Hi,
I was curious about the statistics for the Geany project so I hacked together some Python code to peel out some values from the SVN log. If anyone cares, the following are some of what I found:
Total commits: eht16: 2466 ntrel: 2182 frlan: 413 colombanw: 87 kretek: 74 elextr: 18 peterscholtens: 9 dmaphy: 8 clytie: 2
Total lines changed: ntrel: 6195 eht16: 4174 frlan: 455 colombanw: 326 kretek: 103 elextr: 34 dmaphy: 9 peterscholtens: 9 clytie: 4
Since the log doesn't attribute patch authors normally, I was going to try and parse the "Thanks, Foo bar" messages to get counts for read-only developers from the commit messages, but it's too much trouble. If anyone is curious, the word 'thanks' occurs 410 times in the log.
Anyway, I thought others might be interested to see these numbers, since I couldn't find them through SF/otherwise.
Happy Coding, Matthew Brush
Am 03.05.2011 03:28, schrieb Matthew Brush:
Hi,
I was curious about the statistics for the Geany project so I hacked together some Python code to peel out some values from the SVN log. If anyone cares, the following are some of what I found:
Total commits: eht16: 2466 ntrel: 2182 frlan: 413 colombanw: 87 kretek: 74 elextr: 18 peterscholtens: 9 dmaphy: 8 clytie: 2
Total lines changed: ntrel: 6195 eht16: 4174 frlan: 455 colombanw: 326 kretek: 103 elextr: 34 dmaphy: 9 peterscholtens: 9 clytie: 4
Since the log doesn't attribute patch authors normally, I was going to try and parse the "Thanks, Foo bar" messages to get counts for read-only developers from the commit messages, but it's too much trouble. If anyone is curious, the word 'thanks' occurs 410 times in the log.
Anyway, I thought others might be interested to see these numbers, since I couldn't find them through SF/otherwise.
Ohloh is offering some stats too: http://www.ohloh.net/p/geany/contributors
Cheers, Frank
On 05/03/11 00:21, Frank Lanitz wrote:
Am 03.05.2011 03:28, schrieb Matthew Brush:
Anyway, I thought others might be interested to see these numbers, since I couldn't find them through SF/otherwise.
Ohloh is offering some stats too: http://www.ohloh.net/p/geany/contributors
I knew it must be somewhere out there! Thanks Frank.
Cheers, Matthew Brush
Am 03.05.2011 10:43, schrieb Matthew Brush:
On 05/03/11 00:21, Frank Lanitz wrote:
Am 03.05.2011 03:28, schrieb Matthew Brush:
Anyway, I thought others might be interested to see these numbers, since I couldn't find them through SF/otherwise.
Ohloh is offering some stats too: http://www.ohloh.net/p/geany/contributors
I knew it must be somewhere out there! Thanks Frank.
You are welcome.
And of course cgit if offering some little stats too: http://git.geany.org/geany/stats/
Cheers, Frank
On Tue, 03 May 2011 11:19:22 +0200, Frank wrote:
Am 03.05.2011 10:43, schrieb Matthew Brush:
On 05/03/11 00:21, Frank Lanitz wrote:
Am 03.05.2011 03:28, schrieb Matthew Brush:
Anyway, I thought others might be interested to see these numbers, since I couldn't find them through SF/otherwise.
Ohloh is offering some stats too: http://www.ohloh.net/p/geany/contributors
I knew it must be somewhere out there! Thanks Frank.
You are welcome.
And of course cgit if offering some little stats too: http://git.geany.org/geany/stats/
Some time back in 2010 I played a bit with mpy-svn-stats and I set it up as cronjob even, the result is at: http://geany.org/svnstats/index.html
Though I never really looked into it to get an idea of these stats are of any use and so I also never announced it. Since I upgraded the server to Debian Squeeze, the mpy-svn-stats package isn't available anymore. Not that it would hard to re-package it for Squeeze, I just don't feel like using such old and unmaintained software.
