Re: amd64 into mainstream
On (22/04/05 05:25), Jaime Ochoa Malagón wrote:
> Beautyfull, this perception of debian is, from my point of view, very
> usual, I have been using debian for many years (>7), the oldest I
> remember is potato, but I'm sure there is something else before, the
> schema of release of debian is pretty well done, in fact when I had to
> take my choice of distribution was between redhat, debian and
> slackware primary, slackware was very buggy because all the packages
> was like sid or worst, redhat used to install to many packages and use
> all the oldest HD, but in that time debian has dselect and that make
> the difference resolving libraries dependencies, knowing what packages
> are installed, what packages needs to be upgraded, etc.
>
> No body had seems to take debian seriously and nobody know it, that
> has changed release by release the history is the same, when the
> testing becomes new stable release everybody can use it but after a
> year or two the old software, the new hadware and all the rumors
> claims to move to testing again and again, this is the principal
> problem...
>
> The others ditributions have shorts release cycles, and the users not
> complain about the new bugs, because they have been expecting this...
>
> With debian is quite different because we are sure the product is of
> better quality, but itsn't clear to the average users.
>
> IMHO as you say it security of testing sounds like an option the
> divorce is another good idea but the point is to give recent software
> to the mass and preserve the quality of the mainstream, this can be
> achieved with the following idea in mind "The average users is
> expecting the software to fail." think about them like beta users.
>
> I think the following schema could be work...
>
> unstable ->
> testing ->
> desktop ->
> stable
>
> where desktop means "every n months wait to see a few complains in
> testing and make a desktop release" this release is waiting a few days
> to catch up the few complains bugs and be the security/testing.
This proposal has great merit, however, I suspect that adding an
additional flavour of Debian could be the 'straw that breaks the camels
back'.
There is already talk about dropping ports for less popular
architectures; the long release cycle is a product of the scale and
complexity of the Debian project.
Adding a further flavour wouldn't simplify or speed up the release
process.
Any voluntary effort has its difficulties and generally it is the few
who shoulder the responsibility for getting things done whilst the
'many', of which I'm one, benefit from all their work. Perhaps
encouraging the 'many' to be part of the solution and recruiting more
people to help with testing, documentation etc. is the way forward.
I'm no geek and so can't offer any technical skills but I can test and
write; there seems to be an absence of a framework whereby developers
and maintainers can coordinate users' voluntary efforts. The
current informal arrangements seem to work fairly well but I can't help
thinking that better coordination of individuals' efforts would yield
greater efficiency.
One only has to think about how many great howtos are out there on the
web, to realise that if they could be brought together in cogent form,
how much more professional and appealing debian would look. If you make
the effort to 'learn' Debian, its qualities are self-evident but to the
outsider it may seem obscure and possibly even elitist.
In summary, improving release times could be achieved by harnessing the
goodwill and diverse skills of the user base.
Regards
Clive
>
> All of this with a legend like:
>
> This version is not for production servers, the sucesive bugs in this
> release have the hope to be corrected in the next release, our efforts
> is to create a really trust release called stable, this desktop have
> the tipical behaivor of other distribution of linux, our aggregated
> value is to have a well probed high trust stable release.
>
> On 4/21/05, Matthias Julius <jnews@julius-net.net> wrote:
> > lsorense@csclub.uwaterloo.ca (Lennart Sorensen) writes:
> >
> > > woody is a perfectly good samba/nfs/apache/dns server. Nothing wrong
> > > with the software in it. A few years ago that software was state of the
> > > art, but now you think it is unusable? Why?
> >
> > Woody certainly is as good as it was 3 years ago. Only with time and
> > availability of new features expectation changes. Apache2 and PHP5
> > for example have some advantages. And so have some other packages.
> >
> > Who would want to buy a new model '80 car? It is certainly as good as
> > it was in 1980. But technical debelopments make it desireable to buy
> > a newer model.
--
www.clivemenzies.co.uk ...
...strategies for business
Reply to: