Stylus

https://social.afront.org/@stylus

Named after my first AD&D character, ca 1991.

I'm a software developer, primarily #Python #CircuitPython

Profile picture: Drawing of an owl

Header: the city of Porto at night, with reflection in the Douro river, from a vacation in 2023

Pronouns: He/Him/They/Them
Github: @jepler
Originally Joined Mastodon: Apr 16, 2017

Public Key

npub1t92ss09zhn0t0347d82mxzlrrxmpecc5xanv0seyns0l9qumh7ssyh32jg

NIP-05 Address

stylus_at_social.afront.org@momostr.pink

Profile Code

nprofile1qqs9j4gg8j3teh4hc6lxn4dnp033ndsuuv2rwek8cvjfc8ljswdmlggpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsr2x5c5

Publishing to

offchain.pub

relay.damus.io

nos.lol

Author Public Key

npub1t92ss09zhn0t0347d82mxzlrrxmpecc5xanv0seyns0l9qumh7ssyh32jg

Show more details

Last Notes

2026-05-06 22:02:33 UTC

Falling asleep the other night I was thinking about derivatives.
I can think of two functions that are their own derivatives: f(x) = 0 and f(x) = ex.
There's a pair of functions function such that each is its *second* derivative is itself: f(x) = e-x and g(x) = -e-x.
There's a group of functions such that its *fourth* derivative is itself, f(x) = sin(x), g(x) = cos(x), h(x) = -sin(x), k(x) = -cos(x)
Are there any functions with other numbers of derivative steps before they become equal to themselves again? 3, 5, 8, etc.
In the light of day, I think I see that we could define an "n-self-derivative function" by just setting the derivatives<code>f(0) = 1 f'(0) = -1/2 f''(0) = -1/2 f'''(x) = f(x) etc </code>
and that defines a unique Maclurin series.
But .. are there functions of this form that would arise naturally?
#maths #math #mathstodon #recreationalMathematics

2026-04-30 17:47:48 UTC

Where in the world are the early reviews of the #Prusa3d #bondTech INDX toolchanger? I know to take such reviews with a tablespoon of salt, but I'm really surprised there aren't a raft of youtube videos out there from folks who got shipped early units by Prusa specifically so there would be video reviews out at the same time the thing went in the store.
Of course, since Saturday it also went back out of the store reportedly due to selling through an initial production run.
#3dPrinting

2026-03-12 16:23:58 UTC

- reply

https://en.wikipedia.org/wiki/Adam_Baldwin#GamerGate Possibly claimed to coin the term GamerGate & popularized misinfo about Zoë Quinn

2026-02-12 23:07:13 UTC

- reply

I know, the p of p-adics have to be primes .. but today I learned that you might refer to the composite adic numbers as "g-adics," these were investigated by Gauss, and in 10-adic numbers there are 4 square roots of 1: https://arxiv.org/html/2510.11559v1#S3

2026-02-12 18:05:08 UTC

- reply

> One Mayan inscription suggest that the Long Count was the truncation to the last 5 “digits” of an even longer count, and that a Long Count value such as 9.15.13.6.9 was in fact 13.13.13.13.13.13.13.13.9.15.13.6.9 in this Even Longer Count (why 13 everywhere? I don’t know!).
for some reason this makes me think of p-adic numbers. infinitely repeating 13s on the left side of the date, why not.

2026-02-07 14:26:54 UTC

- reply

This makes me think back: Around 2005 I wanted to drive XY displays with my PC and realized I had a DAC that was already driving at MHz rates: my VGA card. Of course, the blanking interval would always jerk the coordinates down to (0,0). So among my tricks was setting unreasonable video modes with the most horizontal pixels the CRTC would allow and as few lines as needed for the game's display.
It was pretty neat to be able to run MAME vector games on a proper display, with nothing more than a standard PC and a good analog scope, even if the display was never as crisp as the real thing.
https://social.afront.org/system/media_attachments/files/116/029/788/107/125/146/original/9fec4d10f98cccf9.png

