On Mon, 2003-08-11 at 12:12, Colin Watson wrote: > On Mon, Aug 11, 2003 at 12:00:46PM -0400, tallison@tacocat.net wrote: > > [You seem to be breaking attributions - who said this?] > > > > Why can't the installer ask what the CPU is and if it's an x86 then use > > > kudzu and if it isn't don't? Wouldn't that work on all platforms? Or > > > can't an x86 CPU be reliably detected? > > > > It might work. > > It would clearly work; there's a separate installer build for every > architecture! > > > But it blows off the ideology of consistency. > > I disagree; I think it's fine for the i386 installer to behave a little > differently, particularly in inherently architecture-specific things > like hardware detection. I think the debian-installer people think so > too and will be using discover for automatic hardware detection at least > on i386. > > Cheers, > > -- > Colin Watson [cjwatson@flatline.org.uk] I'd also offer that unlike most other platforms, only IA-32 processor-based systems currently operates with a reasonably consistent memory architecture but wide ranging off the shelf plug-in hardware at this point. Compare that to M68K systems where each vendor tended to have relatively limited add-on hardware (generally only from that vendor,) but there were significantly different internal memory designs from one vendor to the next. When faced with those different considerations of the environment encountered for hardware detection, do you doubt that it could be anything but complicated? -- Mark L. Kahnt, FLMI/M, ALHC, HIA, AIAA, ACS, MHP ML Kahnt New Markets Consulting Tel: (613) 531-8684 / (613) 539-0935 Email: kahnt@hosehead.dyndns.org
Attachment:
signature.asc
Description: This is a digitally signed message part