Re: Concerns about AMD64 port
On Thu, Feb 05, 2004 at 11:23:03PM +0100, Goswin von Brederlow wrote:
> John Goerzen <email@example.com> writes:
> I think the attitude is that its sid, to hell with compatibility. Get
> it running and then think about an upgrade path or just rm -rf / and
> reinstall the then emerging final port. Your making a new port, there
> is nothing to be compatible with.
> If its just a temporary solution till the multiarch is sorted out and
> takes over that is fine.
I do not see it in such stark black and white as you. I see the pure
AMD64 port as the starting point; the base upon which the multiarch
features can be added later. I do not think that the upgrade path will
be difficult or traumatic for either our users or the developers.
However, it must be said that nobody really knows exactly what form it
will take since nobody can yet agree on what form dpkg/apt/etc will
> > I started this thread responsing to the message on the AMD64 ports page
> > that says that a pure 64-bit port would be of little use. I continue to
> > maintain that this is wrong, and I have ample experience with pure
> > 64-bit systems to prove it. I am sure you could find many more on
> > debian-alpha and axp-list that would agree.
> That has zero weight. Noone ever claimed 64bit support for amd64 would
> not be needed and thats all you can get from other 64bit archs.
I said "pure 64-bit port". And yes, somebody did claim that: see
http://www.debian.org/ports/amd64/: "A pure 64bit port seems to be
academical and of little use."
> You will be stealing away people that would otherwise invest their
> time into getting multiarch support along. If it catches on it will
> considerably slow down the multiarch amd64 port. But at the moment
I'm not sure that either of these are correct, and even if they are, I
don't think it's a bad thing.
People will invest their time in the things that are important to them.
I care not about multiarch support; therefore, my time is being invested
in getting a 64-bit userland going. Those that care about multiarch
will no doubt invest their time there. If multiarch withers and dies
because nobody spends time on it, that makes a strong statement about
how important it really is. Similarly, if nobody else cares about a
pure 64-bit userland, that project will wither and die. Really, though,
I don't think these are exclusive propositions either technologically or
with respect to resources.
However, once a 64-bit userland is running, I don't see how that will
sap people's time to any great extent. Right now, it's sapping *my*
time, but I wasn't going to work on multiarch anyway. Heck, I wasn't
going to *any* AMD64 work until I get my machine next week.