Per discussion in IRC, we should make it hard to find projects that are essentially in the "sandbox" state - they have simply been created. I mention at http://drupal.org/node/961144#comment-3701708 the idea of an initial moderation step that allows a project to be marked as "looking for contributors" (in other words, there might be something here to actually try out or write patches against).

For projects that have not entered into this state, we should potentially disallow crawling by search engines (suggestion from EclipseGC). i'm not sure if d.o has a dynamic robots.txt file, or perhaps we could look at the UA headers and send a 401 or 403.

We coudl also use this flag to weight the project in terms of drupal.org searches. The first step would be to remove the big boost that is now set for all project nodes, and then add status-based boosts.

see: http://acquia.com/blog/delivering-the-right-search-results for some details on how to use such flag or states for search weighting.

For example:
project with releases ^20
project approved for releases ^10
looking for contributors ^5

note that these are potentially additive.

Alternatively, we could add a filter to all searches so that projects that didn't pass the initial "looking for contributors" moderation step don't show up at all in normal drupal.org searches.

Comments

eliza411’s picture

eliza411 commented
Issue tags: +sandbox projects, +git sprint 8

This may be a duplicate of other sandbox issues. Tagging for consideration in Git Sprint 8.

eliza411’s picture

eliza411 commented
Issue tags: -git sprint 8 +git sprint 9

This is critical to launch but not testing. Will get revisited during Sprint 9.

eliza411’s picture

eliza411 commented
Issue tags: -git sprint 9 +git sprint 10

Assigning to Sprint 10 since no one is willing to work on this at the moment.

pwolanin’s picture

pwolanin commented

So there are a couple things to do:

All project nodes currently get a large boost in search results due to the configurations in the UI.

Instead of continuing that, we turn off the boost for all project nodes, and instead find a marker that flags "approved" projects that have a short name and give them the boost instead. One option would be jsut to give a large boost to projects that have made a release - since we will deny that to sandboxes, right?

For how to do this in 6.x, see: http://acquia.com/blog/delivering-the-right-search-results

It might also be useful where one can filter searches for projects both with and without releases (or regular and sandbox) so that you can more readily narrow your search for duplicate projects.

eliza411’s picture

eliza411 commented

Related to #1031578: [Meta] Reflect a consistent strategy for sandbox visibility