So, if anyone knows some cool SVN stats generator or Matthew wants to pimp his Python script to produce some fancy HTML, I'd happy to replace it :).
Regards, Enrico
2011/5/4 Enrico Tröger enrico.troeger@uvena.de:
On Tue, 03 May 2011 11:19:22 +0200, Frank wrote:
Am 03.05.2011 10:43, schrieb Matthew Brush:
On 05/03/11 00:21, Frank Lanitz wrote:
Am 03.05.2011 03:28, schrieb Matthew Brush:
Anyway, I thought others might be interested to see these numbers, since I couldn't find them through SF/otherwise.
Ohloh is offering some stats too: http://www.ohloh.net/p/geany/contributors
I knew it must be somewhere out there! Thanks Frank.
You are welcome.
And of course cgit if offering some little stats too: http://git.geany.org/geany/stats/
Some time back in 2010 I played a bit with mpy-svn-stats and I set it up as cronjob even, the result is at: http://geany.org/svnstats/index.html
Though I never really looked into it to get an idea of these stats are of any use and so I also never announced it. Since I upgraded the server to Debian Squeeze, the mpy-svn-stats package isn't available anymore. Not that it would hard to re-package it for Squeeze, I just don't feel like using such old and unmaintained software.
So, if anyone knows some cool SVN stats generator or Matthew wants to pimp his Python script to produce some fancy HTML, I'd happy to replace it :).
Regards, Enrico
I don't think these statistics are worth a pinch of ....
What is the purpose? If it is to identify contributors to the project then its totally distorted and shouldn't be used.
Contributors like Matthew, Jiri and many more don't get a mention. And all the translators who funnel through Frank are not mentioned in this list. At the very least it needs to grep the thanks from the changelog/commits.
Also counting the number of commits is a poor indicator of contribution, it says nothing about the scope of the changes. And whats the point of counting, unless you are the sort of sad individual who needs the reassurance that their number (or car or house or ...) is bigger than someone else's.
The project needs to thank everyone who managed to make time to do things, no matter how large or small.
Cheers Lex
On 05/03/11 17:11, Lex Trotman wrote:
2011/5/4 Enrico Trögerenrico.troeger@uvena.de:
So, if anyone knows some cool SVN stats generator or Matthew wants to pimp his Python script to produce some fancy HTML, I'd happy to replace it :).
I don't think these statistics are worth a pinch of ....
Agreed, well except they show how many commits/LOC each committer has, which is why I whipped up a little Python script for this to begin with.
What is the purpose? If it is to identify contributors to the project then its totally distorted and shouldn't be used.
Agreed. I just wanted to see how many commits were done by each committer within specific time periods, nothing more.
Contributors like Matthew, Jiri and many more don't get a mention. And all the translators who funnel through Frank are not mentioned in this list. At the very least it needs to grep the thanks from the changelog/commits.
Better yet, use a proper VCS that allows proper attribution :)
Also counting the number of commits is a poor indicator of contribution, it says nothing about the scope of the changes. And whats the point of counting, unless you are the sort of sad individual who needs the reassurance that their number (or car or house or ...) is bigger than someone else's.
LOC is a fair indicator of scope (unless it's all noisy commits). For me, the point was just to get an idea of the activity of committers, I have no desire to go any further with this. The reason I sent to the ML message was just for anyone else interested since I wasn't aware of the other sources of this information that Frank pointed out.
The project needs to thank everyone who managed to make time to do things, no matter how large or small.
And use a better VCS where you don't need to have commit access to be the --author of a patch/commit :)
Aside from that issue, at least the committers are very consistent in putting a thanks in the commit message and names/email in the THANKS file. One improvement would be to do something like Scintilla does[1] on their website, though it would be tedious to maintain unless it was automated.
I honestly have no idea what the purpose of the "Regular Contributors" section in the AUTHORS file is meant to serve. If it's meant to be committers who commit less often than "Developers", it's missing some from what's in the SVN log. If it's meant to be people who, regardless of commit access, contribute regularly, it's still missing people.
