Forum OpenACS Q&A: OpenACS CVS Statistics Report

Collapse
Posted by Roberto Mello on
Hi all,

Coming back to OACS development, and just for fun, I generated a detailed CVS statistical report for our repository since we moved out of Sourceforge (circa 2001). This doesn't include anything in the 3.x tree that lived in Sourceforge CVS from 1999-2001.

The report was generated with statcvs, a Java program, from a fresh HEAD checkout yesterday. It took quite a long time to finish (it includes stats for every single package), but it has some interesting tidbits, IMHO.

Don continues to be our code champion with 18% of lines of code commited. Dan Wickstrom, one of the co-founders of the project, still holds 3.6% of LOC committed, even after over a year of innactivity (IIRC). I will write something with the history of the project and correlate it to this report.

Thanks Don and all who make this project what it is today.

It's still uploading, but you can see it at http://www.nafpik.com/rbm/openacs/cvs-report/

-Roberto

Collapse
Posted by Miguel Marin on
I just see this report and It's fantastic, this kind of things are the ones that keep this project alive.
Collapse
Posted by Don Baccus on
Wow, what fun!  I don't expect to be at the top of the list after another year or two pass by, there are some extremely productive people chasing my lazy ass!
Collapse
Posted by Mark Aufflick on
Cool reports!

I tried to figure out how many lines of code we inherited from ArsDigita - the initial spike is 150,000 (but some would have been work done prior to moving from sourceforge).

There is a later spike of 250,000 that looks like a bulk import.

Anyone know what the real figure might be?

It also seems that I haven't done any real work since getting cvs commit rights - 361 lines of code and I almost deleted as many as I added!!

Collapse
Posted by Steve Manning on
The breakdown here http://www.nafpik.com/rbm/openacs/cvs-report/user_donb.html is interesting.

It seems your a Friday afternoon person Don. And your tucked up in bed between 22:00 and 06:00 apart from 03:00 when you obviously get up for a glass of milk and a quick commit. 😊

    - Steve

Collapse
Posted by Jeff Davis on
Mark, I have some acs 4.2 checkouts and I think 400k or so is about right for the original code base. Also, the 1.5mm lines of code are a bit deceptive in that there are some things like .eps files and .html files in docs which are big but autogenerated from something else.

There is a lot of bias in the simple line counts (eg. don and ben get a lot from initial imports, Joel from docs and the .eps files, danw and some others from the autogenerated .xql files, me from doing merges, etc).

Here is a sum of lines changed this calendar year with some attempt made to correct for the biases:

  author   | files | lines | added | removed
-----------+-------+-------+-------+---------
 lars      |  1324 | 29279 | 19436 |    9843
 joela     |  1749 | 25964 | 17309 |    8655
 peterm    |  1263 | 24303 | 17786 |    6517
 jader     |   866 | 23535 | 15609 |    7926
 jeffd     |  1148 | 21325 | 12407 |    8918
 donb      |   746 |  8921 |  4869 |    4052
 josee     |   435 |  8766 |  5429 |    3337
 daveb     |   594 |  6975 |  4882 |    2093
 maltes    |   432 |  6873 |  4332 |    2541
 timoh     |   402 |  6228 |  4131 |    2097
 dirkg     |   210 |  5864 |  2758 |    3106
 bdolicki  |   204 |  5410 |  4575 |     835
 tilmanns  |    75 |  3756 |  2868 |     888
 gyang     |   233 |  3701 |  2504 |    1197
 alfredw   |   341 |  2956 |  1128 |    1828
 andrewg   |   150 |  2942 |  2265 |     677
 janine    |   153 |  2917 |  2335 |     582
 bartt     |    39 |  1973 |  1852 |     121
 gabrielb  |   113 |  1973 |   864 |    1109
 rocaelh   |   443 |  1884 |  1652 |     232
 tracya    |   141 |  1449 |  1183 |     266
 olah      |    73 |  1070 |   620 |     450
 joel      |     1 |  1028 |   559 |     469
 juny      |    53 |   740 |   320 |     420
 alvaror   |    27 |   631 |   451 |     180
 janines   |    30 |   558 |   377 |     181
 eduardop  |    32 |   503 |    87 |     416
 rmello    |    16 |   484 |   305 |     179
 vinodk    |    12 |   283 |   103 |     180
 tils      |    26 |   209 |   186 |      23
 marka     |    39 |   171 |    96 |      75
 nimam     |     9 |   154 |    92 |      62
 jvdongen  |    17 |   151 |   119 |      32
 carlb     |     5 |   129 |    53 |      76
 skaufman  |     7 |    87 |    85 |       2
 ernieg    |   204 |    63 |    50 |      13
 jong      |     8 |    60 |    24 |      36
 leed      |     8 |    52 |    40 |      12
 jlaine    |     4 |    23 |    17 |       6
 benb      |     2 |    19 |    14 |       5
 carolinem |    94 |     4 |     2 |       2
(41 rows)
Of course, line counts do not really equate with the value of peoples contributions since some of the people who have low line counts fixed hard bugs and things which have a value much higher than the few lines they had to change would imply. The thing that is tremendous about this is the number of people who have actively contributed over the last year; I think you would find that compared to a lot of open source projects we have a much broader base of people with commit. I think a lot of the strength of OpenACS derives from our willingness to open to new contributors.
Collapse
Posted by Jade Rubick on
If these statistics are being kept often, they would be a good thing to link in from the website. These show the strength of OpenACS, and many people will chose OpenACS simply because of the quality of the community around it. I did.
Collapse
Posted by Randy O'Meara on
I think if we are equating contributor value to changes in code base, it's also important to somehow include contributions via patch submission by folks that don't have CVS commit rights...
Collapse
Posted by Andrew Piskorski on
Contributed value can't really be adequately measured by LOC, and I don't think anyone above suggested otherwise. What these CVS statistics show though, is that a lot of code has been (and presumably is being) contributed, and by a lot of different people. That in and of itself is interesting, and like Jade says, maybe useful from a technical marketing point of view.
Collapse
Posted by Roberto Mello on
If others think this is a good idea, I could automatically generate this report once a month (or once a week if desired), and upload to openacs.org.

In the same page we could perhaps include stats and links to patch submitters grabbed from bugtracker.

-Roberto

Collapse
Posted by Jeroen van Dongen on
Hmm - interesting that there are people with fewer lines than files ... especially ernieg is funny
            files  lines  added    deleted
ernieg    |  204 |    63 |    50 |      13

Does the tool even count quarter lines? 😊)

Collapse
Posted by Rocael Hernández Rizzardini on
excellent! go ahead Roberto... ;)
Collapse
Posted by Jun Yamog on
I agree that the stats does not show the real contribution of the members here.  Although people who has been here for a while already know who contributes.  I think just as Roberto has suggested to make a page monthly.  Hopefully it should attract more new members.  So the stats should be just an interesting item for us old folks, but it can be a good indicator for potential new members.

Besides everybody who has been here for a while knows the greatest fart in openacs.  He is not even on the commit list, I don't think he have even commit privs!!! hehehe

Collapse
Posted by Talli Somekh on
I resent being called a fart. I prefer Fart with a capital "F"

talli