September 21, 2026 at 12:00 am
Comments posted to this topic are about the item Make It Routine
September 21, 2026 at 5:02 pm
The harder a technical task is, the more value there is in spending time simplifying it first, then automating it. Especially if it resists simplification.
A black box is great until it breaks. That is why I like to know how something works. This is where sensible naming conventions and well-structured code become essential. If I automate something, I comment it where I think people are going to struggle.
September 21, 2026 at 6:08 pm
My thought process on automating something generally falls into the "how often do I do this / use this vs how hard is it to rewrite from memory / do as an ad-hoc process each time." Sure, I could automate the "infrequent / easy" items, and I'd almost certainly feel a sense of accomplishment from doing such, but how much time did it take me?
One item I scripted out that's an "infrequent / significant(ish) effort each time" was SQL installs. Tracked down the various command-line options, made notes as to which are or are not needed based on the requirements of the hosted databases, and saved the doc. Sure for new versions of SQL there might be some manual tweaking, but knowing such a migration is coming and starting from a ready-to-go command line, it takes a lot less time than clicking through the GUI installer. In an ideal world, yes, we'd be pure Azure Database with Bicep files to tweak and use to stand up new SQL instances, but not yet...
But some of the other things? So infrequent or it takes less time to just whip up the query fresh each time, it's not really worth the effort to automate...
September 21, 2026 at 6:28 pm
Interesting views. There certainly can be value in studying, or slowing down.
I think as we gain experience, we also get an idea of how much work is involved in automation. OFten I can guess something is quick or slow, even if I don't know the details, because I have an idea of the complexity in automating it. There is an ROI calculation to be made. If something will take 10 hours to automate, but it's a hour a year task, is it worth it? No.
At an hour a month? maybe. but my 10 hours might be low, so is it worth it.
Like Jason, installs might be worth it or might not. If I hand it off, and/or we do it every week, perhaps it is. If it's the occasional item, maybe not. The click through is quick, though this is a place I think AI can help. Point it at a system and "create the install config file" is an easy task that humans hate doing, but an AI can do well.
September 22, 2026 at 2:54 pm
Sometimes it is the tasks that are done infrequently that benefit most from being automated. Automating a task ensures that it is done consistently each time. It also de-risks the likelihood that the knowledge of how to perform the task is locked away in one person's head.
Viewing 5 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply