I’d like to start playing the sax again. I’d also like to get to the gym a couple of times a week, and write some more children’s books, and tear up the paving stones in our front yard so that we can put in a garden, but you probably don’t care much about that. If you’re reading this blog, though, I hope you will care enough about what I hope to do after I’m done at the University of Toronto to help me make it happen. Funding would be nice, as would pointers to potential sources of funding, but what I need most at this point is a reality check—your thoughts and comments would be very welcome.
So: I’m involved in a lot of things right now, but the three that matter most to me are (in order):
- Software Carpentry, a crash course in software development for scientists and engineers;
- Undergraduate Capstone Open Source Projects, which gives students from universities in Canada and the US a chance to work in distributed teams on term-long software engineering projects; and
- a collection of essays on evidence-based software engineering that will be published by O'Reilly in 2010.
The second and third are unlikely to turn into full-time jobs, and I’m not even sure I want them to. At this point in my life, what I really want is to make a difference to the world my daughter will inherit, and I think my best shot of doing that is to help scientists do more research with less effort. As I’ve argued many times, teaching them basic software development skills will have more impact than any amount of petascale this or parallel that. The problem is finding someone to fund me for 12 months while I upgrade the existing course so that it can be used for self-paced study over the web. I have a plan; all I need (for some value of “all”) is half a dozen donors willing to kick in $20-25K each to cover salary, travel, video production, and what-not.
I’ve given up on getting government funding through conventional channels (I’m 0 for 5 on applications), and unfortunately, most companies doing the computational side of computational science are only interested in backing showcase “big iron” stuff (because hey, it makes sense to ask scientists to parallelize code before they even know how to modularize or test it). With 139 days on the clock as I write this, I’m open to (practical) suggestions…