,
4 tweets,
1 min read
Read on Twitter
Hypothesis: If you keep your source code as short as possible, you will in the process reduce it to the "Legos" your idea is comprised of, and when you need to add something, you will usually be able to using those Legos plus at most a few new ones.
I don't have a very large sample size for this, but it's been true of every program I've written. One reason it's true is that in practice the things you add aren't random. They're related to things you already have, so can share a lot of the same code.
(The reason I said Legos instead of something more concrete like modules is that in the best case the Legos are of many different types: sometimes modules, sometimes libraries, sometimes language features, sometimes how you represent data.)
(Yuck, I just noticed I wrote "comprised of." That should be "composed of.")