2026-02-04 14:03:35 UTC

- reply

ARM CPUs literally have a ["big if true" register](https://developer.arm.com/documentation/ddi0406/c/Application-Level-Architecture/Application-Level-Programmers--Model/Execution-state-registers/Endianness-mapping-register--ENDIANSTATE )

2026-01-30 02:13:10 UTC

@npub1nf4…nqe4 re https://johncarlosbaez.wordpress.com/2026/01/28/tiny-musical-intervals/
The only problem I spot with the Python code is that the check for "non-trivial" triples is redundant. Except for <0,0,0> there should be no non-trivial triples, and that is already excluded. Instead, this means no triples closer than 10 micro-cents would be printed.
Removing this check and extending the search range gives some more values on the frontier:<code>0.000884 cents | 1+5.103919569e-07 | [-573, 237, 85] | Tenney: 1146.0 0.000187 cents | 1+1.077468461e-07 | [2033, 90, -937] | Tenney: 4351.3 0.000159 cents | 1+9.185490346e-08 | [399, -1412, 792] | Tenney: 4475.9 0.000028 cents | 1+1.589178367e-08 | [1634, 1502, -1729] | Tenney: 8029.2 0.000006 cents | 1+3.442373941e-09 | [-7470, 2791, 1312] | Tenney: 14940.0 0.000002 cents | 1+1.373041014e-09 | [-22109, 20086, -4189] | Tenney: 63671.1 </code>
none have as big a jump in cent accuracy as the quark of Baez.
#note1fz2…8j0h

2025-12-25 01:58:58 UTC

- reply

fwiw When I wrote it I had in mind the phrase "wacky tabacky" that I was riffing on.
It's entirely possible I don't know how to pronounce the snack brand "takis". In my dialect (US great plains/midwest, and very very white) I'm hearing "wacky tackies".

2025-12-25 01:18:04 UTC

- reply

Hi! A couple of thoughts from someone who is fascinated with musical tuning but in a fairly superficial way.
First, in looking at the hexagonal grid pages about 2/3 of the way through I wondered: You touch on scales with more than 12 tones, but only in the guise of equal temperament (29-TET). Would it be possible instead to develop a family of "just" intonations with 29 tones? Would it be described by a similar kind of grid, just with more elements, or would the shape be different?
Second, I went looking in OEIS for the list of "best approximation to 3/2" and found https://oeis.org/search?q=5%2c7%2c12%2c29%20id:A060528 -- but it goes 1,2,**3**,5,7,12,29,... and I can't figure out what difference between the two statements of the problem leads to 3 being/not being on the frontier of best approximations.
Finally, just an observation: It's unusual to see a discussion of tuning without an appeal to the overtone series somewhere. I assume this was a conscious decision.

2025-12-24 23:30:37 UTC

- reply

wacky takis

2025-12-15 14:39:40 UTC

I know this is a weird request but if you know of any queer-themed barbershop songs that I can buy sheet music to, please send them my way in the next 24 hours
#vocalMusic #barbershopQuartet #TTBB #queer #lazyFedi

2025-11-24 16:01:58 UTC

- reply

I've tried & failed to write a commandline program to upload to a codeberg repo via the /contents/ API... I just get a "500" response. And I've failed to find any examples of anyone else doing this. If I could just upload to a public repo without bothering with having a git repo locally that gets me most of what I would want from a gist replacement except possibly embedding. but it just doesn't quite work...

2025-11-24 14:41:22 UTC

Hey, what are we using as a replacement for github's gist functionality? Ideally it'd be something long lived with embed functionality. History, multiple files & comments are less important to me.
#lazyweb #lazyfedi

2025-11-12 14:55:56 UTC

Aurora photos from last night. Near Wildwood Lake State Wildlife Management Area north of Lincoln Nebraska.
#LNK #Aurora #auroraBorealis #northernLights
https://social.afront.org/system/media_attachments/files/115/537/292/685/867/204/original/ea8267b2cc49acd9.jpg
https://social.afront.org/system/media_attachments/files/115/537/292/684/872/353/original/da0f2ca56dad25b3.jpg
https://social.afront.org/system/media_attachments/files/115/537/292/721/196/348/original/b0d1af81e94a427f.jpg
https://social.afront.org/system/media_attachments/files/115/537/292/994/720/512/original/b3c2295c6cc0dc7b.jpg

2025-10-24 02:21:11 UTC

I had a brain idea so I made a picture of it. #pride
https://social.afront.org/system/media_attachments/files/115/426/744/982/949/787/original/768fe2b5587c38fe.jpg

2025-10-23 13:04:29 UTC

- reply

a stl file without a photo of it printed should be ineligible for such a contest.

