One of the successes of homebrew seems to lie in how simple it is to fork the repo, create a new recipe and submit a pull request to have that recipe included in the main repo. This is something that I think is missing in the land of client-side Javascript.
At the core of BakeJS that is the idea. If you need a client-side library and an existing recipe doesn’t exist, then you fork the repo, get to Writing Recipes and then finally issue a pull request to have it included in the main repo.
It’s a simple process, and you can contribute recipes even if you aren’t the maintainer of the library. Everybody wins.
I absolutely tip my hat to Dustin Diaz for Ender. It works really well, has some great command-line tools for querying what is available and lots of other great features, however, it has one problem:
Ender packages need to be managed in NPM
Don’t get me wrong, I love NPM, and I’ve written more than a couple node modules. I just don’t think it’s the right place for client-side libraries.
A great case in point for me is a library that I do a lot with: eve. At this current point in time the npm deployed version of eve was not put there by Dmitry Baranovskiy and as such it cannot be maintained by Dmitry.
While I’m not going into the ins and outs of it here, its a problem, and working around the problem creates more clutter in NPM and confusion for all of us.