Re: Another about SID in ftp.fsn.hu/pub/CDROM-Images/debian-unofficial/sid/jigdo/
On Tue, 14 Dec 2004, Attila Nagy wrote:
> Santiago Vila wrote:
> > > You say people should not use SID at all.
>
> > Excuse me, where in earth I've said such thing?
>
> Maybe it's just me who thinks that those images help a lot of people to have
> an up to date Debian system at home, so if you are against it, in my
> interpretation you say users should not use SID.
There are a lot of people who use sid without those CDs, so your claim
that without them we would have no users using sid and nobody would
report bugs in sid is completely unfounded (not to say ridiculous).
I would even argue that the people who *need* a sid CD to have sid
installed on a computer is *usually* not the kind of people who produce
the most useful bug reports. A good bug reporter would check that the
bug may be reproduced with the latest available version. If you are so
deprived that you don't have internet access, it is much more difficult
to upgrade to the latest version (unless you receive a sid DVD daily).
> > I've just said that sid CDs/DVDs should not be distributed, as the
> > Debian FAQ says. I'm just asking for a little bit of respect for a work
> > which is not yet finished.
>
> I don't see too much difference between Sarge and SID in this topic. For a
> long amount of time Debian refused to release unstable images, because of the
> things you say above. Now there are official Sarge CDs. Sarge is also an
> ongoing project which is called unstable.
For the purposes of being unfinished work, there is a great difference, yes.
I would even say that there is a huge difference.
The sarge distribution is supposed to be in an always releaseable state,
while sid is not. In sarge, dependencies are met, in sid, they are not.
In sarge, there are not brown paper bag bugs of the kind which are fixed
in several days. In sid, there may be, even if they are fixed in a day.
It is particularly the possibility of brown paper bag bugs the thing
that makes the existence of sid CDs especially undesirable.
If I make a big mistake in a package and fix it a day later, I don't
want this bug to be propagated and distributed widely more than it is
already by being in unstable for one day, so I can't be glad that
people distribute unstable CDs containing the buggy package which I
fixed a day later.
To me, the ten days delay which happens for packages in unstable to
enter testing is exactly the kind of difference which makes testing CDs
to be generally acceptable and unstable CDs generally not acceptable.
Reply to: