Time / Location: Wikimedia Hackathon 2016, Saturday 15:00 - 16:00, Room 1.3
Etherpad Link: https://etherpad.wikimedia.org/p/WikiHack16-Redefining_Hackathons (content below in this task description)
Do we like hackathons they way they are? What can we do better? What should stay the same no mater what?
This session is intended for long term Hackathon veterans who have opinions on what works and what does not work.
Etherpad content from Wikimedia Hackathon 2016 session:
Session name: Redefining Hackathons
Meeting goal: Talk about the future of hackathons, how can we improve?
Meeting style: Problem Solving, Discussion
Phabricator task link: https://phabricator.wikimedia.org/T124262
This session is about long term changes, not about feedback or complaints specific to this 2016 edition.
Topics for discussion:
- What works?
- What does not work?
- Ideal hackathon size?
- Key roles to fill? What needs more attention?
- Should we continue to hold one big hackathon a year or would it be more beneficial to hold multiple smaller ones?
- How can we support local events?
- What documenting is missing?
What do we know about people's opinions on hackathons
- We run a survey after every hackathon, but the responses are quite specific to each event. Let's think more strategically, more longer term in this session.
- There hasn't been a survey about what people expect about hackathons in general. Maybe we should organize one.
- Asking also about the Summit, that could be useful.
- Qualitative surveys, like talking 20 minutes with a group of individuals.
- cf. the requirement gathering about moving away from Bugzilla: https://www.mediawiki.org/wiki/Project_management_tools/Review
- Having specific technical goals e.g. Community Wishlist or Wikidata tasks. Avoid "hackathon projects" that are never touched again.
- Informal, flexible, having plenty of space for changing plans based on the demands ad hoc.
- Scholarships to volunteers that come and find quiet rooms to work on what they already had in mind and now have plenty of time to work on.
- Mixing junior and senior developers, so the former can get advice, and the later are happy to help directly in fresh projects happening now.
- Showcase, a platform to present results. That brings some pressure, and some satisfaction.
What does not work
- Why the buddy system didn't continue? Last year was basically mandatory, we really pushed it, and this make many people not comfortable, or theoretical buddies never actually got to work together. Mathcing people was also very complex.
- Organization of projects and teams before the hackathon. People didn't find a clear way to show interest and get organized.
- Allow people to promote their projects beforehand, provide people clear steps to sign up and get organized.
- The information is there, but it is difficult to get an idea of all the info, and the call to actions.
- Duplication when sending links to people; have a clear structure for the Hackathon page
- Add <tl;dr> at top ("The 3 things you must do before coming here")?
- Less splitting things into sub-pages.
- Add IRC office hours, or talkpages, where newcomers can ask questions a few months beforehand.
- Skills are not visible: who knows about what? For instance, we need a designer...
- Programming languages are not visible, and in fact this defines a lot where a new developer will and will not be able to work.
- Specific instructions for newcomers beforehand. Vagrant, git, phabricator, explanation of gadgets vs tools vs extensions,
- Avoiding disruptions nearby - e.g. loud music playing during working times. Making it easy to get food at any time (not forced to get lunch within a single hour)
- After the hackathon, how to continue the work, how to pickup work that was left half-backed. (There is a connection with Wikimania hackathon, Phabricator tasks help continuity.)
Ideal hackathon size?
What about hackathons on topics/specific areas? Little ones focussed on things like content extensions or usability or whatever?
- Now we are aiming for about 150 in Wikimedia Hackathon, and Wikimania's is out of our control. (also Wikimania is problematic because the profiles are a lot more mixed, and more people is around just because Wikimania is next)
- Why 150? It is probably the maximum operational size. Larger will probably break the "one dimension" of the event, and would probably break into silos. The number of local developers (newcomers) may increase, though.
- % of newbies, it is a tricky balance. We want oureach, but too many might lower the quality of the event.
- Focusing hackathons by tech (i.e. Lua, Tools, Python, Mobile) could also help. Or a track?
- WMF and WMAT to work on this based on the specific use case of the Wikimedia Hackathon 2017
Action items with owners
- @Rfarrand to explore the possibility of organizing a (qualitative?) survey.
- @Rfarrand to explore possibility to have scholarships / prizes to send great Wikimedia Hackathon developers to Wikimania Hackathon.
- @Qgil to file a task about branding, for discussion. Possibly get chapter-hackathons to test the concept.