[asterisk-dev] Asterisk commit stats: unscientific glance at activity

[John Todd]

I went through all of the SVN logs of the last year (2006/10/23) and 
did a quick, very unscientific examination of the volumes of commits 
to the core parts of Asterisk.  This data is fundamentally flawed 
since it conflates commit volume with patch authorship, which often 
are different.  However, it's interesting anyway and I figured I'd 
post it after my curiosity made me actually do the study.


Method:
   Take all mail messages sent to the svn-commits list (before and 
after mail server move) between 2006-10-23 and 2007-10-23.  Remove 
all messages with "trunk" in the Subject: line.  Remove all messages 
with "addons" in the Subject: line.  Remove all messages with the 
word "branch" in the Subject: line.  Sort by author listed in 
Subject: line.  Count incidents for each author.

Notes:
   1) Compares only raw volumes of commits.
   2) This counts bogus commits and repairs as doubly-weighted, since 
a fix also counts as a commit.
   3) Clearly many of these commits are patches that were written by 
others but committed by those with SVN commit privs, so this data 
obscures who actually WRITES patches for Asterisk.
   4) Some people have jobs of taking patches and putting them towards 
* core code.
   5) Some people <cough> who only had one patch had to have someone 
else re-write it the same day.
   6) No consideration is given to complexity of commits - it is 
possible that large numbers of trivial commits artificially creates 
larger apparent volumes of changes.
   7) No validity is given to my mail archives.  I believe I have all 
the messages that should be considered in this calculation - but 
maybe not.  I don't delete anything, and I have no reason to believe 
they are incomplete, but mail is not the best data store for this 
type of analysis.
   8) If I've done this incorrectly, please feel free to let me know 
how this offends you.


27 total committers, sorted by number of updates:

russell       : 768 (d)
file          : 660 (d)
tilghman      : 242 (d)
qwell         : 239 (d)
oej           : 196
mattf         : 161 (d)
murf          : 154 (d)
rizzo         : 149
kpfleming     : 145 (d)
mmichelson    : 125 (d)
tzafrir       :  64
critchter     :  44
dhubbard      :  28 (d)
phsultan      :  25
mgoroman      :  21 (d)
pcadach       :  14
dcb           :  13
transnexus    :   6
dbailey       :   6 (d)
nadi          :   4
bbryant       :   4 (d/intern)
bweschke      :   3
pari          :   1 (d)
markster      :   1 (d)
junky         :   1
jrothenberger :   1
jamesgolovich :   1


(d=current Digium employee)

- end

JT