On the benefits of failure

The Digital Scholars Group at the U of Alberta organized a nice half-day conference On the Benefits of Failure. The first speaker was Quinn Dombrowski who spoke on her experience with different types of failure.

She had a nice survey of the types and value of failure including:

  • Technical Failure – Choosing the wrong technology.
  • DH Failure – Not getting a grant, having a course go wrong.
  • Career Failure – Getting stuck in academic inertia where you don’t do anything but what everyone else does as a grad student.
  • Arbitrary Failure – People moving on, getting jobs. There is randomness in the system.
  • Failure to Probe Assumptions – That crowdsourcing is the way to go.
  • Failure to forge a shared vision – Not really working out a vision together

She talked about how the “canonical bad digital humanities project” – which is often a good project that for the engineers seems bad. In some ways good projects are the ones that persist and iterate and advance scholarship, not the ones that follow all the best practices.

The she talked about Project Bamboo and “cyberinfrastructure to the rescue”. She talked about the variety of visions.

She talked about how some aspects of Bamboo are now showing up. Perhaps it was a matter of timing. Bamboo may have been great at “fertilizer” – getting people talking to each other that bear fruit.

She then talked about doing it herself, doing Bamboo DiRT and DH Commons. DiRT and DH Commons were both based on assumptions that seemed true, but weren’t. For example:

  • Crowdsourcing is the way to go
  • People want credit and will do stuff for credit
  • People will jump at new models
  • People who agree to be on an editorial board will do something
  • We can always do more volunteer projects just because it is fun

This led to having to let go. She had to write people to apologize when she had to let go of DiRT and DH Commons. These were hard to write, but one has to learn to put things to bed.

The last type of failure that she talked about was the most important: Failure to do right by others. There are all sorts forms of exploitation in IT.


I was struck by how many issues this raised. Here are some of the points that came up or occurred to me:

  • How can one learn from failure?
  • How can one fail gracefully?
  • How does the system (as in the University) encourage certain types of failure, or at least know that a lot of people fail at certain points (like writing their thesis)?
  • Is it failure or just a lack of sustainability? Why would not continuing a project constitute failure?
  • Why do feel bad for not working all the time?