OOPSLA 2009 happened a few weeks ago. OOPSLA stands for Object-Oriented Programming, Systems, Languages & Applications and I've always been quite interested in the conference. The proceedings of the conference aren't put online, but I've managed to find two interesting papers:
A Market-Based Approach to Software Evolution (PDF) tries to imagine an open market which is targetted around fixing bugs and improving software. It's quite interesting as it's quite similar to a proposal from Nicholas on spending other people's money. The authors point out many potential flaws.
The Commenting Practice of Open Source (PDF) analyses projects on Ohloh and tries to spot commenting trends. "We find that comment density is independent of team and project size", but they find that it varies from language to language. "Java has the highest mean of comment lines per source lines at.. one comment line for three source code lines" and "Perl has the lowest mean with.. one comment line for nine source code lines". They list as future work to find out why this might be the case.
Why would we need to comment our code? I think this speaks for itself.
push@A,$_ for reverse q.e...q.n.;for(@A){$_=unpack(q|c|,$_);@a=split//;
shift@a;shift@a if $a[$[]eq$[;$_=join q||,@a};print $_,$/for reverse @A
"Java has the highest mean of comment lines per source lines at.. one comment line for three source code lines" and "Perl has the lowest mean with.. one comment line for nine source code lines"
And they say that Perl takes 10 times less code to do the same thing, so....
