Sponsor shortage

Michael Schwendt mschwendt at gmail.com
Sat Jul 11 10:56:18 UTC 2015


On Fri, 10 Jul 2015 17:24:26 -0600, Stephen John Smoogen wrote:

> This is a vicious cycle. A lot of sponsors are burnt out on trying to
> deal with new people who don't seem to have a clue.

Laziness, lack of activity, lack of interest, sloppy packaging,
dumping-ground/fire'n'forget mentality, there are various factors.

Sometimes it's cluelessness, yes. Paired with no interest in reading
guidelines, since the self-made package builds. And lots of people think
why create an RPM package, if compiling from scratch in trial-and-error
fashion seems to produce something?

Over the last years I've talked to quite some people. Some simply find
the package review process "too embarrassing", because the tickets are
world-readable. Once they learn that the package they offer is full of
mistakes, they consider it "public shaming" and would like to delete
the embarrassing ticket and restart from scratch. In some cases this
has lead to doing early reviewing and guiding in private email, but
with mixed results, such as people starting to argue about guidelines
or how to do something.

Sponsoring someone based on a single package only to find out the person
leaves the project again before handling the first few bug reports is
very disappointing for sponsors.

> And a lot of
> potential new people feel burnt to the crisp by reviews and
> expectations of being wed to a package for life when all they wanted
> to do was help someone else use some software.

That's painted in too dark colours. The review process turns up too many
mispackaged pieces of software, where the packages have not been tested
at all. Offering such packages would be a poor choice. You can hope for
random volunteers to take care of thousands of packages in a dumping
ground whenever they feel like spending time on arbitrary packages, but
that won't work. It is doomed to fail.

Once a package is included in "the distribution", there is much more work
to do compared with the review process.

And "wanting to help"? Lots of packages would benefit from better bug reports
(more responsive reporters) and communication between upstream and downstream.
A dumping ground won't help here. All you achieve by talking about lowering
the hurdles it that the current new contributors prefer waiting for the
Fedora Project to announce something, such as getting rid of the review process,
a dumping repo for unreviewed packages, or automatic blanket-approval of new
packagers.


More information about the devel mailing list