Michael Cummings wrote: [Sat Mar 24 2007, 07:54:51AM CDT] > On Sat, Mar 24, 2007 at 01:50:19AM -0400, Mike Frysinger wrote: > > On Friday 23 March 2007, Josh Saddler wrote: > > > I'm very strongly against using Gentoo SoC time and resources for things > > > that are not officially part of Gentoo (yes, this statement could be > > > spun however you wish) or are not official Gentoo projects. And no, just > > > because a project has Gentoo developers in it doesn't mean that it's a > > > Gentoo project -- let's avoid the gray areas now, shall we? Just because > > > we have Gentoo devs who are also Gnome upstream doesn't make their > > > Gnome-related packages that happen to be in our tree official Gentoo > > > projects. > > > > i'd have to agree here > > -mike > Ditto. Gentoo SoC projects need to be for Gentoo developed and > sponsored code/projects, not third party projects, no matter how much > they would whither and die without a gentoo core. There was an example > of gentoo+gnome integration (i think) in a previous email - that > wouldn't be any more appropriate. Unless there's the Gentoo Inc > copyright in the header, it isn't eligible in my opinion.
Okay, let me explain why I think all three of you have the wrong idea here, although I have sympathy for your argument. First, there's the issue that hosting projects that are only tangentially related to Gentoo drains our resources. To some extent that's true, but it's a minimal effect. What resources are we talking about? Infra provides cvs or svn for the SOC students, we have a gentoo-soc mailing list, and I suspect there will be a gentoo-soc planet again. Getting that set up requires significant effort, but the difference in effort between 5 students and 10 students is not very much. (One could imagine web-based projects that would also require Infra to provide various web-based or network-based apps, but those are likely to be for true Gentoo projects, so I'm discounting those for this discussion.) Gentoo also provides mentors, who choose to volunteer their time. If nobody wants to mentor a project, it's not going to be accepted. You could argue that by allowing these sorts of projects we are encouraging devs to spend time on non-Gentoo stuff. *Shrug* Our devs are volunteers, so I figure they're going to spend their time doing what they want to do anyway. (Incidentally, if that gnome+gentoo student chose to submit his or her proposal to Gnome, nothing would stop one of our devs from officially mentoring that person as long as the Gnome folks agreed (or unofficially mentoring if they didn't). Of course, for many my above argument is beside the point. It isn't the resources, it's the principle of the thing. SOC projects hosted by Gentoo should be Gentoo projects that clearly benefit Gentoo and have "(c) Gentoo Foundation, Inc" stamped on them. I have sympathy for that argument, but I respectfully disagree, because I think that argument misses the essential point of Google's Summer of Code program. The primary goal is to get students involved in developing open source code, and thus bringing new blood into the community. Even if our students don't become Gentoo developers, if they have a good experience they are likely to be friendly to open-source software, at least, and perhaps even long-term active contributors. My view is that we are providing an altruistic service here to benefit the community in which we reside, not to get free labor and a bit of cash (Google pays the hosting organizations as well as the students). (That said, we nonetheless did pick up some nice code and at least two devs from last year's program. Also, our being chosen to participate nicely enhances our reputation, both as being a significant player in open-source and as being one of the "good guys" in the community.) It's possible that I'm not being terribly convincing. After all, a student who submits a proposal to nmap is almost certainly going to be working on nmap, not, say, honeynets. Why should Gentoo be different? Well, for one thing, our main product is a distribution, and we spend most of our time integrating existing code instead of writing new code, so we're pretty much a natural umbrella organization anyway. Last year one of the proposals that was submitted to a number of distributions (including ours) involved porting Sun's ZFS to Linux. It was a very well-written proposal, and it was accepted by several different organizations. (I don't remember which organization he went with, but it wasn't Gentoo.) Clearly having ZFS support in Linux would benefit Gentoo in the long run, even if it wasn't an obvious Gentoo project, and I'd have been perfectly happy supporting it. (In case you're curious, you can follow along the progress of that proposal at http://zfs-on-fuse.blogspot.com/, and he just released the first beta at the beginning of this month.) Finally, let me be more specific about pkgcore- or paludis- or gnome+gentoo-related proposals. If not us as a mentoring org, who? They're clearly all Gentoo-related, even if not "pure" Gentoo projects (whatever that means), and it's not as though somebody else is going to pick them up.... Feel free to tell me why I'm completely off my rocker. It probably wouldn't be the first time. -g2boojum- PS. So, anybody have any actual technical comments about this proposal? My personal opinion is that I wouldn't characterize it as a high-priority project for either Gentoo or Paludis, but the quality of the proposal itself is decent and real thought has been put into it. (I can't figure out why this project would need both boost.python and the raw python C api, but that's just a detail.) Moreover, it also seems reasonable to me that a hard-working student could make real progress in this over the Summer. -- Grant Goodyear Gentoo Developer [EMAIL PROTECTED] http://www.gentoo.org/~g2boojum GPG Fingerprint: D706 9802 1663 DEF5 81B0 9573 A6DC 7152 E0F6 5B76
pgpbXjn3wJk0r.pgp
Description: PGP signature
