How about creating a policy that people can refer to? One of the great things about Guix is that is is accessible and hackable, unlike most other packaging systems. Guix encourages contributions and I think it is great that someone comes out of nowhere and submits a package in the Guix way. If there is nothing wrong with the package expression we can take the submission seriously.
I don't think we can accept/reject packages on use or usefulness. I, for one, will submit bioinformatics packages which may be meaningless to 99.9999% of (eventual) Guix users. Are you saying it does not have a right to go in because core maintainers do not see the benefit? I think we should focus on compliance, i.e., the package expression itself and the license, and maybe audit suspect software for not being a spam/worm/virus. Otherwise I think we should be neutral in what other people consider important. It is a bit of evolution too. Unmaintained, defunct or nasty software gets removed eventually. So, what would be a good policy for accepting/rejecting software packages? I favour accepting packages by default, provided the license is acceptable, and remove them again if too many people complain *after* using a package or *after* auditing the software itself for maliciousness. We need to accept by default since we don't have the time to go through all software source code *every* time. I don't favour a policy of core maintainers deciding on what package should go in, or not, in ad hoc fashion. I also don't favour a policy of rigorous mentoring and acceptance, such as used by the Debian project. So, what is great about Guix? And what would be a suitable policy? Pj. On Wed, Jul 29, 2015 at 10:58:18PM -0400, Mark H Weaver wrote: > Mark H Weaver <[email protected]> writes: > > > Orchidaceae Phalaenopsis <[email protected]> writes: > >> * xscreenshot which takes a screenshot and saves it as .if format > >> * imagefile which contains a few programs to convert .if to gif, png > > > > Honestly, I'm doubtful whether our users will have any interest in these > > programs. > > I received a private email suggesting that my response here was > unfriendly. I'm sorry that it came off that way, because that was not > my intent. I'm just trying to express my opinion honestly. > > I've since looked closer at xscreenshot. Its final commit was the day > after its initial commit, which was almost exactly one year ago. In > that time, the only references I can find to it in a web search are: > > (1) a few messages on the suckless email list within a day or so after > it was written > (2) the current discussion on guix-devel, a year later > > and that's it. So, what we have here is a program with a two day commit > history that produces a novel format that apparently nobody uses, and > I've been unable to find any discussion about it in the year since then. > > Do other people think that such programs belong in Guix? > Am I being an ungrateful ass? > > Mark > --
