[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: pure64 and biarch64: lessons in biology

Josh Hansen wrote:

As for me, I am extremely excited about John's work and will probably find myself using that for now. Best wishes to everybody of both camps. May peace reign among AMD64 developers!

Very well said Josh.

I have been lurking on this mailing list for about a month now. Frankly I was very disappointed to hear that AMD64 support was not going to be included along with the sarge release.

I feel the multiarch project in general, and multiarch-amd64 specifically, is very important and a Good Thing. I have felt for a while know that it would be nice to have i586/i686/Athlon optimized packages which could coexist with i386. That said, on all of the debain systems I maintain or use, I do not have a need for software which I could not use on a 'pure64' system mainly because the source is available, if there isn't a deb package I would simply compile it. I use debain on my PowerBook and have never felt it lacking because it can't run x86 code.

It seems a real shame to put off being able to run a 64bit version of debian because multiarch isn't complete. The reason I purchased a new AMD64 system was so I could run 64 bit software. I've been loath to install Fedora but until a few days ago I was very close to doing this.

I am very pleased by this so-called 'pure64' AMD64-only port! I think havening a pure64 based installable system would be good for debian which would in turn be beneficial to any multiarch efforts. I've already 'lost' some debian systems I co-maintain to Fedora due to the current age of woody. Not having AMD64 support for over a year would really make it hard to convince people debian is keeping up.

I don't see how these projects are in any way mutually exclusive, on the contrary for the most part they are complementary. I do feel that perhaps the overlap is causing some friction. Perhaps one solution to this might be to disentangle the multiarch and amd64 efforts.

Perhaps it would be helpful to create two or three separate cooperating projects.

debian-multiarch, for all architecture generic work
debian-amd64, implementation of multiarch for amd64
debian-pure64, 64bit only mono-arch

Multiarch is the fine goal and I see pure64 as one step along that path.


Reply to: