Re: Languages don't matter. A mathematical refutation

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

Actually I thought about it.  That experiment only tell us what programmers at student level can do with both languages.  

But if a programmer needs years of experience and practice before being able to do what a student, perhaps even a first year student, can do then language made a difference.  Specially if you are the employer of the programmer who is paying him until he gets that level of expertise.

I agree it would be better if we could a have significant sample of programmers programming the same thing in Ada and C at their first paying job.  It would be a better mesasure of their early real, ie paying, world performance.  Unfortunately this is not going to happen.  

Another similar objection would be:  _Good_ C programmers could have succeded   But if you need a _good_ programmer, one who belongs to the top 1 perscent to do what 75% of Ada programmers can do then language made a difference.

A more serious objection would be that the Ada programmers in Professor McCormick's experiment were working in the 90s with 90s era development tools while the C students were working in the 80s.  Of course the drastic improvement from the first year belies this.    Solution would be to a comparison last C year against first Ada year but then samples of size six (my hypotheis) would be too small so it could be necessary to ask Professor McCormick about the real number of groups.  From the wording he used it seemed to be significantly larger.  
Another reason to believe evolution of development environments didn't play a role is the fact that both in both the Ada and C subgroups the success rate didn't evolve with time.