Nice PC-museum..

On 17 May 2007 at 5:27, Ensor wrote: You forgot the 432, which was very revolutionary in design and contemporaneous with the 8086. Which shows you how far dramatic innovation will get you. No apologist for Intel, I think that context is important here. At the time that the 8086 (and 8088) was deployed, there was already a large body of x80 software (CP/M, WordStar, SuperCalc, etc.) that were largely written in assembly. The problem with building a product around any processor is what one does for software. With software already available for x80 platforms, and Intel (and others) offering automated translation of x80 code to x86 (and it generally worked well), Intel offered an attractive alternative. Add to that the fact that existing x80 peripherals would work just fine with the new processor and you have a winning combination. Indeed, long after the debut of the IBM PC, there were PC products that were largely composed of automatically translated x80 code. I really was a booster for the 68K--and programmed for it. But no one ever represented that there was a simple and straightforward way to translate x80 assembly to 68K code, nor was it clear if it was going to be simple to use x80 peripherals with the 68K. Zilog had trotted out the Z8000 at about the same time, but it wasn't clear if they were all that serious. I really liked the NS16032 architecture, but National at the time had a reputation to pulling the the rug out from under developers (SC/MP, PACE) without notice and being overly optimistic about release dates (In my case, "real soon now" was about 2 years between the time I got the marketing spiel about the 16032 and when they actually started to be manufactured in quantity). So you saw a dichotomy--existing CP/M applications that were run on the Kaypros and Osbornes made it into the 8086 world. The 68K tended to host the result of whole new development efforts--including Unix ports. Sometimes it seems that around 1983-84, you couldn't throw a stick without hitting a company that was doing another Unix port for the 68K. I don't really blame the guys in Boca Raton for choosing the 8088. There was software for it, but you keep costs down with an 8-bit bus and use commonly-available peripheral chips. Had the the 68K been chosen (and it was a strong rumor, particularly after the IBM 68K- based lab computer was announced before the 5150), it might never have been as successful. Cheers, Chuck