Charlie Loyd @vruba@everything.happens.horse

@kgjenkins It takes some willpower, but it’s easier when I think about how useful old notebooks are compared to old stacks of looseleaf notes.

@kgjenkins hello

@tim That seems fair. i was trying to do stuff in JS, a language that is uh not optimized for that, and it felt pretty uphill at times.

Should I finish a devlog about this even though it’s getting a bit out of hand and probably makes errors about the math I had to learn?

@jenlowe I mean, it might fix me.

Linear algebra even more than most math feels like “Some of this is astonishing patterns in reality itself and other parts are arbitrary notational conventions that a guy named Lumbago Muttonchop von Hoobastank invented in 1852, and we’re not going to tell you which is which until you’re at the postgraduate level, teehee!”

They should make a kind of linear algebra that I can enjoy.

@pnorman Someday that will change, and that will be a new question. For now, as I expect you agree, arguing for ASCII as the only way to do strings is like arguing for 32-bit unsigned integers as the only way to do numbers. “But they’re fast and easy to reason about!”

@pnorman Right – the point is not that Unicode is perfect, because it has various obvious warts and arguable major design flaws. It’s that its utility vastly exceeds any realistic alternative’s, especially ASCII’s, for basically all user-facing software.

Very few people should implement all of ISO 8601 – it’s more complex than you might think – and I have reservations about recommending any standard that isn’t freely available anyway, but RFC 3339 is right there. Right there! And it’s really good!

Sometimes when I’m a joyless pedant about using SI units, formatting them in standard ways, etc., I’m dialing it up to 130% and making a little bit of fun of myself and others.

This time, I’m actually mad. Don’t do this kind of thing. https://mastodon.social/@lorentey/111263794546558934

@at@sigmoid.social Any opinions or reflections on this one? I skimmed it and found it odd for reasons I couldn’t put my finger on. Maybe I just didn’t understand it, or maybe it’s my relative disinterest in ViTs as currently imagined.

@dnomadb Processing artifact.

The Seattle area at night, in color, from a stack of Landsat 9 scenes. Denoised with some clumsy FFT editing. I can’t imagine why you would use this for anything over one of many other options, but it’s neat that it works.

@migurski It does run on touch devices! But there are bits I enjoy that only really come through with a big screen and a pointer.

I say objectively overkill because Euler was computationally faster, even though it took a few more steps.

Also, there’s a really wide range of h parameters that are tied for optimality. Taken together, this makes me think that it’s the lumpiness of the topography that’s the main limit here. But I’m in way over my head mathematically, so I could easily be misimplementing or misinterpreting.

Anyway, this is easily the silliest thing I’ve ever done with ground control points.

Various fixes. It now uses fourth-order Runge–Kutta integration to place the labels, which is objectively overkill and also I probably put many bugs in it, but I had fun and that’s the important thing.

@bil Reminded of this instant classic: https://m.youtube.com/watch?v=kkfIXUjkYqE

Forgot to add the warning: this is intended for, and only tested on, desktop browsers.