[gentoo-dev] Re: Democracy: No silver bullet

[Wiktor Wandachowicz]

Richard Fish wrote:

> > I suppose that there is a way that Gentoo can follow, only that its leaders,
> > developers and users need to see it clearly. Is there a publicly visible
> > page that contains current goals for new releases? Where all sub-project
> > leaders could add their own goals, coherent with the general vision?
> > I couldn't find it, but maybe I haven't looked in the right places?
>
> The problem I see is that for Gentoo the releases are not really
> useful milestones for most projects.  A release is really significant
> for a few core packages, but what is the real downside for users if
> Xorg 7.2 is stabilized one week after a release?  Outside of the fact
> that they have to compile it themselves instead of using the GRP
> package...not much that I see.

After using Gentoo for a good while I appreaciate very much the "constant
development" policy, which prevents the need to upgrade my system to new
releases. I've seen one Ubuntu user dist-upgrading its installation - it went
with some problems, and they were substantially bigger than I'm having doing
occasionally "emerge -avuD world".

But to be honest, stabilization of packages was not my point. ((BTW, stable
X.org, KDE or GNOME would IMO delay the release for a week, so users wouldn't
need to upgrade in such a short time frame - but that's what I think))

I was rather thinking about bigger, user-visible changes. Obviously a big
version bump of widely known and used package would fit this category, too.
Good news could include, for example, new stable kernel + udev (with better
support for [many-nice-features]), GNOME/KDE/XFCE/etc, even easier installer,
Gentoo-branded themes (Grub, splash, gdm theme, wallpaper, icons, colors (?)),
stable porthole, improved portage... These are the things people are looking
for - better, faster, easier. Opportunistic? Yes. Drugery for developers to
come up with such a list and then hold their word in time? Yes. Is it needed
at all? IMO, yes.

> For a distro like Ubuntu, a release is very significant, as it is the
> platform that users will be running for the next 6-18 months.

And for Gentoo it's about 6 next months where new blood, umm.. new users
/the beloved newbies ;)/ come to the project based on the reviews in news
sites. I, for example, got to know about Gentoo after reading a good review
on the site I was visiting quite often (linuxnews.pl). When I took a look
at the Handbook by the first time I was sold immediately. I was thinking
for a long time about installing LFS and only the time was an issue. Then
here came Gentoo and my world changed... for better.

Having said that, releases are targeted mostly for new users. Release media
become more and more filled with features and are more user-friendly than
ever (GNOME running from LiveCD, graphical installer, and so on). Lots of
*visible* changes (even though they are minor or trivial) buy new hearts and
minds for Gentoo. Do you now see what I've meant?

I'm not imposing that Gentoo development should depend on a time-based
milestones but new releases of installation media do happen and are needed.
It would be easier for journalists, newbies, etc. to compare Gentoo against
other distros if some kind of list of features that would-be-nice-to-have
before every release existed.

> Do you think Ubuntu roadmaps would be useful without being tied to a
> release?

Of course not. But that's exactly why people know beforehand that Dapper
would contain one list of features and be stable, while Edgy (advertised
as developers' dream) can be somewhat rough but most probably will contain
another list of new and exciting features. Example [1].

> Or could project status reports (as discussed here recently) fit the
> same bill?

Thanks for pointing this out. Need to re-read the archives.

With best regards,
Wiktor Wandachowicz

[1] "Upstart in Universe"
    http://www.netsplit.com/blog/work/canonical/upstart.html

-- 
gentoo-dev@g.o mailing list