Google Summer of Code

[Adam Gołębiowski]

On Sat, Mar 03, 2007 at 01:41:19AM +0100, Daniel Mróz wrote:
> On Saturday 03 of March 2007 00:39:22 Adam Gołębiowski wrote:
>                              ^^^^^^^^ Whoa!
> > We have just a few days left, we really should think of any projects that
> > we would like to propose for this year's edition. I looked through SoC 2006
> > thread that took place here and nothing much changed since then:
> >
> > * GUI for poldek
> IIRC there was such project already started by one of our developer/user. I 
> don't remember what happened with it.

I am aware of pacman (pacman.spec) and kpoldek (which was recently
started). Both suffer from the same failure - they wrap poldek instead
of using libpoclidek (which would be better, faster, and, well, better).

> > * an installer that does not look like one from Slackware 3.0 age
> Maybe we should start with an installer that works ;)
> There's nothing wrong with text-mode, light installer. I think that instalka 
> was the best installer ever. Simple, fast and working everywhere, anywhere.

areq's rescuecd is the best one, but I realize it may be the best one for me,
not necessary for everyone. Having a modern, gui installer wouldn't
hurt. 

> > * pld stats
> >
> >   client / server collecting information about users' preferences, or
> >   better, packages popularity. I see it as an additional package which
> >   user has to enable himself (default configuration == do nothing),
> >   which would generate some sort of system's unique id and would
> >   frequently sent a bunch of information to the server:
> >    - architecture
> >    - detailed hardware info (no of cpu, memory, hdd)
> >    - list of packages installed
> >    - IP (for GeoIP localisation (country only))
> >    - ...
> >   User could choose which information to send.
> Naaah. Nobody will ever use it. Waste of time IMHO.

why? I would, and I believe we could find a group of users who would
install and use this too. 


> > * pld/freebsd (if trojan agrees)
> >
> >   work on PLD's port to freebsd (to be specified), a working port could
> >   attract some potential users/developers.
> What for? We have enough problems with different hardware platforms (Sparc for 
> example).

geek points, PR, name it yourself. There aren't many initiativies like
that. If google agrees on that, we would not only benefit from being SoC
participant, but also 

> > * pld infrastructure redesign
> >
> >   re-write or extend builder's infrastructure so that requires almost no
> >   interaction from RM. I know there are some TODOs (like hold an upgrade
> >   until a package is built on all architectures). That could be part of
> >   the project - other ideas include adapting bug tracking system to our
> >   needs (like...  automatically assigning a bug to last commiter of the
> >   given package), or some rss feeds of new packages.
> This is a really great idea. Useful and complicated enough. Maybe we should 
> add one more feature to this list: design to easy deployment in other 
> distribution projects so it can evolve into new, distributable system for 
> Linux developers.

Hmm... maybe not a bad idea. Providing a complete set of tools (like
trac) which lets you to 'build your own distro in 5 easy steps'.

> >   buildlogs redesign - there was some ideas on one of the devel lists,
> >   yet there was no response.
> Like...?

RDBMS based one - grep devel's archives. If I remember correctly, it was
witekfl's idea. The problem with what we have now is speed - takes a lot
of time to search through it, interface could have some 

That, as well, could be a start for a packages.pld-linux.org - which was
proposed a year or two ago. I would see it as a mix of
packages.{gentoo,debian}.org with information about package, and some 
links - homepage, ra/ac/th/xx rpms, spec file, buildlogs and other.

> > * network manager (don't know how appropriate that one is)
> >
> >   adapt NetworkManager to PLD's /etc layout (if it needs it) or design a
> >   new one from scracth (gtk/qt based). Project aims on creating an
> >   easy-to-use tool for network configuration.
> There are similar projects/products on the Intenet.

But I guess none that fit PLD in 100%.

Do not think these about projects as wasting our (limited) resources on
some useless project. This is not important (well, not the most
important).  What really matters is the opportunity on doing some
positive PR to PLD.  And if we can somehow benefit from it - that's even
better.

-- 
 http://www.mysza.eu.org/ | Everybody needs someone sure, someone true,
   PLD Linux developer    | Everybody needs some solid rock, I know I do.