The only worthwhile hackathons I've ever seen or participated in were related to the release of a new API, one that none of the participants had any access to b

The only worthwhile hackathons I've ever seen or participated in were related to the release of a new API, one that none of the participants had any access to before release. This arrangement benefits both the hackathon attendees and the sponsor: attendees know that they're starting from a level playing field, that if there is a competitive angle to the hackathon, then everyone will be judged on how well they use the new API in their software and not on what they wrote before they showed up. For the sponsors, it is a trial by fire, testing their API architecture, backend, documentation, and usefulness. I completely agree that the optimal result for an engineering-driven hackathon is a long list of bugs and a couple of surprisingly awesome demos. But I would add a room full of excited-about-pioneering developers, and a battle-tested developer support team.