Comments on: Metrics for success of a DVCS http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/ Random stuff Wed, 28 Oct 2009 02:53:32 +0800 http://wordpress.org/?v=2.8.5.2 hourly 1 By: Hal http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/comment-page-1/#comment-698 Hal Sat, 02 Aug 2008 17:13:08 +0000 http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/#comment-698 James, I have better things to do than contribute patches when they will be totally ignored, regardless of whether they fix actual bugs for which bugzilla entries existed prior to my attaching a patch. Now I don't mind if after review the maintainer says this isn't the way forward, but why should I spend 5 or 6 hours coming to grips with some part of a codebase to fix a known bug in GNOME if GNOME maintainers can't be bothered to look at bugzilla then apply or reject the patch? Seriously how on earth does the choice of revision control system used affect me in this situation? It would be the same with source tarballs as with the funkiest DVCS imaginable. Ignored work is ignored. So please don't use "patch contributor" efficiency as an argument for anything unless you're going to actually do something about getting the patches, whatever method is used to generate them, reviewed. Reviewing the patches is the win, VCS is just NOT REMOTELY interesting AT ALL to "patch contributors" until that happens. Changing which Editor I use won't help me write better code. DVCS won't get patches reviewed. If it did, many of us would even use visual source safe if that would help. Oh by the way I can generate patches with cvs, svn, git, baz, bzr, hg and diff. Just about any "patch contributor" will be in the same boat. At the risk of labouring "labouring the point" the VCS I prefer is the one used by the project that reviews the patches submitted in bugzilla. James, I have better things to do than contribute patches when they will be totally ignored, regardless of whether they fix actual bugs for which bugzilla entries existed prior to my attaching a patch. Now I don’t mind if after review the maintainer says this isn’t the way forward, but why should I spend 5 or 6 hours coming to grips with some part of a codebase to fix a known bug in GNOME if GNOME maintainers can’t be bothered to look at bugzilla then apply or reject the patch?
Seriously how on earth does the choice of revision control system used affect me in this situation? It would be the same with source tarballs as with the funkiest DVCS imaginable. Ignored work is ignored.
So please don’t use “patch contributor” efficiency as an argument for anything unless you’re going to actually do something about getting the patches, whatever method is used to generate them, reviewed. Reviewing the patches is the win, VCS is just NOT REMOTELY interesting AT ALL to “patch contributors” until that happens. Changing which Editor I use won’t help me write better code. DVCS won’t get patches reviewed. If it did, many of us would even use visual source safe if that would help. Oh by the way I can generate patches with cvs, svn, git, baz, bzr, hg and diff. Just about any “patch contributor” will be in the same boat. At the risk of labouring “labouring the point” the VCS I prefer is the one used by the project that reviews the patches submitted in bugzilla.

]]> By: James Henstridge http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/comment-page-1/#comment-697 James Henstridge Sat, 02 Aug 2008 11:26:49 +0000 http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/#comment-697 xclaesse: I agree that almost any use of DVCS is a gain. The question is how much gain the various tools give, which I think would be worth investigating as there are some objective metrics that can be checked. Hal: I didn't mean "patch contributor" as a negative. Many people who submit plain patches simply have better things to do than learn about DVCS features. I think a good DVCS is one that makes it easy for such a person to start using just the features that make them more productive without needing to invest a large amount of time. jldugger: I don't have any evidence, but I believe that it should be possible to gather such evidence (either for or against). xclaesse: I agree that almost any use of DVCS is a gain. The question is how much gain the various tools give, which I think would be worth investigating as there are some objective metrics that can be checked.

Hal: I didn’t mean “patch contributor” as a negative. Many people who submit plain patches simply have better things to do than learn about DVCS features. I think a good DVCS is one that makes it easy for such a person to start using just the features that make them more productive without needing to invest a large amount of time.

jldugger: I don’t have any evidence, but I believe that it should be possible to gather such evidence (either for or against).

]]> By: jldugger http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/comment-page-1/#comment-693 jldugger Thu, 31 Jul 2008 17:31:16 +0000 http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/#comment-693 The question is, do we have any evidence that the VCS Users group has any members outside "core devel"? The question is, do we have any evidence that the VCS Users group has any members outside “core devel”?

]]> By: Hal http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/comment-page-1/#comment-692 Hal Thu, 31 Jul 2008 11:41:50 +0000 http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/#comment-692 CVS vs SVN vs DVCS is really a non-issue for me as a "patch contributor" The big win would be if maintainers of GNOME packages actually opened bugzilla, reviewed and then either rejected with reasons or applied the multitude of patches there. Patches that represent rather a lot of work done by volunteers who are not being treated with basic politeness when the patches are just ignored. Right now patch review isn't happening. Why not? Fixing that is the win. Without fixing that you're just pissing into the wind. I don't want to fix more GNOME bugs and waste more of my time if this is the attitude of GNOME devs. DVCS will not fix that and will yield zero benefit as far as getting more people more involved. As ever the start would be treating people with the proper respect, which is not at all a function of their patch quality. Are you with me on this? BTW hope all is well with you in the west and this rubbish of you being a proprietary software developer nowadays ends with a proper release under and OSI license. Maybe catch you for a beer at the next LCA. CVS vs SVN vs DVCS is really a non-issue for me as a “patch contributor”
The big win would be if maintainers of GNOME packages actually opened bugzilla, reviewed and then either rejected with reasons or applied the multitude of patches there. Patches that represent rather a lot of work done by volunteers who are not being treated with basic politeness when the patches are just ignored. Right now patch review isn’t happening. Why not? Fixing that is the win. Without fixing that you’re just pissing into the wind. I don’t want to fix more GNOME bugs and waste more of my time if this is the attitude of GNOME devs. DVCS will not fix that and will yield zero benefit as far as getting more people more involved. As ever the start would be treating people with the proper respect, which is not at all a function of their patch quality. Are you with me on this?

BTW hope all is well with you in the west and this rubbish of you being a proprietary software developer nowadays ends with a proper release under and OSI license. Maybe catch you for a beer at the next LCA.

]]> By: xclaesse http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/comment-page-1/#comment-691 xclaesse Thu, 31 Jul 2008 10:35:42 +0000 http://blogs.gnome.org/jamesh/2008/07/31/metrics-for-success-of-a-dvcs/#comment-691 In my opinion DVCS gives the exact same workflow than VCS, you can just use the diff command of your tool and attach the patch on bugzilla. DVCS gives more power to contributors/commiters that wants/need more complex things. So that's a 100% gain and nothing is lost... There is nothing to study, discuss, troll, etc. DVCS is simply better and it's obviously a shame that GNOME still use the worst system ever (equality with CVS). Any DVCS can do the job, all we need is an admin that pick a random one and make the move without asking what the community wants because the community will never agree. In my opinion DVCS gives the exact same workflow than VCS, you can just use the diff command of your tool and attach the patch on bugzilla. DVCS gives more power to contributors/commiters that wants/need more complex things. So that’s a 100% gain and nothing is lost… There is nothing to study, discuss, troll, etc. DVCS is simply better and it’s obviously a shame that GNOME still use the worst system ever (equality with CVS).

Any DVCS can do the job, all we need is an admin that pick a random one and make the move without asking what the community wants because the community will never agree.

]]>