Thanks

Zbigniew Łuszpiński zbiggy at o2.pl
Sun May 22 18:34:43 UTC 2005


> --- Couannette <couannette at free.fr> wrote:
> > I agree 200% !
> >
> > That's a multipath communication that could improve
> > Lunar: devs or contributors
> > talking to each others improving modules they take
> > care off. Given that small
> > "thematic" groups could *emerge* around a particular
> > set of modules ;)...
> >
> > Example : I take care of e17 or kino. People that
> > use / know theses software
> > could ask me and we can talk off the list of those
> > topics returning to the main
> > discussion channel with improved modules / ideas.
> >
> > Whereas today all moonbase stuff is done by a *very*
> > small group of devs that
> > can't handle the thousands of modules we have.
>
>   Well i agree with most of the facts here, but not
> with all of them :) I admit that we need an AUTHOR
> field in the module definition to keep track of the
> original author of the module. Too many anonymous
> contribution seems to diminuate the interest from the
> contributors. But, being the maintainer of that module
> is another story. First, devs in any Linux
> distribution are elected by a democratic debate and
> vote of the existing devs (at least that was my case
>
> :D ). So, any person which want to be a dev can demand
>
> that evaluation. 

Yes. This is right way to elect skilled people. But they (like in all 
democratic elections) must present skills. How today lunar contributors can 
present skills if their names, e-mails are erased by one developer who get 
the module and on web page there is info that this developer added this 
module? Everybody will think that this module was created by developer who 
added it to moonbase.

> Of cource, this need some work to be 
> done first (like contributing modules and patches, a
> good relation with the devs, a persistence in using
> and trying to improve the distro).

Yes. But developers put contributors modules as anonymous work. So there is no 
chance to see who from anonymous contributors crowd is good or bad. There 
should be "top list" consisting of: number of modules sent, status 
good/buggy, how many work was done to make it working in lunar. The more hard 
to compile programm, the bigger number of contributed modules, the more good 
than buggy modules sent - the better contributor is. The best contributors 
could be converted to developers.

> Second, a 
> maintainer, which in the Lunar Linux case is a
> designated dev, have some responsabilities versus the
> maintained modules.

If I post something with my name or e-mail address inside I feel responsible 
for it. Otherwise I would post it anonymous. If a devel make it anonymous my 
responsibility becomes developer's - if I find a bug in a module I made I 
probably will blame developer not myself.

> It must assure the needed security 
> and stability, update the module when new versions are
> released (after testing it carefully, and, many times,
> asking for aproval and opinions of other devs) and
> receiving and testing the proposed patches from other
> devs or users of the distro. Now, when we used CVS for
> maintaining the moonbase we couldn't restrict the
> maintainer access to the designated modules only, so
> any dev could modify any module. This can be a big
> problem and can lead to abuses (there was at least a
> case in the past).

The developers CAN NOT give free cvs access if they want lunar keep living.
Opening cvs to everybody is suicide. I only want to preserve rights of 
contributors. There is very important difference between developer and 
contributor (as I understand it):

developer - lunar guru. Knows everything about lunar. Linux hacker (this is 
positive adjective, do not understand it negative as television or other mass 
media try to tell you). Uses lunar almost since birth of lunar :-).

contributor - lunar newbie, but knows linux very good. Started using lunar 
when noticed that more programms in Mandrake or other popular distro he/she 
compiles than installs from rpms. Likes lunar very much and want to make it 
more powerful. Usualy angry when he see that working modules he sent stuck on 
mailinglist instead being included in moonbase or if they striped his e-mail 
address again.

> So, this is the main reason that 
> the devs group is very small for Lunar Linux. Every
> dev must have a great responsability when he commit a
> change in the moonbase. I don't know if we can provide
> now restricted maintenance access to the moonbase
> repository.

Do not do this. Even do not think about it. Experienced cvs user tells you 
this. People can ask you for cvs access, they have in general good will but 
sometimes trying to make good they make worse.

> I know that the main reason to use 
> subversion was exactly this need. We have our
> benevolent dictator Auke which can light our minds in
> that matter ;) I am open to make a team with any
> serious and responsible man (and women :) which have a
> sincere desire to improve the Lunar Linux distro. So,
> don't hesitate to contact me.

Summary of my point of view:
cvs access only for developers.
Contributors work is signed by them to offload developers from dirty work: 
answering newbies, updating modules to new versions, patching bugs.
Developers look what contributor send, do not cut contact informations placed 
by contributors and commit to cvs or not. If not a contributor must be told 
what's wrong so can improve in future.

>   Best regards to all,
>
>    Florin

Thanks for comments florin.

greetings,
zbiggy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://foo-projects.org/pipermail/lunar/attachments/20050522/5b7ef8d9/attachment.bin


More information about the Lunar mailing list