Comments on: [opinion] Fedora needs an architect http://www.jonmasters.org/blog/2010/12/15/opinion-fedora-needs-an-architect/ World Organi[sz]ation Of Broken Dreams Thu, 01 Dec 2011 20:35:15 +0000 http://wordpress.org/?v=2.9 hourly 1 By: [opinion] Fedora needs an architect (part 2) « jcm's blog http://www.jonmasters.org/blog/2010/12/15/opinion-fedora-needs-an-architect/comment-page-1/#comment-133881 [opinion] Fedora needs an architect (part 2) « jcm's blog Sat, 18 Dec 2010 19:34:50 +0000 http://www.jonmasters.org/blog/?p=949#comment-133881 [...] jcm's blog World Organi[sz]ation Of Broken Dreams « [opinion] Fedora needs an architect [...] [...] jcm's blog World Organi[sz]ation Of Broken Dreams « [opinion] Fedora needs an architect [...]

]]> By: jcm http://www.jonmasters.org/blog/2010/12/15/opinion-fedora-needs-an-architect/comment-page-1/#comment-133815 jcm Wed, 15 Dec 2010 20:01:23 +0000 http://www.jonmasters.org/blog/?p=949#comment-133815 Conversely, Fedora should not be held hostage to the whims of various disconnected upstream projects. Instead, it should help steer those other projects as a major user, especially in the case that the upstream exists primarily as a result of Fedora folks to begin with. Conversely, Fedora should not be held hostage to the whims of various disconnected upstream projects. Instead, it should help steer those other projects as a major user, especially in the case that the upstream exists primarily as a result of Fedora folks to begin with.

]]> By: jcm http://www.jonmasters.org/blog/2010/12/15/opinion-fedora-needs-an-architect/comment-page-1/#comment-133814 jcm Wed, 15 Dec 2010 19:59:22 +0000 http://www.jonmasters.org/blog/?p=949#comment-133814 Take a look at other large projects with multiple upstreams that *do* handle this. A random example might be the Yocto project, which has the architect role I am suggesting, but there are many other projects. Just because there are upstreams to work with doesn't mean you can't have someone who helps figure out how it all goes together, gets involved with those upstreams to ensure the software will have the features relevant to Fedora, and works with packagers to ensure that the integration of those upstreams is appropriate later on. Take a look at other large projects with multiple upstreams that *do* handle this. A random example might be the Yocto project, which has the architect role I am suggesting, but there are many other projects. Just because there are upstreams to work with doesn’t mean you can’t have someone who helps figure out how it all goes together, gets involved with those upstreams to ensure the software will have the features relevant to Fedora, and works with packagers to ensure that the integration of those upstreams is appropriate later on.

]]> By: Radek Vokal http://www.jonmasters.org/blog/2010/12/15/opinion-fedora-needs-an-architect/comment-page-1/#comment-133805 Radek Vokal Wed, 15 Dec 2010 14:15:53 +0000 http://www.jonmasters.org/blog/?p=949#comment-133805 Honestly, I'm desperately seeking for this kind of role in Fedora as well. Fedora recently does things, that are against old Unix rules and of course, rules are here to break them, but some of the changes should have been reviewed from a broader perspective and verified with affected components. I thought that FESCo or board should stand up in such cases and drop feature that are against these rules, but that's not really happening Honestly, I’m desperately seeking for this kind of role in Fedora as well. Fedora recently does things, that are against old Unix rules and of course, rules are here to break them, but some of the changes should have been reviewed from a broader perspective and verified with affected components. I thought that FESCo or board should stand up in such cases and drop feature that are against these rules, but that’s not really happening

]]> By: Karel Zak http://www.jonmasters.org/blog/2010/12/15/opinion-fedora-needs-an-architect/comment-page-1/#comment-133801 Karel Zak Wed, 15 Dec 2010 11:36:31 +0000 http://www.jonmasters.org/blog/?p=949#comment-133801 I have a better idea :-) -- create a huge collection of *detailed* use-cases, ideally with regression tests. The distro should be defined by such use-cases and verified before each release. Jesse is right, we have no clue about upstreams changes -- if you have tests for use-cases than you can verify that upstream development matches with our (downstream) goals. I have a better idea :-)

– create a huge collection of *detailed* use-cases, ideally with regression tests. The distro should be defined by such use-cases and verified before each release.

Jesse is right, we have no clue about upstreams changes — if you have tests for use-cases than you can verify that upstream development matches with our (downstream) goals.

]]>