>>>>> "kk" == Kris Kennaway <kris@xxx> writes: kk> Sorry, what was your point? When you work on a marginalized project that has to interface with others, like FreeBSD/sparc64 with FreeBSD/i386, or like BSD ports/pkgsrc with the Linux-centric mainstream of free Unix software, in practice your work is not ``what interests you'' or ``scratching an itch'' or ``what you make of it''---larger and larger chunks of your agenda get chosen by what the dominant project is breaking. It interests some random guy to rototill libcairo for some reason which he finds interesting, and which I would probably find silly if I knew what it was. And that's fine because we all work on our interests, right? But it interests NetBSD's Michael Lorenz to run X11 on Creator3D framebuffers, so does that mean he gets to learn how Creator3D works and write drivers for it? Somewhat, but more so it means he gets to spend his time frantically following around randomguy with a dustpan and cleaning up everything he breaks. If you've done this, you know very well what it's like---the fastest way is often to go through the commit log looking for impatient- or confused-sounding messages from i386-centric people, or just changes in general, and ask, ``could this be it?'' People working on the dominant architecture will say ``I don't have a sparc64'' or ``I don't have a FreeBSD box,'' which means to them more-or-less ``I don't have to worry about what I break here, because after all, how could I?'' Yes, it's always going to be like that somewhat. No, these people who are breaking things definitely aren't bad or necessarily even doing anything wrong. But it's not just an ``we're all volunteers so no one can complain about anything'' situation. It's a software engineering problem, and certain habits, best-practices, social pressures absolutely will influence how much you have this problem, and the ``we're all volunteers'' mantra I think sometimes squashes that process before it can begin.
Attachment:
pgpsWOnyqEwVA.pgp
Description: PGP signature