Re: lenny+1 and the future of the alpha port?
On Thu, Aug 28, 2008 at 02:49:49AM -0700, Steve Langasek wrote:
> So your point in posting was to make insulting rhetorical comparisons and
> browbeat developers into carrying on supporting an architecture that has
> ceased to be useful to them (and almost everyone else in the world)?
No my point was that the desire to keep a machine running isn't always
about efficiency or cost. I have never been under the impression that
anyone involved with Debian is being made to do anything. If Debian was
only interested in what is useful to the majority of people, then it
wouldn't support most of the architectures that it does or even most of
the packages, so that would appear to not be a relevant criteria.
> If people love being able to keep their alphas running, let them step up to
> do the work. Right now all I see is a dying port, and I'm trying to make
> sure that the last one out turns off the lights.
That's fiar. I should do more work. Winter is less busy than the
summer generally, so time to do something.
> There's the list of failed packages, as well as
> (which is a lot more informative than I expected it to be, really). And
> then there's the need for someone to take care of the kernel builds on an
> ongoing basis, and the particular issues that you quoted from my mail.
> Does "doable" mean that this is something that you're going to try to do?
I should do something, so I will check what video cards are in the
various alphas I have (I haven't had these that long) and if one of them
has the relevant card I might as well try to get that working. So far
my first fight was trying to even install anything on these alphas,
given two require milo, and one has the lovely IDE DMA bug (I think I
will switch to scsi on that one) which seems to affect many pws433au's.
> It's great that Debian can fill so many different needs for different
> people, but frankly, if your goal is to show that you can keep old hardware
> working, you don't need a supported, recent OS to do that, and I don't think
> the Debian project should carry a port along for no other reason than this.
No it doesn't need to. I still think that argument would tend to work
against the majority of new ports too where there are very few users
too, although they do have the difference than they are trying to get
new hardware going rather than keep old hardware going.