Hi. Thanks for answering! I'm going to look at Bicycle Repair Man -- looks interesting. I've been writing python extensions to a text editor (http://www.sublimetext.com) and BRM may be very useful for that.

I'm really enjoying python. It seems very well designed and I can knock stuff out quickly. Thing is, I feel like I'm only doing small things, and that the benefits of static typing only really kick in for big projects. The article you linked to on artima.com makes explicit what I'm worried about ("The initial productivity gain of working with a dynamic language can decline as a project's codebase grows, and as refactoring becomes increasingly a chore.")

Unit tests; yes, but not 100% coverage. I've found 100% coverage to be too much. Getting to 100% can involve test code more complex than the situation you are testing, at which point the test becomes most suspicious, and you have to test it...

The problem I've chosen is one that favours static typing, I know, and I think a perfectly reasonable answer is 'this particular task takes more time. However, solving problem X, difficult for static languages, is easy for dynamic languages, and that'll solve more problems over the long term.'

Anyway, thanks for an interesting discussion.