MCA licensing (was: Public mainframe system)

Ian King, among others, mentioned: This was just the first battle of the war. Though IBM had no patents associated with the PC or XT, they did have some (seven in Brazil, for example) for the AT. They spent the first half of the 1990s going after chipset makers (practically all in Japan at that time) to get royalties on those and, more importantly, get them to license other patents (whether relevant or not) that weren't going to expire in 2000 and 2001. Having achieved that goal, they spent the second half of the 1990s going after the PC makers themselves. This was a long time ago, so I don't remember all the details. Most of the patents were just silly in that they described stuff that the Apple II already did, like splitting a screen into a graphics part and a text part. Others were trivial to eliminate, like the algorithm to detect 40 vs 80 track floppies by moving to track 60, going back 59 tracks and checking to see if we got to track 0 or not. Most BIOSes just ask the user instead. The ISA related patents are the ones I remember the least. I think there was one related to having both latched and unlatched versions of some address bits. And the other was about the MASTER signal, which not only didn't really work but was shamelessly copied from an Intel application note for the Multibus. Anyway, as far as I know IBM has made as much money from PC patents as it would have if its MCA strategy had been totally successful. And though the original patents have now all expired, I wouldn't be at all surprised if the money is still coming in. -- Jecel