Hackathons

[kdungs]

As discussed with @betatim and @seneubert on various occasions, some LHCb-internal hackathons / week(end)s of hacking / … could prove fruitful.

Not sure about concrete implementations just putting the idea out there.

[particleist]

I like the idea of hackathons in general, but we have to be somewhat careful with respect to inclusiveness. I can see various issues to do with being cern-centric (in other words people not at CERN might struggle to get funding to travel for this), weekends being difficult for people with families, and so on.

I think one way around this is to convert some of the existing computing workshops into hackathons, at least partially. One would have to keep some of the talk-and-listen structure to keep those paying happy, but maybe it might be a step in the right direction? What do you think?

[betatim]

You could also copy the structure of the Mozilla global sprint. Several physical locations where people go, linked with video connections.

I think LHCb hackathons would be "work", therefore they should not be at the weekend. This would just further the idea that they are some kind of "geeky, free time, not to be taken serious" thing.

We should separate hackathons from sprints. Both are a good idea, but I think they are different beasts.

Hackathons: try crazy ideas, no obligations, no limits, maybe a common theme, lots of prep to think up projects, ask for X and possibly get Y, expect most projects to die after, organise incentive structure to encourage people to produce something tangible at the end (not just slides).

Sprint: either focussing on recruiting new contributors or serious nose-to-the-grindstone core contributors hacking away. In either case you need to prepare very well. Creating a suitable list of issues for new people to try out, spend a lot of time hand holding, polish your contribution guidelines before hand, make sure they work. Hardcore sprint: make sure the right people come, pick a good topic, make sure it is achievable, remove all distractions that remove one of the people you need.

[particleist]

OK so what we are talking about for this data processing workshop is definitely more "sprint" than hackathon. I need to learn my cool kids terminology...

Agree both would be good to organize anyway, however am a bit sceptical of links with video connections. We actually tried having a 1 week HLT sprint way back at the end of 2011 IIRC, with non-CERN people linked by VIDEO. By far the most productive people were the ones in the room. But maybe that was just a fluctuation...

[kdungs]

I think what Tim meant with the video connections is more of a Masterclass-like approach where you have small groups of people working together and then discuss via video.

My thoughts behind the hackathon were really less about planing and more about doing, hacking if you so want. Just get together for a defined period of time with an idea and work on it, see where you can take it. Could inspire people to try out new things they would otherwise not "have time" for. At the end, let them present their results and give out little prizes and a lot of applause for everyone.

Whether projects are continued afterwards should be a secondary consideration.

[betatim]

All in the same room is best, having a video link between different locations (each of which have a critical mass) could be a solution to the not-being-able-to-travel problem.

Not sure I understand what you mean re: planning vs doing @kdungs. You don't want a hackathon to plan things, that seems so obvious you must mean something else? You do need to plan a hackathon quite well though. Having a theme, some pre-prepared projects people could just join, organised infrastructure are a good idea. You should maintain the "if you have your own idea, just do it!" option, the pre-prepared things are to help people get started if they don't have their own idea.