Drupal code stats - preliminary
The image at left is a slide from a deck I'm putting together for possible lightning talk at DrupalCon SF. So far there are 11 slides showing the historical complexity of Drupal in terms of some simple metrics, and comparisons with the current release of Joomla!. The point of this is to try to understand the complexity of Drupal, both historically and in comparison with other frameworks, in terms of software metrics.
Comments, questions, pointers, and insights/opinions are very welcome.
BTW - The Joomla metrics (and the Drupal 7x) metrics are probably not quite right, yet - this is a work in progress. I'm working on metrics for Wordpress and a few other platforms also.
UpdateI presented this at DCSF - thanks to everyone who commented and discussed this with me. The attached file is slides as presented Sat Apr 17, 2010.
| Attachment | Size |
|---|---|
| Drupal-Stats.pdf | 207.79 KB |
Comments
Of course, you are correct
Download the PDF and have a look at the other 11 slides that I have so far.
As mentioned, this is a work in progress, but for Drupal code, and more generally PHP code, I can evaluate lots of other metrics, based on number and type of variables, number and type of function definitions and calls, control flow (including e.g. cyclomatic complexity). Currently I'm thinking about various process metrics, and SNA-based metrics. What questions about the Drupal code bases would you like to see answered?
I think about things like "Is Drupal becoming too Complex?". What kinds of things are you thinking about?
I like your comment about qualitative evaluation. Finding a reasonable way to correlate static metrics with qualitative measures would be great. Any thoughts about this? For example, do you have some candidates for "beautiful code", or "really hairy, complex code"? We could look at regressions for some collections of stats, or?
Cheers - Djun
Post new comment