ALL,
Some interesting chat from Keith Wootten on USENET:
Keith Wootten author profile email reply view thread post new ú post reply subscribe Email: Keith@wootten.demon.co.uk Date: 1998/09/04 Forums: comp.lang.forth view for bookmarking ú text only
--------------------------------------------------------------------------------
In article <6sp2e4$uqp$1@home.acesales.com>, Michael Coughlin <mikc@home.acesales.com> writes
[discussion of CM and MISC chips snipped] [..The rest of the world will get the idea eventually.]
> The rest of the world will not get the idea if Forth >programmers keep hiding their work under a bushel. A computer >chip is no good unless it gets produced by the millions. >Current designs may be less than optimum, but they get shipped. >If you keep thinking of new ways to make a better chip, and >changing the design, then you will get nowhere. Forth programmers >are always doing the same with their software and their Forth >systems. Forth is constantly undergoing change. Combine that >with computer people's habit of putting off documentation, >and we get nowhere again.
Some time ago Chuck Moore designed a Forth chip, this was the RTX20** based on the Novix and was a very useful part. It was taken up by Harris and had some small commercial success. A while later Chuck Moore designed a better Forth chip. It was taken up by Patriot Scientific and is being manufactured and touted as primarily a Java chip. It's very fast, inexpensive, available, documented etc etc and implements some of the ideas now being extended in the MISC range while using standard production techniques.
My point is simply that it has been done, that there have been many concrete results from a life of almost continuous development. The PSC1000 is a current design, and it's not far off 'optimum'. I've little doubt that Chuck's ideas will continue to produce real chips; it's at least partly our responsibility, and in our interest, to ensure that they are used.
Look at www.ptsc.com for information on the PSC1000 (ShBoom) processor and at www.forth.com for very impressive benchmarks using their SwiftX Forth compiler.
Cheers -- Keith Wootten |