac consoles look very much like a prop from the original Star Trek TV
show, with a hint of the Cyber 205 console added for good measure.

I have an early model Imlac PDS-1. The minicomputer section is up and
running, and I'm working on the display and display processor section.
I'm sure it will be fully operational soon.

I first saw an Imlac at VCF East, and was impressed enough to 'have to
have one'!
Now that I'm becomming more familiar with their design and role in
shaping what we know as the 'intellegent terminal' today, I've decided
that these rare and special machines deserve preservation beyond what
little original Imlac hardware exists.

I am aware of an Imlac emulator program under development, but no
emulation on a modern raster-based display will quite match the beauty
of a true vector based stroke-generated display. To address this, I've
been discussing a re-implementation of the PDS-1 with an ASIC chip
designer where I work. He's convinced me that we can reduce the PDS-1
to a small handfull of Xilinx chips at a very affordable cost.

My plan would be to build a replica of the early model PDS-1 programmers
front panel that contained all the processors and memory (the original
Imlac was a peice of furniture with the electronics hidden in the back
under a small desk).

This Re-Imlac would use a real X/Y vector based display, just like the
original.
Received on Sat Aug 18 2001 - 11:09:58 BST