Are *Forge sites worth it?

Is it just me or are GForge, Sourceforge and Savannah and the likes very much lacking on usability, or simply being useful, side?

At least I find myself constantly going to the “real” home page to find something rather than going to the Documentation and Screenshots sections in those sites. One reason, I think, is that even though the site lacks content it feels busy (but consistently) and the “real” home is often much more useful at glance. Another reason is that those dedicated sections are usually empty.

Which would suggest that they’re not so convenient to update, either. I don’t even want to think about the different trackers and horrible mailing list archives. And making releases in Sourceforge is interestingly complicated as I recall.

So I have to wonder, do people, either as users or maintainers, find these sites making their lives easier?

5 comments ↓

#1 will on 04.24.06 at 9:57 am

I have several projects on SourceForge. In general, I find their system clunky and difficult to use. There’s a project called ReleaseForge which takes a lot of the pain out of releasing.

6 years ago (when I started some of the projects) there weren’t any alternatives to *Forges, so the only other option was to roll your own.

If I were to start a new project today, I’d seriously consider using something like Google-groups for the mailing list and Trac/Subversion for bug reports, documentation, the web-site, and version control. I think that combination would be a lot easier to manage and a lot more useful for users and developers. Pylons does something like this (http://www.pylonshq.com/).

The nice thing about SourceForge is that if I decide to give up a project, it’s easier to transfer it to someone because the system is on a third-party’s server.

That’s my brain dump regarding *Forges.

/will

#2 Tuomas Kuosmanen on 04.24.06 at 5:31 pm

I think this is common to a lot of other interesting projects too: The a decent idea of combining “everything a free software project needs” into one platform, but the problem is indeed the usability – the interface is very complex and probably works fine for someone who is already familiar with everything needed (“Where’s that CVS link?”) but it probably is horribly complex to someone who has no existing knowledge.

Plus I agree the “project page” is very “busy”. I also just try to find the “project homepage” – those very often are nicely done.

We have some usability work going on in the community, and some more inside companies working on open source, but I wonder if it is possible to do it in the community the same way as we do the coding? How to do the quality checks for user interface? At least it needs “UI maintainers” just like code has the maintainer to decide. Maybe it could work? It would be interesting to explore.

#3 Tuomas Kuosmanen on 04.24.06 at 5:35 pm

Duh. And of course I clicked “post” too soon. The reason I wish to see if a community process could work, is that a team inside a company usually has to be very product / deadline centered and there would be a lot more things to fix than they can focus on – this could be a nice opportunity for volunteers etc – just like people contribute bugfixes and code. Of course it also means filtering out a lot of strange ideas, plus also having different interesting but contradicting approaches perhaps happening at the same time and someone needs to coordinate the whole thing. But interesting nevertheless. And it doesnt really differ _that_ much from reviewing code and patches either. Maybe it could work?

#4 Murray Cumming on 04.27.06 at 7:49 am

They are better than nothing. And shit as they are, it’s easier than setting up your own separate cvs, bugzilla, ftp, etc.

#5 Tommi Komulainen on 05.01.06 at 9:14 pm

Hmm, didn’t get any notifications the comments… Need to bug jrb again

Murray, yes, I agree the infrastructure is nice to be provided for, it’s just the glue that ties all the things together that’s sucky.

Tuomas, usability checks for UI aren’t too complicated. Little subjective, and time consuming unless you’re seasoned in doing evaluations, but it’s pretty much the same with code. Helps to keep end user goals in mind. I wouldn’t
be surprised if making the user and maintainer interfaces completely separate would allow streamlining both.

I think having a benevolent UI dictator here could be a good thing.