Lightnews — Scholar-powered news

✩ milky resonances ✿ @sp1nnylights.bsky.social · 17d

i feel so sad and angry at all the immigration stuff going on. it hurts everyone. immigrants in the U.S., people in other countries, everyone in the U.S. itself, just for the sake of meaningless, pointless racism and xenophobia. i wish i had something more productive or insightful to say, just upset

7

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

it is maybe worth noting perhaps that some mammals do shed their fur in pulses of course…like how many mammals have summer and winter coats, or juvenile and adult coats—no idea if there's any direct evolutionary relation there but seems interesting to consider

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

although i'm having kind of a hard time finding a really solid reference, i'm also getting the impression that at least some early synapsids also had scutes, that hair likely developed from the keratinous layer on scutes, and that some synapsids had both (e.g. news.uchicago.edu/story/125-mi...)

125 million-year-old fossil reveals early mammalian hair and spines

news.uchicago.edu

2 2

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

other reptiles have scutes, which shed their outer keratin layers piecemeal; they may still do so in stages as opposed to continuously (e.g. i'm seeing that turtles tend to shed in pulses as they grow, but from each scute separately as opposed to in a sheath or w/e)

1 2

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

(also i feel like i should note just to be precise, i meant like, after synapsids and sauropsids split off from each other or w/e, since they're sister taxa, just so as not to confuse anyone; this is an early differentiation of amniotes)

1 3

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

okay so actually, reading more of pmc.ncbi.nlm.nih.gov/articles/PMC..., only lepidosaurs shed their skin in the way i think you're talking about, and only snakes do it all in one piece (lizards do it more in flakes or scraps)

The Process of Cornification Evolved From the Initial Keratinization in the Epidermis and Epidermal Derivatives of Vertebrates: A New Synthesis and the Case of Sauropsids - PubMed

During land adaptation of the integument in tetrapods, an efficient stratum corneum was originated through the evolution of numerous corneous proteins in addition to the framework of intermediate fila...

pubmed.ncbi.nlm.nih.gov

1 2

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

so, i'm getting the impression (from e.g. www.sciencedirect.com/science/arti..., pmc.ncbi.nlm.nih.gov/articles/PMC...) that what we think of as "reptile skin" evolved in sauropsids after they had split from synapsids

The Process of Cornification Evolved From the Initial Keratinization in the Epidermis and Epidermal Derivatives of Vertebrates: A New Synthesis and the Case of Sauropsids

During land adaptation of the integument in tetrapods, an efficient stratum corneum was originated through the evolution of numerous corneous proteins…

www.sciencedirect.com

1 3

✩ milky resonances ✿ @sp1nnylights.bsky.social · 25d

i really wonder this too…we do shed skin of course in a sense, all the time 😛 not sure how relevant that is

1 2

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

Okay here's one part for starters! This is a "text lexer" written in Ruby that takes in a text file containing English text and returns a list of tokens together with their frequencies of occurrence spinnylights.net/misc/textlex The class I wrote to do this has a bit of other potential use here too

spinnylights.net

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

(I'm just writing some additional documentation and things to make what's here easier to follow btw before I send it…this is good work for me to do at this point anyway so don't fret on that basis 😋 sorry about the wait though, it shouldn't take long)

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

I'm going to make the core logic easy to break out though ultimately since the game Lily and I are working on right now is in Unity and I'll probably use my C++ multifractal implementation as a plugin there in order to benefit from the performance attention I'm giving it now :P

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

The C++ code I've written as a part of my new game engine as a side note, so there will be built-in support for multifractals as a standalone, general-use feature in that engine :P or there already is I guess in some sense

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

The CPU-side code will also benefit from multithreading most likely, and I haven't done that yet either. All of this stuff is in a really early stage as far as performance goes basically, although on the bright side that might make for clearer code at least in terms of showing the multifractalness

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

As a side note regarding this code, I have a strong feeling (as you can probably already kind of get a sense of from what I've said so far) that it will be advantageous to do at least part of it on the GPU, but I don't have any of that yet…ultimately I will try it though, using e.g. Vulkan + compute

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

properly getting my bearings for the day at this point hopefully 😛

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

(b) in practice the self-similarity is applied at a variety of different lengths, so it doesn't necessarily just square the window length, and the easiest thing to do will simply be to figure out what size the window needs to return to post-expansion and interpolate back to that (maybe cubically)

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

also i got a little mixed up there i feel like 😅 i honestly would say like, (a) it's a little difficult to actually define what resolution means in this context maybe without arbitrary setting a length somewhere first, and

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

(🦔……… goes off to code) oh wait hold on I'll send you my current stuff XD okay I know what I'm doing next though :P

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

In some ways the problem is like that inherently, because each time you make the generator self-similar, you're basically squaring its size—/however/ (this is good thinking for me rn actually 😅) you're also basically doubling its resolution, so perhaps you can then "downscale" by half at that point

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

*quantitative 😅

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

That's only if you use a truly massive generator, like one with e.g. 1000 elements 😅 Mandelbrot's example uses a 3-element generator so I'm maybe kind of pushing it with that 😛 but I want a quantitive sense of "how much am I pushing it"

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

My current code or the ultimate finished article?

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

For example it's capable of consuming e.g. 40GB of RAM at once 😅 so memory efficiency could use some attention too

1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

(I don't promise anything close to any definition of optimal for my current code yet though on that basis 😛)

1 1

✩ milky resonances ✿ @sp1nnylights.bsky.social · 28d

Part of why I think that is that naïve implementations can suffer from rather drastic performance problems with large generators 😅 I think that's somewhat inevitable but I'm curious about /how/ inevitable 😛 and I've already gained a little ground there

1