@mjg59 Since you bring up microcode as canonical example: I hypothesized in patrick.georgi-clan.de/posts/o… that the cut-off point for on-ROM microcode is quite arbitrary and definitely before development is complete, with microcode updates available before market entry.
That has been true for ages, too: Some Via CPU (remember those?) failed to boot into the OS without a microcode update because the OS would start putting parts of the CPU to sleep, and the ROM-version of microcode was having none of that, so the CPU hung.
With that I think it's hilarious that the FSF has a "Defective by Design" campaign against DRM, while the same terms perfectly describe their general approach to computing.
The only way this approach would make sense is if this was putting some objective pressure on CPU vendors to either ensure that their ROM contains good microcode, or to make it Free Software. I don't think the FSF could ever exert that much pressure - and its influence only shrank over the years.
https://retro.social/@patrick/113640830183755694
https://patrick.georgi-clan.de/posts/on-microcode/
No replies.
────
────