Op woensdag 2 augustus 2006 15:42, schreef Khalid B:
Here is an idea, it may not fly because it is "grunt work" and resource intensive, but how about we get a few people to do what Morbus did to the top 10 (or 50 or whatever) downloaded modules?
It is not exactly fun. Nor rewarding. Turning this into issues is even less fun, especially since the topX modules all seem to have rather large issue queues with lots of rather pressing issues. Dries once made some automated scripts which could be found under your project page. Where did they go? Using these reports is a first big step. I doubt half of the developers were aware of these code reports, let alone looked at it on a regular basis! I -personally- was not even aware that they were gone! So much for good intentions :)
This can be a step towards the top tier/certified/gold modules?
In railfrog we have this policy that nothing gets into the stable branch without unit-tests. (Same in the blog system typo[1]) Without passing these tests, but also without creating these tests. This might sound like an awfull high barrier, in fact it /is/ an awfull high barrier. But AFAIK it is a required first step (demanding tests and passing tests) when you want to offer Real Quality. So, some maths: A: we have a unit test engine. But I guess, just like me, most of you thought: "yea, I should really have a look at them soon". Enforcing such tests for the top tier/certified/gold modules sounds like a good idea to me. It might get some of us/me to actually write them :) B: we know the top X modules, by pingback, download stats and more. C: Dries has or had some automated "thingies" to find common mistakes. Not sure if its possible to tie this to project module again, though. B + C makes a good list of todos for the top tier/certified/gold modules. A + C makes a nice todo for us to get some quality control in. A + B + C makes Drupal look a lot more professional to the outside world. Bèr [1] http://www.typosphere.org/trac/wiki/PatchingGuide