Hey all,

I'm working on an open source data-sciency library. It's been a little while since I've done a big JS project not for the web, and I was wondering what the current best practices and tools were considering the following:

• Library should be available both in the browser context and in node.
• Library will be object-oriented with a few different classes -- what are the current best practices around modules and module loading?
• Namespacing: how are people doing it these days?
• Test coverage: will need 100% unit test coverage, though I'm probably not too interested in BDD. Since it's a data science library I'll need to make sure the maths, algorithms, and edge cases are tested.
• Publishing: bower? NPM? Both?
• Documentation: are there good documentation generator tools that read from docblocks and render to markdown out there?
• What do you guys appreciate and want to see in an open source library?

And of course any other advice you wish to share would be helpful!