Path: csiph.com!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail From: rbowman Newsgroups: comp.os.linux.misc Subject: Re: AI-Based Coding Taking Over Date: 26 Oct 2025 18:40:47 GMT Lines: 71 Message-ID: References: <10crldd$lkb4$1@dont-email.me> <10ctb13$12oei$2@dont-email.me> <10ctrdc$16st9$1@dont-email.me> <10d24n9$2bb83$2@dont-email.me> <10d2dos$2dnk0$3@dont-email.me> <0g8hslxmvj.ln2@Telcontar.valinor> <10detk2$2b3jr$1@dont-email.me> <10dhm2v$33b7e$1@dont-email.me> <10dkuvq$3uqs3$4@dont-email.me> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Trace: individual.net HmmVBRZCoB+61deWJn9AiATBlpAtWNSfH5ErQJ8ggwlYQRVbH4 Cancel-Lock: sha1:y/1howiU4Rrz7QUj/rNY6+s+ftI= sha256:DR1Ps8X1er1kzAKUsMoO3jaRUC1JpUt3BEm4Ar5GZck= User-Agent: Pan/0.162 (Pokrosvk) Xref: csiph.com comp.os.linux.misc:76720 On Sun, 26 Oct 2025 11:00:10 +0000, The Natural Philosopher wrote: > Having programmed on both, they were. > CP/M was very short of RAM and the typical machine had no standard > expansion slot. https://en.wikipedia.org/wiki/S-100_bus Many didn't allow for expansion but it wasn't a limiting factor in most cases. For S-100 there were prototype boards that plugged into the back plane so you could roll your own. http://www.s100computers.com/Cards%20For%20Sale.htm > 8088 was a good step up. 386 was the next big one The 8088 could only address 64K. You could point CS, SS, DS, and ES to different 64K blocks. CP/m 3.0 could use external logic bank switching to accomplish much the same. Intel formalized it in a processor that was supposed to be a temporary fix before the iAPX 432 arrived. The original PC ran at 4.77 MHz. Odd speed, right? The CGA adaptor needed 3.579545 for the color burst. (I don't know about the European market). 4 times 3.57945 is 14.31818. Divide that by 3 and you get 4.77277. Couple of dividers and you save the cost of a 8284 or an extra crystal. https://en.wikipedia.org/wiki/Intel_8284 The 8088 also used the cheap 8bit peripherals from the 8080, 8085, and Z80 designs. Of course that meant limiting yourself to a 8bit data bus but money is money. https://vintagecomputer.com/osborne-occ-2-executive.html I'll go with the article rather than my memory but I remember the Executive as using the Z80B, which was a 6 MHz part. I had an Osborne 1 but bought two Executives from the Boston Globe when they were moving to PCs. For all practical purposes the original PC was not faster than a CP/M system and had a hell of a lot less available software. Much of that had been ported from CP/M and was less than efficient. Part of that was the memory models. https://digitalmars.com/ctg/ctgMemoryModel.html Building an executable was fun since there were 5 sets of libraries for the 5 models. Don't forget the 5150 (IBM Personal Computer) came with 16 KB of RAM. Usually you needed more slots than it had since the mobo was limited. The 5160 (XT) at least kicked the base memory up to 128 KB. It had more slots but good luck getting some of the cards into them. You might want to toss the cover and point a fan at the thing. The 5170 (AT) was starting to get real with wider buses and a 6 MHz processor. Let's forget the PS/2. You couldn't tell the horses apart without a program on that mess. I think they built them from spare parts and, arguably, were just about out of the PC business. So, yeah, given a CP/M 3.0 system versus a bog stock 5150 I'll take the CP/M box.