Many of our 'worklog' messages currently assume a lot of context

November 4, 2019

The primary way we keep track of things around here is our worklog system, which is a specific email log both of changes that we make and of how to do things (such as rebuild systems). Also, a while back I wrote about how keeping your past checklists doesn't help unless you can find them. In the process of recovering the checklist from 2015 that I was looking for, I wound up re-reading a bunch of our worklog messages from around that time, which gave me a learning lesson.

What I learned from re-reading our old messages is that most of our worklog messages assume a lot of implicit context. This makes perfect sense, since we write and send in our messages when the context is fresh in everyone's minds; it is the water that we're currently swimming in, and we're as generally oblivious to it as fish are. But looking back a few years later, that context is all gone. When I re-read our worklog messages I had to carefully rebuild the context, which I fortunately mostly could by reading enough messages from both worklog and our internal sysadmin mailing list (which is also archived, and which we try to discuss most everything on or at least mail in summaries of in-person discussions).

I don't think that we want to try to put down that implicit context in every worklog email; that would be tedious both to write and to read. But for worklog messages which we expect to refer to much later, for example as 'how we did this last time' directions, putting in much more explicit context and explanations seems like a good idea. Of course this is a bit tricky to actually do in practice for two reasons. The first is that what context needs to explained for the future isn't necessarily clear to us, since we're immersed in it. The second is that it's not always clear what worklog messages we're going to want to refer back to in general. Some things we can predict, but others may look like one-off things until they come up again. Still, I can try, and I should, especially for big things that took a lot of planning.

(This is similar to the lessons I learned from some planned power shutdowns many years ago, in part 1 and then part 2. We've had some other power shutdowns since then, and the lessons have been useful and I've tried to carry them out.)

PS: Just in general it might not hurt to put a bit more context into my worklogs, if only in the form of a link to the first iteration of any particular thing, where I'm likely to have written out a bunch of the background. Especially if I put it in as a postscript, it's easy to skip most of the time.

Written on 04 November 2019.
« Using personal ruleset recipes in uMatrix in Firefox
Systemd timer units have the unfortunate practical effect of hiding errors »

Page tools: View Source, Add Comment.
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Mon Nov 4 00:17:19 2019
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.