Temporary REVU package storage

[Thomas Leonard]

On Sat, 18 Aug 2007 11:06:09 -0400, Barry deFreese wrote:
[...]
>> Additionally, before you start demanding that I do things in the way
>> that's easy for you, remember that I'm a volunteer.
>>
>> On a related note, I really don't think we need to make it easier to
>> submit packages.
> 
> I have to agree with Scott here for the most part.  You have to remember
> that "community" development works both ways.  It's mighty easy to throw
> a package up on REVU/LP/wherever and walk away.  It's another thing to
> get it packaged properly, make sure it meets Debian/Ubuntu standards wrt
> to licensing, packaging, dependencies, etc.

As a contributor, though, it's also hard to get your packaging done 
properly when the review feedback doesn't come until months after the 
upload. Maybe some of the people who gave up would have become good long-
term maintainers if they'd had a good first experience.

I know I'm guilty of this too. As an upstream author, I don't have time 
to look at many of the patches people send me, and I never hear from them 
again. Back when I was a student (and had more time!) I replied to 
everything and built up a good developer community.

> There are over 7,000 bugs filed against packages in Universe and
> Multiverse.  Add to that hundreds of packages that possibly 1 individual
> thinks should be added to the archive.  And this is all to be supported
> by a few dozen volunteer MOTUs?  Many of whom have full-time RL
> jobs/school/wives/husbands/girlfriends/boyfriends/children, etc.

> Tell me how we are supposed to manage that?

You can't. But, these people still need a way to distribute their 
programs. If someone makes a package that's only useful to 10 Ubuntu 
users in the whole world, then the only recommended way to get it to them 
is to get it into Universe. Yet, the full review process doesn't make 
sense for such a small audience.

How about suggesting that these packagers first create a Zero Install 
package (http://0install.net) and maintain that on their own web-site for 
a while? Then MOTU could *invite* authors of desirable packages to get 
them into Universe, rather than having people submit everything and the 
reviewers being too polite to turn them away?

This is similar to the development model used in the distributed version 
control systems: let people publish whatever they want on their own site 
and the 'official' maintainer pulls the bits they want. If you don't have 
to be accepted by MOTU to distribute packages with security, updates, 
dependency handling, etc then there's less pressure to get every trivial 
package in Universe in the first place.


-- 
Dr Thomas Leonard		http://rox.sourceforge.net
GPG: 9242 9807 C985 3C07 44A6  8B9A AE07 8280 59A5 3CC1