Interesting.

This kind of thing can be quite useful, but I think it's got more difficult that it used to be because there are more tools out there so there are not as many small useful projects to produce a tool that does something existing tools can't do and there's the risk that not completing something will have a negative effect on morale that's as big as or bigger than the positive effect of being able to try. Thirtyfive years ago I wrote a simple interactive submitter for batch jobs and an associated stream scheduler one afternoon and it was something new and exciting and gained acceptance and wide spread use within weeks. Now you've got someone trying to produce something better that VISIO in 4 days - that's much more difficult, even if the tools he has at his disposal are a bit better than what I had way back when.