2025-07-06 12:38:58 UTC

- reply

Now you can witness one of the first quickdraw calls from MicroPython on mac... 🎉 ⬛
#micropython #python #retrocomputing #retro68 #macintosh #m68k
https://social.afront.org/system/media_attachments/files/114/806/322/961/932/731/original/ba1b8177dc7cd9a2.mp4

2025-07-03 10:33:27 UTC

- reply

asyncio is working on mac micropython!
also (not shown here) the "sys.stdin" object is pollable, and you can double-click a python file to run it as the main code file.
#micropython
#python
#retrocomputing
#retro68
#macintosh
#m68k
https://social.afront.org/system/media_attachments/files/114/788/719/831/837/208/original/46ef0e430fd4bb64.mp4

2025-07-01 07:37:26 UTC

- reply

Are there any folks with vintage Mac programming experience who'd like to collaborate on figuring out what's next with Micropython on Mac?
I think that might be: What's the minimal set of APIs to do a fun graphical demo.
Pointers to "how to"-ish documentation in Pascal and C would be great as well, I have failed at finding this kind of stuff. (I do have Inside Macintosh but it is not a great tutorial)
#micropython
#python
#retrocomputing
#retro68
#macintosh
#m68k

2025-06-30 20:09:30 UTC

- reply

oh my yay, it now works well enough to save and edit a file, then import and run that file.
you could basically use an old m68k mac as a python3 (well, micropython) development system now.
don't mind the debug messages, they're as scared of you as you are of them.
#micropython #python #retrocomputing #retro68 #macintosh #m68k
https://social.afront.org/system/media_attachments/files/114/774/116/027/437/902/original/3ca9ebd1c7ecd08d.mp4

2025-06-30 20:03:55 UTC

- reply

yup. putting in a nasty hack so that pointers can be found at multiple-of-2 addresses makes it work a LOT better.

2025-06-30 19:20:17 UTC

- reply

I thought I had m68k GC collection working, but I don't.
There may be multiple problems. I'm worried about one possible problem though: micropython assumes pointer types are sizeof(pointer) aligned, but I'm not sure that's the case on m68k macs. and indeed the compiler tells me alignof void* is 2. When it comes to the content of the stack, this may be trouble.

2025-06-29 18:48:25 UTC

- reply

Whee! Now we can import files from the filesystem, as well as reading and writing files. Still haven't wrapped my head around how folders work, still can't enumerate files ..
as usual, work pushed to my micropython fork on github (which is called circuitpython, because you can't have two different forks of the "same" original project, argh).
#micropython #python #retrocomputing #retro68 #macintosh #m68k
https://social.afront.org/system/media_attachments/files/114/768/135/684/925/644/original/bfbad54c1ca149ce.mp4

2025-06-28 10:15:24 UTC

- reply

I hope to use adafruit_editor (simplistic python full-screen editor) as the editor for .py files.
https://github.com/adafruit/Adafruit_Learning_System_Guides/tree/main/CircuitPython_Text_Editor/adafruit_editor

2025-06-28 10:13:04 UTC

- reply

OK, well, it was much easier to hack out part of RetroConsole that was pulling in locale-sensitive routines. Now a "full features" build is under 400kB on disk and uses about 1MB when first opened (with a 100kB heap)
Next task is filesystem I/O I guess?
ugh, Unicode support? micropython is utf8 only, while mac has its own pre-standardization encoding I suppose. (yup: https://en.wikipedia.org/wiki/Mac_OS_Roman)
https://social.afront.org/system/media_attachments/files/114/760/431/811/634/705/original/426318612362de15.png

2025-06-28 08:55:02 UTC

- reply

Trying to be excited about the opportunity to write my own terminal emulator for m68k mac... and failing.

