July

I'm still not done with that set of essays about approximation quality, the Markov constant, and the pattern, but I'm really very close now. There's just one little essay left (The Discriminant), and I'm pretty sure it won't need to be subdivided. Where does the new part begin? Maybe you remember that I was just about to get into the random thoughts.

I said a lot about threads last time, but let me say a little more. The average thread length is roughly 4, but what does the distribution look like?

2		5
3		8
4		9
5		3
6		1		Objects and Identity
7		1		Concepts for Persistent Objects
8		1		On Authority
10		1		Rubik's Cube
16		1		Approximation Quality

My first reaction is that length 3–4 is just right for a thread. Length 2 seems too short, and length 5 or more seems too long. But, from that baseline I could go either way.

On the one hand, part of the idea of urticator.net is that it's supposed to be a network of loosely-coupled essays, not a linear narrative. When one essay depends on another, it should just link to it. The set of essays How I Cleaned My Room is a decent example of how to fit a single large thought into that framework.

On the other hand, sometimes the dependencies between essays force a set of essays into an essentially linear order. That's more or less the case with Approximation Quality. I could have made The Markov Equation stand alone, but the equation seems pointless outside the context of continued fractions. I could have made Group Elements and Matrices and the random thoughts be nominally independent subessays, but even there, there are enough dependencies that the order isn't very flexible.

Still, in the future I'll try to do a better job of keeping the thread length under control.