[OpenIndiana-discuss] IPS pkg popularity contests

[Richard L. Hamilton]

On Oct 8, 2010, at 6:51 PM, Rthoreau wrote:

> Dear: fellow OpenIndiana users developers;
> 
> After installing a lot of programs not listed in the repositories, such
> as msmtp, aspell, hunspell, and trying to get a basic desktop
> environment.  I was wondering how Illumos, and OpenIndiana are going to
> handle the every increasing packages that people will want to include or
> have included into the release cycle.  Which brings up the desire for
> some kind of popularity contest, which would give priority to the most
> used packages and could cut down on the ability and security
> vulnerabilities of those packages.
> 
> I could see that OpenIndiana could end up in a situation such as Debian
> where every release has more and more packages without maintainers but
> packaged to scratch a need then left in the repositories until someone
> comes along and claims the package, or it gets dropped due to
> inactivity.
> 
> It would also be nice for new users to have some good documentation on
> popular packages not in the repositories.  Hanging out in the IRC
> channel and following some of the lists it might save a lot of people
> some headaches if we had a wiki page on common packages that people
> might want.  This might include such things as the KDE repositories, and
> those for mplayer and the codecs for those not using the Fluendo
> codecs. Like how to compile lame, and other basic desktop packages that
> might not be included due to restrictions and license issues.
> 
> I ran into this problem as I tried an article on the sun observatory
> about how to compile mplayer which failed to compile.  Then in the irc
> channel I found that someone had already had a repository with the
> included package.  The same goes for kde, just following the lists it
> seems that a few posts describe the repository and a few directions on
> getting it to work.
> 
> As we all know that documentation can make or break a project and can be
> a point of keeping people interested. That is one reason why the BSD's
> has done so well over the years the documentation has been very good and
> usually works between releases. My concern right now is to see if we can
> get a general consensus on the proper method to make an outstanding
> resource for users and developers.
> 
> Rthoreau

I think it's not enough to have a list of apps and votes per app.
When there are multiple apps that do _roughly_ the same thing, it
might be nice to have them all, but I think it would be good to
go for those that provide the most essential (for some definition
of that word...) functionality first, and so get a broad range of
functionality before going for alternatives that might have a little
better functionality or usability for some people's needs, but not
have as much to contribute to the whole.

I'm assuming that no Solaris derivative will ever have the
multiple tens of thousands of packages that some Linux distros
may offer...at least not unless a whole bunch of people get really
busy to make it happen.  So, however priorities should be set, I think
it should take into account three things:
* demand
* a good spread of functionality
* who is willing to do the work

In the end, it may well be the last that decides...