June 2010 Archives

Popular 1960's British computer still putting on a show

| No Comments
| More
TNMOC volunteer Peter Onion reports on the success of one of Britain's iconic machines, both in industry and as one of the larger, working exhibits at the National Museum of Computing.

This coming weekend at the Vintage Computer Festival at TNMOC, I will as usual be found tending the Museum's oldest original operational computer, an Elliott 803. In fact this weekend, I'll be giving formal talks on it too.

The Elliott 803 was a very popular machine in its day with over 200 being sold between 1960 and 1965 (in 2011 we are hoping to arrange a celebration for the 50th anniversary of the first shipment of an 803B). The 803 was used in a wide variety of business, industrial, scientific and military applications and luckily a list of Elliott's 803 customers has survived, which included:

  • The GPO who used an 803 at their Goonhilly Downs satellite earth station to calculate the path of the first communications satellite, Telstar enabling reception of the first transatlantic satellite TV pictures in 1962.
  • Corah Knitware in Leicester who managed their orders and production on a pair of 803s.
  • A poultry farm in Yorkshire who used their 803 to analyse each hen's egg production to aid their chicken breeding program.
  • Several 803s were shipped overseas with some eventually finding their way to Russia and other Eastern Block countries. Some even went to the USA as part of industrial process control systems!

The TNMOC Elliott 803B was manufactured in 1962 and apart from a period of about 12 years in a barn has been in regular use ever since. It was the first "large" machine to be installed at TNMOC.

Almost every week, I have the pleasure of hearing TNMOC visitors say: "I used one of these when I worked at .......". Some of them were employees of Elliott Automation who installed or maintained 803s and some were students who used an 803 at university to analyse their research.

Of course most of our visitors have never even heard of the Elliott 803, so it is always a pleasure to give them a potted history of the machine. For many of them it is the largest computer they have ever seen close-up - that is until they move on past the 803 and are confronted by TNMOC's huge ICL 2966 installation.

More technically minded visitors can get treated to an in-depth discussions of the finer points if the 803's design. Once explained, its unfamiliar serial architecture and seemingly bizarre 39 bit word length are often described as "clever" or "ingenuous". But those were the days before the byte ruled over machine word sizes, and when the lower gate count and smaller physical size of a serial computer made a machine affordable to many more potential customers. The 803 was the last of Elliott's serial machines with all its successors using parallel architectures.

The ultimate demonstration of our 803 is to see and hear it running its Algol 60 compiler. The compiler was produced by a small team of only three programmers led by Tony Hoare (now Professor Sir Tony Hoare), and the availability of such an advanced high level language made a major contribution to the commercial success of the Elliott 803.

Peter Onion leads the Elliott 803 restoration team at the National Museum of Computing. You can find him at the VCF this weekend in the Large Systems Gallery where he will also be giving talks on the 803 at 12.15 and 15.15 each day.

Restoration: Impossible

| No Comments
| More

With Britain's first-ever Vintage Computer Festival at TNMOC, Bletchley Park on 19-20 June, TNMOC volunteer Delwyn Holroyd speculates that over the next 30 years a computing museum may be forced to become mainly a computer museum.

Whilst hammering some new bearings onto a drive motor shaft the other week, a couple of thoughts crossed my mind. Firstly, how computer restoration can often be more akin to car repair than electronics. Secondly, how fortuitous that it's possible to buy new bearings off-the-shelf for less than a tenner to repair a 30 year old disk drive - one of the 14" removable disk pack types made by Control Data Corporation and used widely throughout the mainframe and minicomputer industries of the 1970s and 80s. Visitors to The National Museum of Computing often comment how much like top loading washing machines these drives look, and indeed the technology is possibly closer to washing machine than the modern hard drive.

It isn't just the industrial nature of the technology that aids the restorer tackling computer hardware of this era - manufacturers used to be a lot more open in the information they provided about their products. The technical manual for this particular CDC drive runs to three volumes, including full schematics, wiring lists, and a reference section written in an agreeable university tutorial style: very useful for repairing the drive, but equally valuable for learning about analogue electronic design techniques.

Fast forward to 2040, and the restorer attempting to get a 30 year old Blu-ray drive working will face some significant obstacles. It will be next to impossible to find any schematics, none of the specialized integrated circuits will be available and to cap it all trying to extract the firmware for analysis will probably be illegal on the grounds of reverse engineering.

We are facing today the culmination of a process that started sometime in the 1980s - a steady shift from 'open source' hardware to 'no user serviceable parts inside'. Should we care? After all, what's special about the PC on your desk that might make it worthy of restoration in the distant future? In many cases it won't matter, but it's notoriously difficult to recognize in the present just what will be interesting or valuable to future historians.

The open source community has been very successful in raising awareness of the benefits of openness in a software context, but exactly the same considerations apply in the hardware world. Increased availability of hardware technical information would ensure more of it survives into the future. It's hard to imagine a scenario where the source code to Linux is ever lost, but you may well have just thrown out the last copy of the technical manual for what will be viewed as a landmark machine in 30 years' time.

If the current trend continues, the computer museum of the future will have far fewer working exhibits. Ironically, the machines from the 60s and 70s will still be going strong (chunky bearings will probably still be available) but that shiny hex-core PC will just be another non-functional box gathering dust in a display cabinet.

Delwyn Holroyd is the restoration team leader for the ICL 2966 at TNMOC. You can follow his progress on the TNMOC restoration project page.

Subscribe to blog feed



-- Advertisement --


About Archives

This page contains links to all the archived content.

Find recent content on the main index.


-- Advertisement --