[sciencerobot]

There are a lot of problems in bioinformatics. Mainly, lack of reproducibility (ie "custom perl scripts"), poorly organized and characterized data and plenty of wheel reinvention (I heard Jim Kent, who first assembled the human genome, created his own version of wc [word of mouth, citation needed]).

The fact of the matter is that through high-throughput sequencing, microarrays, what have you, generation of biologically-meaningful results is possible.

There are a lot of problems in bioinformatics that need to be solved. Github has helped. More of bioinformaticians are learning about good software development practices, and journal reviewers are becoming more enlightened of the merits of sharing source code.