Hi all! I wanted to pose some choices for the future of community.haskell.org
It provides the following services (listed on the homepage) * public read-only version-control repository space * shared write-access to VCS repository for project members * personal webspace for members * public webspace for documenting your project * a public Trac instance for project management * a public MailMan mailing list for your project When we set up this box it was the case that A) the Haskell community was smaller and B) there were not many good, commonly used resources of this sort for the open source community. These days, this is not the case. So what we have is a box with lots and lots of users with shell accounts, providing services that are largely duplicated elsewhere. However, there are some high-profile projects that remain users of c.h.o and we don't want to leave them in the cold. Furthermore, it is an option, no matter what we do, to keep portions of c.h.o in "cold storage" as read only so that legacy links don't break. Of the services provided, the VCS stuff (for darcs in particular) is, as I understand it, the most used. Notably this is used not only for public-facing repos, but also private collaborations by academics on forthcoming papers, etc. However, I think hub.darcs.net is now in sufficient shape to take on all needs in this regard. We could then leave in place project and individual webhosting (which, gh-pages aside, is actually generally still a pain to 'get right' these days in a free and convenient way) but restrict access to sftp, rather than shell, which should clean things up significantly. As I understand it, the trac and mailman stuff is seldom used, and rather bitrotted and unmaintained as well. It seems to me that it should be possible to encourage users of these features to use another public bugtracker (hub.darcs has one, as does, of course, github) or google-groups. (the box also serves planet.haskell, and we can continue to host that of course). This leaves in place c.h.o for hosting haskell-community related content, and meanwhile lets us clean up the glaring silliness of hundreds of shell accounts, and more centrally maintain what we _do_ have. Thoughts? Concerns? Disagreements? (Also note that if we come up with a plan, then to go forward with it will necessarily be some work, and more people are welcome to volunteer, especially with regards to communication and coordination with the many current c.h.o users). --Gershom _______________________________________________ haskell-infrastructure mailing list [email protected] http://community.galois.com/mailman/listinfo/haskell-infrastructure
