[python-uk] Future Dojo idea ... randomised trials

Carl Reynolds drcjar at gmail.com
Tue Feb 7 20:23:18 CET 2012

I'm happy to provide informal advise on trial design

Power calculations are the starting point for proper RCTs...

Michael Grazebrook <michael at grazebrook.com> wrote:

>It would be amusing to do a randomised trial, well-designed, and submit
>the results to some prestigious journal.
>Of course we'd probably be turned down. But then again, I bet there are
>very few academic studies based on the use of hardened professionals
>like us.
>Or even ... we could look for papers published by London academics in
>this area, and work with them to design an experiment which would be
>fun, informative, and have a realistic chance of being published.
>Is there an academic amongst us who can advise?
>On Feb 7, 2012 15:59 "Jonathan Hartley" <tartley at tartley.com> wrote:
>> People are always banging on about how programming lacks scientific
>> rigour when it comes to evaluating common practice.
>> Is TDD *always* faster?
>> <http://blog.8thlight.com/uncle-bob/2012/01/11/Flipping-the-Bit.html>
>> Is it *ever* faster?
>> <http://www.davewsmith.com/blog/2009/proof-that-tdd-slows-projects-dow
>> n>
>> Who knows? Fortunately, we have the tools to answer the question once
>> and for all. We, at the London Dojo, could run a randomised trial:
>> <http://www.guardian.co.uk/commentisfree/2011/sep/30/run-your-own-scie
>> ntific-trials>
>> I'm curious what results we'd get if we randomly made some Dojo teams
>> use TDD for the assignment, and others not. Our usual time contraints
>> result in a mad scramble for the finish line. Would TDD make the task
>> harder, or easier? Would the results be more functional, or less?
>> Perhaps it doesn't make sense: A team assigned to do TDD might only
>> have
>> members who were not practiced in it. But I can't help but wonder what
>> results it would produce. Is anyone else curious?
>> Jonathan
>python-uk mailing list
>python-uk at python.org

More information about the python-uk mailing list