Our Installation Profiles section has a lot of projects that never had releases and haven't been touched in years, or never even had commits: http://drupal.org/project/installation%2Bprofiles

New Drupal users sometimes gravitate towards these because they promise an out of the box "this does what I want!" experience, but such a large percentage are not maintained.

Is there some cleanup that can be performed here as a GCI task? Perhaps following guidelines from http://drupal.org/node/251466 ?

Comments

vosechu’s picture

vosechu commented

Yes!

Need a section on deliverables in particular but also need to change to the task template

stephthegeek’s picture

stephthegeek commented

Looking for some community input on the deliverable for this. It's become a bit of a wasteland and I'm not sure how precisely we should adhere to the guidelines for abandoned projects.

vosechu’s picture

vosechu commented

Yeah, that's what I was hoping for as well but I don't have that particular vision. Any idea if there's anyone already working on this area? I'll ask Grendzy but maybe we'll have to rope in some advice on IRC.

cwgordon7’s picture

cwgordon7 commented

Looks like this is still being worked out but I want to say that I think it sounds like a generally good idea - keep brainstorming! :)

vosechu’s picture

vosechu commented

Talked to Grendzy about this and he said roughly the following:
Being unmaintained isn't really enough to flag a project, it usually has to be broken in other ways.

Could we say something like:
Build a script to scrape the installation profiles and evaluate their dependent modules. After the scrape create a mapping of install profile to dependent modules and find out if any of those modules are flagged for having an outstanding security problem. Also create a mapping for some threshold of installed users such that flagged dependent modules get a big fat red flag, but dependent modules having less than 100 installed instances get some amount of yellow flagginess.

I guess this is sort of a weighted point system. Each low usage module is worth .3, Each flagged module is worth 1. Find out how many points of evil there are and do NUM_MODULES / EVIL_POINTS and you get some amount of disgrace. Disgraceful install profiles should be compiled and posted to the webmaster issue queue in batches of 10?

Other potential metrics could be, failures from the coder module, last commit date on the project, number of installs.

Script feels reminiscent of Flog which tries to detect places to pay attention to when refactoring.

That all said, Grendzy wants to talk more over lunch about this. I'll post again after lunch.

webchick’s picture

webchick
she/they
Primary language English
Location Vancouver 🇨🇦
commented
Status: Active » Needs work
vosechu’s picture

vosechu commented

I sorta think this is dead unless we can get some more community support. Is it possible we could move this to the webmaster queue or something until we get more guidance? I'm just not sure what we need to do with this.