Back when I was a teenager, I would have absolutely gone down a rabbit hole like the author did. From "Upgrading and Repairing PCs" to reading all the technical manuals, usenet, etc. I definitely nerded out over this stuff! Glad to see folks still take an interest.
These days, I've an acquired brain injury. Between that an old age, it was a bit hard to read, but also, just a little bit familiar, so I enjoyed it.
Now I am expecting "256 color VGA programming in C" to resurface at some point! :D
320 x 200 CGA allows you to have only 4 colors at a time. Of course you can change the fixed palette on a certain scanline to have more colors. See for example California Games.
Yes, I did consider writing "4 colors", but thought someone would say "but you could have 16 colors on-screen" citing the early 80s titles which did scanline palette tricks. So, I went with "16 colors" as it was technically the upper-bound of 'official' documented colors (excluding composite artifact, dithering and other cool tricks). I didn't really want to get into explaining more since my goal was just linking the "8088 MPH" demo for the GP and anyone who hadn't seen it. Also, I have no idea if a technically correct, definitive and complete statement about CGA's maximum possible colors in a single English-language sentence is even possible.
But after wasting... oh, nearly ~800ms on this internal debate, I realized I'd already cleverly chosen to not link directly to the 8088 MPH video but instead a site containing both the video and links to explications revealing the myriad brilliant tricks and unnatural acts behind 8088 MPH. And from that rabbit hole, one may learn more than any mortal should know about CGA graphics. Especially apropos since the author of the OP was instrumental in creating 8088 MPH (although the OP's post is about BIOS things).
That was one option I thought of at first (mentioned in the first section), but the info I found indicated that the /370 models used the same firmware as the "plain" 5170s - if there were any BIOS extensions, they were probably somewhere on the add-on cards. The AT/370 also had 512K of on board RAM, while this BIOS seems to indicate 640K.
Plenty of people remember, and used, them. Just not people who tend to hang out here. I knew several IBM VM dev types who had them as light dev/remote mainframe access machines, usually at home. They were popular enough there was a followon product: the PC/390 which was the same idea, more advanced processor, based on a PS/2 microchannel platform (and, AFAIK, OS/2).
You want really obscure? Unisys had the same idea with the "Micro-A", which was a PC running OS/2 with a coprocessor card with a single chip implementation of an A-series mainframe. I know of 2, possibly 3, still around.
Details: The IBM AT/370 used standard bios on the motherboard, and the two 68k custom cards had their own bioses. The 68ks were very heavily modified by one of the motorola engineers.
Its the second version of the AT Bios that was disgusting was verion 2, that ran on 6mhz 286s and prevented you from swapping the crystal for a 16Mhz/8Mhz speed up. The first version had bugs, and the third version was for the 8Mhz machines. ( still a few bugs ).
Oops. Anyway, I remember attending a talk by one of the IBM engineers back when they first released the XT/370. He said that they looked at all possible ways to integrate their production line as a kind of secondary track off of one of the main production lines for the PC/XT, but the most economical option ended up being a separate facility that would receive normal pallets of regularly boxed, end-user XTs from the main factory, unbox them, make the mods, and pack them back into XT/370-labeled boxes for shipping.
These days, I've an acquired brain injury. Between that an old age, it was a bit hard to read, but also, just a little bit familiar, so I enjoyed it.
Now I am expecting "256 color VGA programming in C" to resurface at some point! :D
Old hardware was always so much fun...