Nice page, I see concrete useful guidance here which will make the difference in the future. I particularly appreciate «Interns must submit their contributions soon and often». Deferring everything to a big import/merge of all your code from your own branch (or worse, repo; even worse, local code) at once is extremely dangerous. --Nemo 07:26, 17 October 2013 (UTC)
Having two mentors
Is the two-mentor requirement really a "lesson learned"? Looking at the set of past GSoC projects, I don't see any strong statistical evidence that two-mentor projects have succeeded better than one-mentor projects, either at creating usable code or at retaining the student afterwards. Yaron Koren (talk) 21:30, 17 March 2014 (UTC)
Wikimedia's pyramid of reports
Does such a thing still exists? Cf. http://thread.gmane.org/gmane.org.wikimedia.foundation/75430 --Nemo 20:43, 13 January 2015 (UTC)
Proposed amendments to current workflow after GSoC 2015 and Outreachy Round 10
These are some things I felt we should have done:
- Asking all students to create "proposal" tasks in Phabricator and then merging the selected ones with the original project ideas did not go so well, in my opinion. A lot of students didn't bother copying over important information from their proposal to the main task, leaving conversations scattered in two places, leading to quite some confusion. A better way, I think, might have been to just add proposal tasks as blockers to main project task and leaving it at that. The main project task can be resolved when the project completes (i.e. the blocker is resolved).