February 8, 2024



6:42 AM

Today’s goals:

  • Finish the last little bits of hardware registration
  • Start building the meat and potatoes.

That latter bit I’m not ready to say much about yet, but it’s going to be tricky. I thought for a moment that I’d overreached when I realized just how complicated it’s going to be, but then I thought up what I hope to be a reasonably simple and elegant solution.

We’ll see if that pans out…

This is one of those products that should be able to start off with a simple, solid core, and then grow with time. Not actually what I hoped for – I would have preferred a simple “set it and forget it” kind of business – but I’ll take it anyway.

After all, if I get this thing truly going, I can always hire developers to do that work for me!

Of course, my local production k3s host just randomly decided to go out to lunch on me; it looked like it was swapping madly, but there’s no swap on it. Not sure what caused that; couldn’t get in to find out. A hard reset seems to have solved it… for now.

What a waste of twenty minutes…

8:39 AM

Device association is in a Good Enough state to continue. I’m going to go grab breakfast, and then things will get challenging. Time to implement the hard part of this…

10:52 AM

Making good progress; have that ever-important starting point done, and now I need to get the actual bits and pieces put together. That’s where it gets complicated.

I’m feeling pretty crappy, so I’m going to take a break for a while. It’s “lunch” time anyway.

1:49 PM

Had a very long lunch; made (and ate!) some yummy soup with some freshly harvested basil, planted some oregano and chives for the future, and then went and got a very overdue haircut.

That all gave me time to think through the issues with the next phase of the project. There was (and still is) a particularly evil bit, but I’ve managed to come up with a strategy that will solve the part I need solved right now. Just need to implement it now.

And I have discovered MariaDB JSON support and virtual columns, which solves a different problem for me. My only complaint is that they don’t use the same syntactic sugar as PostgreSQL (or MySQL) does, but oh well. The likelihood of switching to postgres at this point is 50/50, but so far everyting will be easily fixed up to work with it.

I find it very interesting that you can generate a column and then index it, though. Useful stuff! Too bad one can’t update through them, though; that would be pretty damn nifty.

Oh well. Moving on…

5:28 PM

Okay, serious progress happened today. This little corner of the project is turning out to be a project all on its own. I’m currently working around that fact since this is just the proof of concept, but it’s going to require some serious thought & effort to make it flexible enough for what I’ll need in production.

To make matters worse, it’s all behind-the-scenes stuff. If it’s done right, the end user will never know just how much work went into it.

In any event, I’ve reached a good stopping point, so I’m done for the day. I’m mostly on target for having a workable PoC by the end of tomorrow, though it may stray into Monday or Tuesday depending on the depth of this little rabbit hole.

And whether or not I end up working over the weekend…

9:56 PM - “I need a life...”

Yeah, couldn’t stay away. Got some fairly critical database bits done, so hopefully I really will be able to finish this up enough tomorrow that I can start working on the hardware.