I divide the two missions into "need to know" and "want to learn" and focus on producing when building something. For example, I needed to learn how to use Cou

I divide the two missions into "need to know" and "want to learn" and focus on producing when building something. For example, I needed to learn how to use CouchDB to upgrade a web app I maintain so I spent the time necessary to get that done by working on it a piece at a time and learning what I needed to know as I went along. CouchDB does a lot that I didn't need to do, and while I want to learn more about some of it, and will get to that, I didn't need to learn about it all in-depth to be productive and get the job done. I wanted to learn a bit about using HTML5 animations to create graphic gauges for digital sensors so I spent enough time narrowly focused on that until I reached a point that accomplished what I was wanting to do. What I ended up with was a hack because the first thing I did was look for existing code examples and libraries to interface with the sensor I purchased (a magnetometer). This approach allowed me to prototype a digital compass very quickly and gain an high level understanding of what's necessary to accomplish that, and it worked surprisingly well.