On Thu, Mar 4, 2010 at 11:23 AM, Michael Arnoldus <[email protected]> wrote:
> Thank you for this Robert.
>
> I find it frustrating how people within computer science argue for the most 
> important area being increasing programmer productivity only to shortly  
> after start discussing subjects where the person doing the work is completely 
> left out of the equation.
>
Yes, it is frustrating. My current working definition of Software
Engineering is 'Computer Science with the people back "in"' while the
slogan for our research is 'Putting the Engineer back in Software
Engineering' ;). I think that focusing on and understanding the humans
in concert with the technology might be more productive than focusing
purely on technology, on process or specific development methods.

> I suppose it's bounded in the (mis)understanding that "true science" should 
> look somewhat like physics and/or mathematics.
>
Yes, I have a strong sense that the large amounts of funding going
into formal approaches and "pure" CS stems from a kind of physics and
math envy. I think it will be more fruitful to see it as an applied
and multidisciplinary mix of present-day-CS, Psychology, Sociology and
Economics/Business.

> So it's great to see a physics quote saying that a reasonable definition of 
> complexity is highly subjective. I can see why it's tempting to look for 
> solutions where the programmer is left out - it does complicate the problem 
> considerably. However it reminds me of the old joke about the drunk looking 
> for his keys under the lamppost 
> (http://dushkablog.blogspot.com/2008/04/drunk-under-lamppost.html).
>
I think that goes to the root of it; it's conceptually cleaner to keep
them separated.

> (c) and (d) has the advantage that it would probably be possible to construct 
> experiments to test this and this gain a bit objectivity. One problem is of 
> course we have no "standard programmer" to do the testing with :-)
>
Randomization, multiple study subjects and statistical tests to the rescue...

Cheers,

Robert

_______________________________________________
fonc mailing list
[email protected]
http://vpri.org/mailman/listinfo/fonc

Reply via email to