[spacelizard]

This article also leaves out:

- waiting for compiles

- waiting for the CI server to finish running tests

- waiting for network transfers

- waiting for VMs to spin up

- waiting for slow algorithms to process data

- et cetera

Even if it were physically and mentally possible to problem-solve 16 hours a day, we are still not at the point where our processes and machines can keep up with us, and we probably never will be. The dev-test cycle in itself is very time-consuming and repetitive. I don't know any way of solving this that doesn't involve spending even more time writing a lot of unit tests.

[auvrw]

provided there are a few tasks to handle at once, some of the time under the general "wait for compute" heading can be reduced by switching between tasks ... just recently (my dumbfoundedness over why i didn't think of this sooner being the main reason i'm posting) i realized "hey, why not clone the local repo into _another_ local repo?". the thing to figure out is how to decide whether waiting for compute or context switching will be more boring (or "less productive").

while i'm blathering, this too:

"good programmers are lazy." -- anon