Motivation gap and pkgsrc for developers ...

[Jens Rehsack <rehsack%googlemail.com@localhost>]

Hi all,

as the one or other may have recognized, I'm a bit unmotivated last
months. This has no special reason - it's more a fundamental gap
between my personal goals and the goals I mean to see from the loudest
developers and the most named packages in the quarterly announcement.

Looks to me like pkgsrc targets NetBSD desktops at first (good choice!),
than some kind of server admins (not only on NetBSD) and than a long
time nothing else.

I'm a contract developer. I like to develop on NetBSD (thanks to
Thomas Klausner who forced me into it with his feedback on my commits),
but there is no commercial effort for me in NetBSD packaging. But
there is commercial effort in maintaining Perl5 modules at CPAN.

I spent around 6 weeks in 1st quarter of 2010 to update important
Perl5 packages like DBIx::Class, Catalyst or Padre. We were at only
40 Perl5 packages in pkgsrc needing an update from CPAN at the begin
of the 1st freeze this year! The only answer I get when making
proposals for the announcement is: there are some less important
plugins of Padre which fail to configure without an X-Server - so
we name at least nothing of the Perl5 packages.

Have a look at the presentation of the last Perl Survey:
http://blogs.perl.org/users/holy_zarquons_singing_fish/2010/06/perl-survey---initial-data-analysis-and-presentation.html

Where is NetBSD named? Where is pkgsrc named? And I'm sure, other
languages from the Top 10 of the TIOBE index will create no other
picture. But we have a very good Perl5 packages infrastructure.
We have excellent developers support (compared to other packaging
systems). But first choices for development are Linux and Windows.

But naming only Firefox, Gnome, KDE, Samba, OpenOffice.org updates
in the announcement will not transport this excellent support to
the mind of the developers. I don't want to minimize the fame for
all the packagers of Gnome (drochner), Firefox (tnn), TeTeX (minskim),
Samba & Apache (tron) and all other very important packages. I praise
Joerg every day using pkgsrc for his (most times) impressive work
at the pkgsrc infrastructure and see the time which is spent by
the admins (spz) and the pr review crew (seen wiz, dh, asau often).

Meanwhile (I thought about this the last 3 month) I think, either
pkgsrc has other goals than I want to reach or we need to improve
something in the project itself:

- we need to identify the goals of the pkgsrc users
- we need to identify the goals of potential new users
- and we need to identify the goals of the committers

I'm not writing these lines to get responses like "Yeah, great
work with the Perl5 packages" - to be true, I write them to
remind the developers that some people (at least me) might have
no job which allows to work on pkgsrc for business but they do
it for some kind of reputation.

Currently - there is no reputation for me continuing my engagement
in updating Perl5 packages for pkgsrc. And currently the pkgsrc
infrastructure is moving in a direction which makes it more and
more cumbersome to use it.

Best,
Jens