On Mon, Mar 12, 2018 at 10:05 PM, Connor S <basicprogrammer10@gmail.com> wrote:
Thanks so much for this excellent background! I will admit I had to read your e-mail 3 times to make sense of it. Connor S. helped explain it to me and I'm learning. :) Connor was excited to hear more about CP/M as it's something he's researched, but not experienced in programming. The computer arithmetic was really interesting as we'd never heard of single or double precision. This information provided so much more context on why the computers were stopping at 33! Now, Connor wants to learn CP/M and retest and also try test on Mainframe. Seems to be a trend of Connor's having interest in Mainframes around here! :)
If you have data you could share from the Kaypro 2 and Amiga 1000, Connor would be interested in adding this to his report. I'll ask him to share the results from the other computers later this week.
Thank you!!! Jen (& Connor S)
your very welcome, sorry for the late reply I was trying to find more options in CP/M to get double-precision math I shared with you the results using screenshots with my camera on Google Photos. And I confirmed your's son's results on my Amiga 1000 with the maximum Factorial possible with double-precision math is 170! If your'e interested in trying more programming on CP/M, with 8 bit computers, there's several more flavors of BASIC available to try this experiment. However I became disappointed after checking their manuals. These 3 were also available beside MBASIC [Microsoft], in order of their release, for the 8bit version of CP/M BASIC-E= developed by Digital Research [maker of CP/M] CBASIC = developed by Digital Research [maker of CP/M] SBASIC = compiler developed by Kaypro computers I ran the experiment on the SBASIC version since I have this on my Kaypro-II and since the manual says this offered double-precision math too. But this had the same flaw as with MBASIC, the highest factorial that it will calculate is only 33!. After checking the owner's manuals for the others, I became more discouraged about the lack of true double-precision support. Because they attempted to offer this in some of these flavors of BASIC but still lacked the number range. All three of those are lacking in this. They had the same flaw as with MBASIC. There were some other computers in that early time period. If your son was interested in these others too, then you can put a call out specifically for any of these computers. Maybe somebody on here can offer to run the experiment too. There were other 8bit and 16bit computers in the early 80s with CP/M-86 (16bit) and the IBM-XT (8bit) and IBM-AT (16bit) . So far I only checked CBASIC-86 software manual for the CP/M-86 16 bit computers. And this shows the same flaw with double-precision math. At least their manual was descriptive to decipher their version lacked this too. But I haven't checked all of those manuals yet to see if there were any hidden surprises. But I'm betting those 3 kinds of computers still did since they offered an option to add the Floating Point co-processor on the motherboard. And this extra hardware contained true double-precision math. Your son's project has opened my eyes a lot more about finding a wide disparity in computer arithmetic from the early home computers(pre-1985). So I was still curious myself but I'm only checking the manuals for now as I don't have all of these other computers. I was going to attempt this at my friend's museum here, Dave McGuire, with an IBM-XT and add the Floating Point co-processor. Dan