Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: notesfiles Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!vax135!cornell!uw-beaver!tektronix!hplabs!hp-pcd!orstcs!richardt From: richardt@orstcs.UUCP (richardt) Newsgroups: net.arch Subject: Re: What if IBM...(more 8086 v. 68k) Message-ID: <12200023@orstcs.UUCP> Date: Sun, 1-Dec-85 18:21:00 EST Article-I.D.: orstcs.12200023 Posted: Sun Dec 1 18:21:00 1985 Date-Received: Fri, 13-Dec-85 08:24:44 EST Organization: Oregon State University - Corvallis, OR Lines: 49 Nf-ID: #N:orstcs:12200023:000:2513 Nf-From: orstcs!richardt Dec 1 15:21:00 1985 Yum, Yum. Next? One issue which has een beaten to death in this argument was that IBM wanted to be able to get CP/M software to run on the PC right off the bat. My opinion is that compatibility was a big issue. Evidence: 1) IBM wanted CP/M-86. When Microsoft informed them that DRI was respnsible for CP/M, IBM went and talked to DRI. DRI refused to give them exclusive rights to name and marketing, along the lines which PC-DOS now has. IBM went back to MicroSoft and said "Write us an OS that can use CP/M calls and looks like CP/M-86." Voila, MS-DOS is born. Note that the CP/M "Call 5" convention is *still* supported in MS/PC-DOS. 2) 8080 code is fairly easy to migrate to 8086, *if you have source*. I'm aware that one 8080 instruction can expand into as many as 3 8086 instructions. In the same vein, I can give my C compiler pathological cases that make that look silly. However, an 8080-to-8086 source code translator can still be written *quickly and with minimal effort* == *cheaply*. Thus, 8086 Wordstar is *still* a somewhat larger and slightly stupider product than 8080 WordStar. 3) Given that 8080 programs which migrated would be stuck with 64k, how many PC/clone programs do you see that can handle more than 64k? Moreover, how many do you see that can handle more than 64k *efficiently*? The answer is very few, because the 8086 has a programmer-hostile architecture! You really have to work to use more tha one data segment with any degree of efficiency. 4) The 8080 peripheral market was well developed. No waiting for new peripherals and drivers for a 68k. Most drives which run on a CP/M machine can be used with a PC with little or no modification. Thus, IBM had its reasons for choosing the 8086. They backed down to an 8088 because that meant that they didn't have to think about a 16 bit bus. This is silly, since they decided not to go with S-100, but who ever accused IBM of anything more than a good marketing department. I will aggree with Brad to a point: If IBM had used the 68k, it would not have been fantastic, as there were, *then*, a number of technical and marketing hassles which still had to be dealt with to use a 68k. However, I think that the 4004 and all of its derivatives, up to and including the iAPX family, should be buried in defense of mankind. Anybody else got a shovel? Stepping Off the Soapbox: {hp-pcd | tektronix} !orstcs!richardt Richard Threadgill 1230 NW 23rd #7 Corvallis Or