My guess is that the brain is trying new combinations, so that if you hit a new idea during the day, it will get reinforced (you get "excited"). So the brain "shuffles" stuff during the night (dreams) and during the day selects. It is like evolutionary modification of behavior, in a way.
But to do it, you have to reinforce a little the nonsensical stuff for a while (and at speed). This semi-conscious state in which there are inconsistencies is "pre"-dream, in which almost everything is quite nonsensical, but it kind of "makes sense", so it gets reinforced a bit. The interesting thing is that this is done in parallel, so what you remember about a dream is like a small part of what your cortex was "thinking", a massive recombination of ideas.
Well, I could go on and on with my speculations... ;)
> Eventually my motivation will return to whatever it was that I should have been doing
How do you do to keep a coherent direction, then? My problem is precisely this, and I am so driven by motivation that I end up doing a kind of Brownian motion, in retrospect... ;)
2. Start using emacs (maybe with "Learning Emacs", O'Reilly, as mechanical_fish suggests).
3. Learn some Python (maybe with "Learning Python", O'Reilly too).
4. Use webpy (http://webpy.org) to develop a simple web application.
5. Learn how to program SQLite (and thus basic SQL) using Python and make your web application use a database.
6. Make your application public (http://vpslink.com might be enough, $7.95/month).
The 4th and 5th are "minimalistic" (for a web framework and a DB engine), least "complication", because the goal should be (I think) completing a simple project in a few months. From this you might get a glimpse of "the forest", or "connect the dots" (of web programming).
After that, if you want more, you can now go learn all of what others have very appropriately suggested: SICP, some system administration, PostgreSQL, Ruby (+ on Rails), git (or Bazaar), Data Structures, Algorithms, Test Driven Development, PHP, etc.
What I would love to read is an essay about "absolutism": the belief that there's only One Truth and that it can be really proved. I think many people disagree precisely about this, they think that being an absolutist is bad. Among programmers, the incidence of absolutism seems to be much higher (the closer to math, the higher?).
So I think that PG's essay is quite good (as most others), in that it looks for this "absolute truths", he tries to analyze stuff and make a contribution, say something that will last. I find this very scientific, which is good in my opinion. The problem is that many people don't see it that way. It is very difficult to stick to the principle of "egoless disagreement".
But to do it, you have to reinforce a little the nonsensical stuff for a while (and at speed). This semi-conscious state in which there are inconsistencies is "pre"-dream, in which almost everything is quite nonsensical, but it kind of "makes sense", so it gets reinforced a bit. The interesting thing is that this is done in parallel, so what you remember about a dream is like a small part of what your cortex was "thinking", a massive recombination of ideas.
Well, I could go on and on with my speculations... ;)