A reading from the book of true things we all know but need to be reminded of sometimes:

Don't spend a lot of time on a decision that can be easily changed.

I found myself overthinking an architectural question: if services A and B each need a Q, is it better to spin up a multi-tenant Q server as a shared component given that it's more resource-efficient and I have a lot of experiencing managing Q; or do I create composed AQ and BQ services at a slightly higher resource cost but with simplified routing, no shared state, and the ability to manage their versions independently.

The reality is that either one of these patterns can be converted to the other with very little effort, especially if I'm aware that I might want to do so when I set it up. The even harsher reality is that A doesn't exist yet and B is purely hypothetical, conjured into existence by the ever-present urge to try to solve tomorrow's problem today.

I'm a fairly heavy user of job control in the shell. By which I mean one thing I do constantly is press control-z to suspend what I'm doing - reading a man page, editing a file, etc. - in order to run another command, and then use fg to resume.

Or: how I burned a day of my holiday break doing a peculiar kind of tedious computer maintenance.

I've decided to start a blog. Again, I guess. Once in a while, I get the urge to do this, and it hasn't stuck with me yet. I'm not sure this one will last either, because it's prompted more by the urge to play with some new software than the actual commitment to regularly write something.