2025-06-27 18:19:47 UTC

- reply

"retro console" pulls in a substantal amount of code, maybe from libstdc++?
Oh hi, I'm std::time_get<wchar_t, std::istreambuf_iterator<wchar_t, std::char_traits<wchar_t> > >::_M_extract_via_format(std::istreambuf_iterator<wchar_t, std::char_traits<wchar_t> >, std::istreambuf_iterator<wchar_t, std::char_traits<wchar_t> >, std::ios_base&, std::_Ios_Iostate&, tm*, wchar_t const*, std::__time_get_state&) const and I'm adding 4.5kB to your program and so are each of 100 of my closest friends.

2025-06-27 17:53:06 UTC

- reply

the basic 'hello world' produces a ~750kB file when the main object file is just 872 bytes. The "retro console" object files are only about 32kB., so I guess it's coming from retro68 including a lot of stuff as "standard library". Is it supposed to be that chonky? Can it ... not ?<code> 740 -rw-r--r-- 1 jepler jepler 755712 Jun 25 18:04 HelloWorld.code.bin </code>

2025-06-27 17:42:20 UTC

- reply

It looks like the failure to start under system7 is just an artifact of a too-big executable file, caused by enabling micropython language & stdlib features willy nilly. Which prompts the question: why is the executable file already around 1MB !?

2025-06-27 16:23:09 UTC

- reply

Darn, I think all the easy parts of the micropython mac m68k port are done.
unfortunately, vfs_posix doesn't seem to be a good fit. I did manage to create a "file" but couldn't actually write data to it.
And while it was working in system6 I just tried it again with system7 and it's failing to start at all. boo.
https://github.com/jepler/circuitpython/blob/ports-m68kmac/ports/m68kmac/README.md#key-todos
#umac #micropython #retro68 #macintosh #retrocomputing

2025-06-26 13:18:09 UTC

- reply

Build updated at https://emergent.unpythonic.net/files/sandbox/micropython.dsk
Almost all of the "full features" support level of micropython are enabled, "mpz" for long integers and "float" for floating point numbers.
Backspace works in the repl. arrow keys don't.
A TODO section in the new README lists key items to work on:<li>Correctly implement GC collection (stack + registers)</li><li>Add Mac API support (e.g., quickdraw, arbitrary traps)</li><li>Support larger heap</li><li>Address GPL files (RetroConsole), GitHub Actions CI, and other issues that might prevent upstream inclusion.</li>
#umac #micropython #retro68 #macintosh #retrocomputing
https://social.afront.org/system/media_attachments/files/114/749/853/529/003/443/original/13af00685a4e4767.mp4

2025-06-25 18:48:20 UTC

Micropython on m68k mac? why not! It's reached the point of sorta working.
#umac #micropython #retro68 #macintosh #retrocomputing
https://social.afront.org/system/media_attachments/files/114/745/493/664/056/987/original/4364b6da3f7c837d.mp4

2025-06-16 14:54:37 UTC

Yayay! today my PR to build #microPython with gcc's "undefined behavior" sanitizer landed (which also means micropython is free of detected undefined behavior problems! (except for one class of warning we deliberately silenced))
I've only been working towards this goal on and off since .. geez, 2018. This is one of the earliest PRs I opened against micropython: https://github.com/micropython/micropython/pull/3799
I hope the "address sanitizer" will follow soon; I'm working on it now. One of the newer checks (stack_use_after_return) will have to be disabled because it interferes with assumptions micropython makes about how C program stacks work, but even so!

2025-05-23 08:51:48 UTC

- reply

From the arxiv paper:
> The orbit of 2017 OF201 also presents an intriguing challenge to the Planet X / Planet 9 hypothesis. Previous discoveries of TNOs with extremely wide and elongated orbits suggest an apparent clustering in the longitude of perihelia ϖ≈60∘, motivating the idea that an undetected planet may be shepherding the extreme TNOs. However, with ϖ = 306∘, the orbit of 2017 OF201 is an outlier of that clustering (Figure 2) and likely unstable to gravitational influence from this hypothesized planet. Our N-body simulations suggest that the presence of the Planet X / Planet 9 that produces the clustering will cause ejection of 2017 OF201 in a short timescale around 0.1 Gyr (Figure 4). Nevertheless, more investigation is required to decisively rule out that hypothesis.

