We've done them regularly at FusionAuth, every 3-4 months. I like it for a number of reasons:
* fun to see what other people experiment with
* lets folks investigate latest tech without having to justify incorporating it into core work
* offers freedom to explore something with low risk
We've definitely built things that made it to production, though often it was more the concept than the code. It's also been great for exploratory work that let us understand the domain or possible solutions more.
Examples of stuff from our last hackathon:
* building a rust SDK
* re-imagining in-app search
* hacking on a low-level different approach to building our client libraries
* multiple approaches to simplifying the upgrade process
* setting up a FusionAuth themed appsec CTF
I love them myself, but we do timebox them to 1 day of effort. When I used to run them, I also said that having nothing to show but learning about what didn't work was a perfectly valid use of a hackfest day.
* fun to see what other people experiment with
* lets folks investigate latest tech without having to justify incorporating it into core work
* offers freedom to explore something with low risk
We've definitely built things that made it to production, though often it was more the concept than the code. It's also been great for exploratory work that let us understand the domain or possible solutions more.
Examples of stuff from our last hackathon:
* building a rust SDK
* re-imagining in-app search
* hacking on a low-level different approach to building our client libraries
* multiple approaches to simplifying the upgrade process
* setting up a FusionAuth themed appsec CTF
I love them myself, but we do timebox them to 1 day of effort. When I used to run them, I also said that having nothing to show but learning about what didn't work was a perfectly valid use of a hackfest day.