Jupiter Ace

~ the eulogy ~

The two people who left Sinclair Research in mid 1982 to start Jupiter Cantab were not peripheral. Richard Altwasser had worked on the ZX81 and done the hardware design of the ZX Spectrum. Steven Vickers had written most of the Spectrum's ROM, having already grown the ZX80's four kilobytes into the ZX81's eight. They knew exactly what a cheap British home computer was, because they had just built the one that won.

The Ace went on sale on 22 september 1982 at £89.95. A white case around a Z80 and one kilobyte of RAM, with black and white output to a television. Sound came out of a small speaker inside the box. From the outside it read as a Spectrum with the color taken out, and the reviews said so. The difference was in the ROM, which held a full Forth system instead of a BASIC interpreter.

That was a defensible technical choice. Forth compiled to threaded code as you typed it, so programs ran several times faster than the same thing in interpreted BASIC while using under half the memory, which on a machine with one kilobyte was not a small matter. Dropping into machine code was easy. Anyone who has worked in a constrained environment can see why two hardware people who had just spent two years squeezing a computer into a price would think this was obviously correct.

It is also why the machine sold about five thousand units. The first language you meet is not chosen on merit, it is chosen by what is already installed on the machine in front of you and what the person next to you can help with. My generation met BASIC that way, not because anyone compared it to anything, but because it was in the ROM and in every magazine listing, and the kid who already knew a little of it was sitting right there. The Ace asked a fourteen year old to give that up before writing a single line, in exchange for advantages they had no way to evaluate yet.

~ the dossier ~

~ the forth problem ~

Jupiter Cantab pushed hard at schools, which on paper was the right target: Forth is structured and it teaches you things BASIC actively hides. The push failed for reasons that had nothing to do with the language. Teachers did not know Forth and the exam syllabuses did not mention it, so a school buying the Ace would have been buying a retraining problem. The reverse Polish notation did not help either, since a pupil's existing BASIC was worth nothing on this machine.

The consumer side had its own version of the same trap. Software houses wrote for the Spectrum because the Spectrum was where the buyers were, and buyers went to the Spectrum because that was where the software was. With almost no commercial titles and a display that could not do color at all, the Ace had nothing to argue with except the elegance of its programming environment, which is not a thing that sells a machine in a shop at Christmas.

~ what it left ~

Jupiter Cantab stopped trading by the end of october 1983, thirteen months after the launch. A color graphics board had been designed and never produced. Boldfield Computing bought the brand in 1984 and sold the remaining stock by mail order, along with a handful of games written for a computer that no longer had a manufacturer, and the name has changed hands again since.

Surviving machines are now rare enough to be expensive, which is the usual reward for failing quickly. The more useful thing the Ace left behind is the demonstration itself. It is one of very few times anyone shipped a mass-market computer built around the argument that beginners deserve a better language than BASIC, the argument was correct on the technical merits, and it lost anyway in about a year. That result has held every time since.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »