[Ada_list] [Ada_List] more on the poly

Hi ADAers

In the spirit of adding to what is available online about the Poly computer, I'm
posting another piece on the Poly, written and sent to me by Alec Utting (posted
with his permission).

Also, I noticed that the web interface is not showing the full text of the last
post in this thread.  I don't know why this is.  If anyone is looking for it,
Perce Harpham's text is available in the mailman list archives, at
http://list.waikato.ac.nz/pipermail/ada_list/2007-November/002116.html

Melanie


Poly Computer Software

Alec Utting
November 2007

I worked for Progeni, Lower Hutt as their Systems Development Manager. I had been
a teacher in both primary and secondary schools previously.

The first that I heard about the Polywog Computer development was just prior to a
visit to the USA in Oct-Nov 1980 when Perce Harpham told me that he hoped to do
the demonstration software for the system with the Education Department and the
Wellington Polytech. While in New York, I visited a children’s TV studio that was
creating TV using computers.

On my return, I had some time with Kevin Hearle from the Education Department and
Neil Scott from the Polytech show me the work that they had already done on the
project, before I worked with a group of teachers who had been brought together
by the Education Department to define programs that they would like to see
running on the Polywog. I provided the technical interface with them and kept
them with feasible limits.

At that stage the Polytech staff had designed a prototype had started
construction of it using a 6809 8 bit processor. They had specific ideas about
how the system would be configured and the special features that they wanted in
the software, especially the graphics. They had already decided that they would
use Teletext for the screen display. They had made no decision about the
operating system or any other software. They were using the Flex operating system
meanwhile for testing. I then went away on holiday hoping that they would be more
advanced when I returned in late January.

I was disappointed. Although I was only implementing the demonstration software,
I made the decision then to use Flex as the basis of the system, with extensions
added. We would also use the Flex BASIC interpreter with extensions for the
evaluation.

During February, my basic task was to define the extensions that needed to be
made to BASIC in order to have the computer able to do what had been specified,
and to specify the software interrupts which were needed, and thus alert the
design team to any further modifications needed to the board in order to carry
these out. These were put in the document Specifications for Polywog Operating
System 25th Feb 1981
It also meant that I had to define all the conventions and descriptors that would
be used on the Poly.

During this time I had some programmers working to define the programs from the
teacher’s specifications.

By this time the launch and teacher’s training course had been set for the first
week in the May school holidays - about 10 weeks away. … and the Polytech team
had still not worked out a critical implementation path in order to meet that
deadline and have computers available to my team for testing. Once again I had to
step in and work out the deadlines that the Polytech hardware and software teams
would need to meet. Time was tight and they had to change their priorities if
they were to meet these deadlines.

My problem was that I had to develop the software without any computers for
testing on until a few weeks before launch, and to use a BASIC language which
would not have all the extensions ready by then.

We immediately added multiple parameters to the USR command so that we could
access the software interrupts using this command.

In order to create the necessary code, I had the programmers write the code as if
the new commands  had been made.
E.g.  the LDES$ command which defined the boundary points of an area would be
written exactly as we had defined it. 100 A$ =
LDES$((0,0),(100,0),(100,100),(0,100),(0,0))

We then used a product (PROGENI TOOLS) which I had developed earlier for
generating COBOL code, to generate the exact BASIC code necessary to implement
that command. This code was only a single line in some cases, but in others
created dozens of lines of code.

When the first prototype computers arrived on Anzac Day 1981, we fed this
generated code into them for testing. The next few weeks were hectic with long
hours as the computers were all standalone and not yet been networked, and many
of the commands did not work as envisaged.

I did not see a networked system until midnight two evenings before the course
started in Palmerston North in mid May - and never had one available for testing.
The Menu system was based around random access files on the system disc.

On arrival in Palmerston North late afternoon the following day, I spent the
whole night testing the programs which had been given to me as I left - however,
I only had the prototype standalone computers to use.

After no sleep, I went immediately to the Training College where the course was
being held, gave the tested programs to others to put on the network, and had a
microphone thrust in my hand to tell everyone about the wonderful software. They
would be able to see it after morning tea.

I then went and tried out the networked system, and to my horror found that the
random files did not allow simultaneous access by multiple users. As I had never
had the networked system available to me, I had never been able to test the menu
system with multiple users. I could not get changes made to the file handling
system then, so I sent a message back to the course to delay any access to the
computers for two hours, and I sat down and rewrote the menu program putting all
the program selection options into the code. Amazingly, we never went back to the
original menu and stayed with my hastily written one throughout the life of the
Poly.


>On Nov 12, 2007, at 16:45 PM Melanie Swalwell wrote:
>Dear ADAers,
>
>Herewith another addition to our long running thread on the Poly
>computer.  Some of you may have seen a version of this article last year
>in the Dominion Post (have searched for it in various databases but
>haven't turned up the date yet).  I'm posting it here so it is easily
>available online.  Posted with the permission of the author, Perce
>Harpham (ex-Managing Director of Progeni).
>
>I hope to get info on some of the software that was written for the Poly
>entered into the Early NZ Software Database.  If you have missed my
>emails on this project, please take a look at
>http://nztronix.org.nz/main.php and, of course, add any titles about
>which you know something!
>
>Melanie