[aur-dev] Roadmap

[Simo Leone]

Hey guys,

I would have written this email earlier, but I'm currently overseas
enjoying the effects of global warming (since when does it not snow in
the arctic?). I had a few spare minutes and decided it might be a good
time to sit down and finally write the damn thing.

Right, so I just wanted to get a few things cleared up about where I
want to take the AUR, short and long term.

In the short term, I'd like to keep things realistic. What can and can't
we do with the current level of interest in the AUR? Big sweeping
changes are out of the question, the userbase is quite large and the AUR
is something very prominent in Archland. I haven't read an Arch review
that doesn't mention the AUR in quite a while, so we must've got
something right. We should be somewhat cautious about large
public-facing changes and introducing regressions. Short term fast-track
changes should be things like bugfixes, and admin-side changes, and
small no-brainer type stuff.
Keeping with that, I'd like to tweak the interface a
bit. It's always struck me as a tad clunky, and could use some
refinement. Parts of it are buggy too.
Another small welcome change would be fixing those nasty URLs. Changes
here should be backward-compatible, since AUR links are likely pasted
in all sorts of places.

The AUR's long-term future is less certain. One thing I'd like to get
cleared up is this stuff about AUR2. First off, I'd like to know how
many separate people/groups are currently coding some AUR-successor.
I admit that I probably made a mistake in even introducing the idea
of recoding the AUR, and after some further thought it's probably in
our best interests to improve upon what we have, and be satisfied
with it before any kind of recode is seriously considered. If you're
currently working on a 'new' AUR, now's a good time to speak up, so your
hard work doesn't go to waste.
That said, I've always felt that the AUR's codebase is on a pretty
shaky foundation, and needs some serious rearranging. One of my biggest
long-term goals is to functionalize the code, and separate the HTML from
underlying logic, particularly SQL queries! As it stands, most of it
looks like HTML/SQL/PHP soup, a lot of it quite repetative. This is
really bad programming practice for a whole shizload of reasons,
security among them.
Another long-term issue is cleaning up the data. A lot of the users
are dead, and packages uncared for. The idea of package
maintenence/ownership might be interesting to reexamine and rethink at
some point, the results might really benefit the community.

There is a whole whole lot more that belongs on both lists, but taking
it slow, steady, and simple, is best. The AUR is still growing and
maturing as a project, let's not try to make it grow up overnight.
Lastly, I welcome you to send patches and take advantage of aur-dev.
Developing is always more rewarding when you have others to put their
momentum in the same direction.

-- 
Simo Leone
Arch Linux Developer
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://archlinux.org/pipermail/aur-dev/attachments/20080102/d263978a/attachment.pgp>