Here's a heretical thought:

I've never been 100 % convinced that most of the stuff I have submitted even belongs in fink.

Much of it is rather specialized, like programs for X-ray crystallography and other types of biophysics techniques. These programs have a very specialized user community, and for those, they are absolutely essential, and for everyone else, they are probably absolutely useless.

So if at this point if it is making large demands on people's limited time and resources, especially given it seems to take me about 4 iterations to get the package right, maybe it is really a waste of resources for fink to have these packages.

I keep a web site that has more up to date and often less broken versions of almost everything that I have in fink unstable, along with instructions for users on how to install in fink's local tree.

So maybe the first step would be some sort of formalized executive decision making process whereby those in charge and who have a good sense of the nature of the user community's requirements, could make a decision on whether or not a package should be in fink at all. I actually advise people my web site NOT to use packages of mine in fink unstable, because they are out of date and aren't in the 10.3 branch, as the updates are sitting in the queue. Hence at this point I would be better off if my stuff in fink unstable was simply removed, unless it is updated.

Fink could even host a page on its web sites with links to maintainers of more specialized software packages, like mine, instead of formally including it in the fink distribution. It would be then up to the individual niche user communities to see that the packages work correctly. Fink then becomes a centralized core of support for essential software, like tcltk, and experienced core maintainers can focus efforts on really popular or critically essential packages.

Bill Scott




Reply via email to