On 2/6/07, Christopher Blizzard <blizzard@xxxxxxxxxx> wrote:
Luis Villa wrote: >> I see them as somewhat separate things. To your first point - making >> things easier to install, I happen to completely agree, and it's one of >> the reasons why I'm driving down the path of using bundles instead of >> rpm or deb packages in olpc. This much to the chagrin of people who are >> heavily invested in rpm and similar technologies. >> >> I'm approaching this from the standpoint of "make it easier to find >> software that's interesting and be able to give it to others" as opposed >> to "making it easier and more convenient to make an rpm." > > I agree that the first approach is a vastly superior one which should > be prioritized given finite resources, but that is a very big problem > space with a lot of barriers preventing a solution. Given that, I > don't think you should so casually dismiss the second approach in the > meantime- I think as an intermediate bandaid it possibly has quite a > bit of value. We're trying to make it pretty easy. We have a bunch of tools that lets you build an rpm locally, a "local build service" if you want to play rhetorical games. How we allow third parties be able to distribute and let other people know that they have compatible software is a different issue. That's the kind of thing that I've been pushing us towards as well. The pieces that Jesse and others are working on are the beginnings of that.
Ah, I may be missing information on what jesse and others are doing- pointers?
Our current system of centralized distribution _and_ information about what's available seems to make that harder. i.e. the only way that people can find software that's compatible with Fedora is to find it in Fedora's repositories. It's something that has scaled well for us with free software,
It makes it hard for people who are producing free software as well, since even if you can guarantee that all of your software is available on all distros, it is impossible to quickly get bugfixes and new versions out to users on all distros. I think this is the problem suse's build system is trying to fix- release one tarball, upload it to their system, *poof*, you've got builds for everything that you can point your users at. The current system is release one tarball, wait days or months for distros to pick it up, and in the meantime apologize to your users for their inability to get their software on distro X.
> Luis (who sometimes wonders if web + xpi means we should have moved > the whole desktop to mozilla/xulrunner two years ago) There's a huge amount to be learned from what Mozilla is doing, and I know I'm personally applying it. But it's hard for others to see it. They don't quite have the inside view that I have had to both projects. What's worked, what hasn't, what are the key success factors, etc. I'm trying to prove out some of using olpc as a mechanism, but it's slow going.
I would love to talk with you about this at some point; I'm currently drafting a blog post which basically says that not working with mozilla was GNOME's biggest mistake of the post-2.0 era. (A mistake that I played a significant role in, I might add :/ Luis _______________________________________________ fedora-advisory-board mailing list fedora-advisory-board@xxxxxxxxxx http://www.redhat.com/mailman/listinfo/fedora-advisory-board _______________________________________________ fedora-advisory-board-readonly mailing list fedora-advisory-board-readonly@xxxxxxxxxx http://www.redhat.com/mailman/listinfo/fedora-advisory-board-readonly
