Russ Garrett3d ago
It does feel like organising a festival would be a lot easier if it didn't involve running the equivalent of a medium-sized software business at the same time
Show threadRuss Garrett3d ago
where's the fun in that though
Show threadRuss Garrett3d ago
We have some exciting innovations in the field of conference schedule constraint solving for you this year. But also we are currently doing the refactoring equivalent of laying the track in front of a moving train.
Show threadRichard "RichiH" Hartmann

@russss If only there was a conference for conference organizers; we could call it #confconf! ;)

But in all seriousness, more conferences are starting to steal our #FOSDEM devroom concept and it's working really well for them; e.g. #IndiaFOSS did so quite successfully last year. Never been at #emf #emfcamp so I don't know the programme and space availability, of course.

May 4 at 5:44amWeb
Show threadBonkers2d ago
@RichiH @russss but how do you train the confconf organizers?
Show threadRichard "RichiH" Hartmann2d ago
@bonkers @russss fosdem ptsd, mainly
Show threadBonkers2d ago
@RichiH @russss "to be precise, I'm not a doctor. That's not a PhD in my title"
Show threadSamLR/Cass2d ago
@RichiH @russss how does the devroom work @ fosdem?
Show threadRichard "RichiH" Hartmann2d ago
@SamLR @russss you apply for a devroom. If you get one, we give you all the infrastructure and you run a small conference-in-a-conference in that room. This way, we can scale to 26(?) parallel tracks.
Show threadSamLR/Cass2d ago
@RichiH @russss ah right, I thought devroom was a method for solving "oh gods why're we running a software business" rather than specifically scheduling. We have something with villages who set their own schedule and provide methods for them to add their stuff to our schedule (which is just one source of software for us to write...)
Show threadRichard "RichiH" Hartmann2d ago
@SamLR @russss we keep everything in #pretalx these days and automate around it. That works reasonably well for us
Show threadRuss Garrett2d ago

@RichiH we ended up writing our own stuff just before Pretix and Pretalx came onto the scene, and it's all so tightly integrated and customised that it's now more work to switch, so we're kind of stuck with it all.

https://github.com/emfcamp/Website/

GitHub - emfcamp/Website: The Electromagnetic Field web site

The Electromagnetic Field web site. Contribute to emfcamp/Website development by creating an account on GitHub.

GitHub
Show threadRuss Garrett2d ago

@RichiH to my knowledge, our constraint-solving scheduler is still unique and has saved us weeks of manual scheduling so far. (At the cost of perhaps days of debugging...)

https://github.com/emfcamp/slotmachine

GitHub - emfcamp/slotmachine: A conference schedule optimiser using linear programming.

A conference schedule optimiser using linear programming. - emfcamp/slotmachine

GitHub
Show threadHugh1d ago

@russss @RichiH I looked (not very deeply) at interfacing Slot Machine with Pretalx for MCH... It's just a bit too custom, and making it work with the open-ended configuration in Pretalx was too complicated.

So we just did post-its.

Show threadRuss Garrett1d ago
@hugh @RichiH good news probably, this should make it a lot easier https://github.com/emfcamp/slotmachine/pull/16
Refactor library interface by russss · Pull Request #16 · emfcamp/slotmachine

This draws a clean boundary between the external interface, which operates in datetimes, and the solver, which operates in slots. It improves type checking and makes writing tests a lot more straig...

GitHub
Show threadRichard "RichiH" Hartmann1d ago

@russss we don't have tickets, but for talks, we were deeply integrated with pentabarf. It took us a year or so to be able to start the migration; @johanvdw will have more details.

I didn't know your specific situation of course, I can just say that paid off for us within a short amount of time.