Re: Languages don't matter. A mathematical refutation

["Jean François Martinez" <darkquark99@gmail.com>]

On Friday, March 27, 2015 at 6:36:51 PM UTC+1, Dmitry A. Kazakov wrote:
> On Fri, 27 Mar 2015 04:25:21 -0700 (PDT), Jean François Martinez wrote:

> Why should this imply anything for SW processes as
> performed by professionals?

Sorry the text was submitted before it was complete.  

 While thinking about the implications of Prpofessor McCormick's experience I had thought about that objection but forgot to include: it  only telle us about performance at a certain level of experience.   Of course we
could tell to people who say "Languages don't matter" that since they didn't
restrict the set to "programmers having graduated"  we win and they
lose but statisticians don't say that.  They say "we don't have data for 
market-ready programmers.  Further study is necessary".   I agree that what matters is what happens when people get paying jobs and employers (and users)
have to cope with the consequences of programmers' output.  However I think
it would be wrong to restrict us to "experienced" programmers: if during
first year after university programmers are more productive with language A and during the rest of their carreers they have identical performance with A and B then A > B.  

A final note about randomness:  you can tell the color of window shutters is
decided by house owner and probabilities don't apply.  But stastiticians consider that there is some randomeness in the  population of a city so 500,001 green shutters and 499,999 red ones vs the opposite is not considered significative (within the confidence interval or, if you prefer one guy who moved to that city) while 200,000 vs 800,000 vs the opposite points to people having different cultural habits (well outside the confidence interval).   

Since the discussion is becoming philosophical and I fear we end boring the others to death I suggest we continue it in private.

Regards.

Jean-François Martinez