Cheers, Matthew Brush
On 4 May 2011 10:39, Matthew Brush mbrush@codebrainz.ca wrote:
On 05/03/11 17:11, Lex Trotman wrote:
2011/5/4 Enrico Trögerenrico.troeger@uvena.de:
So, if anyone knows some cool SVN stats generator or Matthew wants to pimp his Python script to produce some fancy HTML, I'd happy to replace it :).
I don't think these statistics are worth a pinch of ....
Agreed, well except they show how many commits/LOC each committer has, which is why I whipped up a little Python script for this to begin with.
Let me be clear that there is no criticism of you doing it, I hope you didn't interpret it that way, my concern is the use of such statistics, especially if published regularly.
What is the purpose? If it is to identify contributors to the project then its totally distorted and shouldn't be used.
Agreed. I just wanted to see how many commits were done by each committer within specific time periods, nothing more.
Ok
Contributors like Matthew, Jiri and many more don't get a mention. And all the translators who funnel through Frank are not mentioned in this list. At the very least it needs to grep the thanks from the changelog/commits.
Better yet, use a proper VCS that allows proper attribution :)
[...]
And use a better VCS where you don't need to have commit access to be the --author of a patch/commit :)
My concern is not about the mechanism or even really the accuracy of the statistics, more how they can be used. I recently saw a project (which I better not identify) almost destroy itself and lose many developers and contributors because such statistics were used to justify ignoring and overriding contributions when there was a disagreement. It even got to the point of commits being reverted in favour of the "more prevalent committer".
I'm not saying that Geany is in the same situation, far from it. Since there have been some rumblings around the areas of VCS and process speed, I think its best not to play with gas near the fire :-)
Aside from that issue, at least the committers are very consistent in putting a thanks in the commit message and names/email in the THANKS file. One improvement would be to do something like Scintilla does[1] on their website, though it would be tedious to maintain unless it was automated.
Yes thats a good idea. And the history is easier to maintain than just those who contributed to the latest release.
And I think that the original and major developers need to be acknowledged as well, without them we wouldn't have anything to argue over.
I honestly have no idea what the purpose of the "Regular Contributors" section in the AUTHORS file is meant to serve. If it's meant to be committers who commit less often than "Developers", it's missing some from what's in the SVN log. If it's meant to be people who, regardless of commit access, contribute regularly, it's still missing people.
Lots of them.
Cheers Lex
On Wed, 4 May 2011 13:01:35 +1000, Lex wrote:
Aside from that issue, at least the committers are very consistent in putting a thanks in the commit message and names/email in the THANKS file. One improvement would be to do something like Scintilla does [1] on their website, though it would be tedious to maintain unless it was automated.
Yes thats a good idea. And the history is easier to maintain than just those who contributed to the latest release.
Good idea. We could add a page "Thanks" which has the contents of the THANKS file from SVN. So there is no need to maintain both lists. I'll set up this page on Sunday (no time until then).
And I agree the current THANKS file in SVN is missing some contributors.
Regards, Enrico
On Wed, 4 May 2011 23:05:34 +0200, Enrico wrote:
On Wed, 4 May 2011 13:01:35 +1000, Lex wrote:
Aside from that issue, at least the committers are very consistent in putting a thanks in the commit message and names/email in the THANKS file. One improvement would be to do something like Scintilla does [1] on their website, though it would be tedious to maintain unless it was automated.
Yes thats a good idea. And the history is easier to maintain than just those who contributed to the latest release.
Good idea. We could add a page "Thanks" which has the contents of the THANKS file from SVN. So there is no need to maintain both lists. I'll set up this page on Sunday (no time until then).
Done. There is a new page link "Thanks" in the sidebar next to "Authors". The current Thanks page is simply filled with the contents of the THANKS file from SVN, synced daily. It breaks the page layout as the ChangeLog site does. This sucks but isn't the biggest problem ever I'd say. We could need a ReST/ChangeLog to PmWiki syntax converter (in PHP) :D.
Regards, Enrico