April 2nd, 2009

Making Long Term Sustainability a Priority

So one thing that came out of the National Science Foundation Workshop of Cyberinfrastructure Sustainability was the clear identified need to teach academic scientific researchers and their students how to use modern collaborative workflows when developing code.  The exact sort of skills and practises which make it possible for the open development system in which Fedora is deeply embedded to exist.

Scientific codebases by and large seem to live outside of those collaborative best practises and people inside of the NSF funded ecosystem are starting to come to grips with that deficiency. The workshop was a starting point that will hopefully drive a cultural change.

What's really interesting to me is that the sort of skillsets mentioned during the workshop as been critical towards long term sustainability of reusable bits of scientific oriented code..are exactly the same sort of skillsets that Google Summer of Code gives students practical experience with. I'm going to do what I can to get the Google Summer of Code directors talking with the NSF as I think there is a lot of synergy there.  We probably don't fully grasp the  long term workforce development value that Summer of Code is actually delivering in terms of spreading real world best practises about how to collaborate effectively.  We need to find ways to do more of that across as many academic fields as possible. Scientific research is becoming more and more about working with digital data, and scientists across a huge number of fields (including the humanities!) are building special purpose tools (software artifacts) in order to do their research. If collaborative best practises aren't put in place we will lose the ability to sustain and use those tools..those artifacts...as things move forward.  That will end up being a significant detriment to the public good.

Given finite resources, I'd much rather see the Fedora Project leadership double down making it easier for the Fedora Project contributors to mentor more students as part of Summer of Code and other outreach programs than to do something more superficial like add to the number of lwn subscriptions  that Fedora gives away to contributors to boost the Fedora presense on lwn.  I love having my lwn subscription, but if it ever came down to a choice between making sure another student was able to work on a summer project and my gratis lwn subscription... I'd give up my lwn subscription in a heartbeat....because that one student's potential as a collaborator is far more important to the larger open source ecosystem than me reading news.  I'd be surprised if any other Fedora contributor disagreed with me on that. Hopefully the communities which support other projects who have chosen to step away from participating in Summer of Code (after being accepted as a mentoring org) will come to the same conclusion and won't be content to be bought off with free lwn subscriptions.  

-jef