It should also be noted that even beyond the issue of generating 'goodwill', there's another reason for people to share their work. Suppose I think GIMP is a wonderful program, but it would be more useful to me if it could do X. If nobody else has modified GIMP to do X, making the improvement myself would afford me an immediate benefit: I'd now have a version of GIMP that could do X like I want.
At this point, I can either share the software or not. If I don't share the software, then any time someone else makes any improvements to GIMP that I'd like to use, I'll have to merge those improvements with my own. If, however, I instead release my improvements to GIMP, then I stand to lose very little but can gain at least four big ways if my improvement is a good one: (1) If my improvement is popular, it will likely get incorporated into a mainstream version of GIMP so I won't have to patch it into newer releases anymore; (2) If there are any bugs in my improvement, others may find them and tell me about them (and maybe offer a fix) before I would have found and fixed them myself; (3) Other people amy be able to further improve my improvement; if they do, I'll get the benefit of their having done so; (4) If my improvement to GIMP is a good one, publishing it reduces the likelihood of other people making changes to GIMP that would be incompatible with my improvement.
Thus, even someone who is merely trying to get work done, and for whom issues of goodwill, self-promotion, and pride are all irrelevant, can stand to benefit from involvement with open source software.
Actually, I think your example is somewhat similar to the point I was trying to make with the examples of frontier "small-c" communism (i.e. a barn-raising). The frontier folks contributed their labor TO SOMEONE ELSE'S "PROJECT" in order to be able to reap a benefit to their own "project" at some future time.