2025-03-18 02:18:39 UTC

more pico-mac stuffs on adafruit fruit jam with rp2350
left: 512x342 resolution on 640x480 framebuffer with hstx
right: 4MB mac emulated with psram (video reversed "because of reasons")
sadly, placing the main mac RAM in psram slows emulation down considerably. I haven't tried overclocking anything yet, that's for later.
#raspberrypi #rp2350 #picomac #marchintosh #retrocomputing
https://social.afront.org/system/media_attachments/files/114/181/015/351/360/372/original/976c2ade818af33e.jpeg
https://social.afront.org/system/media_attachments/files/114/181/015/360/000/733/original/08281085fb5398da.jpeg

2025-01-24 20:50:24 UTC

- reply

looks like TE AMPLIMITE .050 series.
https://www.digikey.com/en/products/detail/te-connectivity-amp-connectors/6-5173277-6/2203215 down to the square holes
mating part: https://www.digikey.com/en/products/detail/te-connectivity-amp-connectors/5174684-6/2202989
there are also maybe idc cable versions? https://www.te.com/en/product-CAT-AM7-2250.datasheet.pdf
fwiw my (rigol) scope's logic probe attachment has a connector similar to this.
https://social.afront.org/system/media_attachments/files/113/885/283/816/430/301/original/3ed7e8ce840c9ede.png

2024-12-29 02:14:22 UTC

hey @npub1vv8…6fhk I know you want to be respectful of my mailbox .. but the checkbox here seems to have gone missing -- I'm at https://psfmember.org/civicrm/contribute/transact/?reset=1&id=45 and it affects firefox 133.0.3 on Linux & chromium 131.0.6778.108 on Linux; firefox 133.0.3 on android & chrome 131.0.6778.200 on android.
thanks in anticipation of a fix!
https://social.afront.org/system/media_attachments/files/113/733/681/979/866/020/original/f3380ab70436f52a.png

2024-12-16 14:06:43 UTC

- reply

does some feature in the current stop the decoupling layer from sliding around in X and Y as the next layer prints on top of it? or are you relying on friction to hold it there?

2024-11-15 03:38:48 UTC

What happens when you die?
Atheist: I decay
Agnostic: IDK

2024-10-14 16:48:05 UTC

In the video during the launch of #NASA #EuropaClipper you could see a silvery material wrapped around the rocket. The material wasn't still but moved around. Sometimes it seemed rhythmic, other times it looked like it was wind-blown. What's going on here and why did it look like that?
https://social.afront.org/system/media_attachments/files/113/306/787/670/258/373/original/d4513ec58e369d11.mp4

2024-09-20 12:53:49 UTC

- reply

look honestly I'd pay good money to be "haunted" by such a person for 24 hours. But it won't happen, "we've always done it this way".

2024-09-13 22:10:21 UTC

My dear friend Binah is lead plantiff in a lawsuit against Aetna for denying her gender affirming facial surgery mandated under the ACA. The lawsuit, filed this week, seeks class action status.
Press release: https://www.cohenmilstein.com/advocates-for-trans-equality-files-class-action-against-aetna-for-denying-gender-affirming-facial-surgeries/
News coverage: https://www.axios.com/2024/09/11/aetnas-denial-of-gender-affirming-care-taken-to-court
I don't understand the details, but if you've been affected by Aetna's denial of health care for FFS you may be able to join the lawsuit (there's a contact form on the law firm's page https://www.cohenmilstein.com/case-study/aetna-gender-affirming-surgery-facial-litigation/)
Please boost! 🔁
#transgender #ffs #facialFemininization #facialFeminizationSurgery #lawsuit #classAction #aetna #fuckAetna #aca #obamaCare