For instance, choosing a method, that's easy. Doing a bunch of brainstorming and coming up with projects to apply that method too, that's easy (for me). It's that next step, where you're actually designing the infrastructure of the system that's difficult. Suddenly you're getting distracted by things like Operating Systems, web hosting, bandwidth requirements, APIs, etc. And it becomes this huge, abstract project.

It's then that you need to stay focused. Even if you only commit to 20 minutes a day, you'll be surprised how far you get. I find that really it doesn't take that long for me to take an abstract problem and reduce it down to a component I can get up and running now. This reduces the amount of mental overhead I'm juggling drastically and allows me to continue.

It's the same thing at work for me, have you ever had that experience where you finish a feature, and it's time to jump into the next one? And it's almost like you can't do it. You need to breathe, because you've just been so focused at such a low level on this one feature, and suddenly you have to open your brain up to imagine the entire f#$@ing system again! And you need to connect end points, think about data structures, algorithms, libraries.

Another thing I recommend is trying to get some work done before you go to work. Your mind is pretty refreshed after a good nights rest and a decent breakfast, but much less so after two commutes and a good 8 hours in the office.

Finally, in my area there are a ton of meetups. A lot of these are just hacker hours. You get a bit of social activity and they tend to keep my focused (since who's going to bust out Hulu and toss on Seinfeld at the monthly ViM meetup? There's someone talking about macros up there! MACROS!). I'd recommend finding some of those. The interchange of ideas, and being able to bounce ideas off other practiced programmers is another tangible benefit.

And a free beer and slice of pizza (well never is definitely the wrong word here...) never hurt anyone.