azul

they/them 🏳️‍🌈🏳️‍⚧️:nb_flag:♾️ #pluralgang
we make games and compilers sometimes

pronouns: they/them

Public Key

npub1e9y64jhalql35k2cfhntmkvnnz5937pwrq0267578qrpty6uafzsc2vdd5

NIP-05 Address

Profile Code

nprofile1qqsvjjd2et7ls0c6t9vyme4amxfe32zclqhps84d020rsps4jdww53gpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds20a9l0

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1e9y64jhalql35k2cfhntmkvnnz5937pwrq0267578qrpty6uafzsc2vdd5

Show more details

Last Notes

2026-05-07 14:13:37 GMT

- reply

But if we stick to ZFC meta-theories, we still have the meta-meta-theoretic idea of a "standard model of PA" as a model that exists across meta-theories, the image of N in V into this version of the meta-theory.
(If we don't do this, the question "is there a model of ZFC set theory where the standard model of PA is different from the initial model of PA" has the answer "no, by definition", which isn't very interesting.)

2026-05-07 14:00:13 GMT

- reply

In both interpretations, 'N in *V is the standard model of PA. This 'N may look different externally but it looks like N in the ZFC set theory (the meta-theory).
Where things get interesting is when we have a separate notion of "standard" set in the meta-theory (i.e. if the meta-theory is IST instead of ZFC), and then a meta-theoretic notion of "standard model" means "the initial model that is standard" as opposed to just "the initial model".

2026-05-07 13:50:30 GMT

- reply

I think the problem is confusing. There are three levels here: Peano Arithmetic (the theory), models of PA which live in ZFC set theory (the metatheory), and models of ZFC set theory which live in some ambient set/class theory (the meta-meta-theory). The "standard model of PA" is both a meta-theoretic idea that is internal to ZFC set theory (usually the initial model), and a meta-meta-theoretic idea that relates different models of set theory (N in V and its images).

2026-05-07 12:19:18 GMT

- reply

Because you can't take arbitrary external subsets of sets in *V. Every infinite set in *V is going to have nonstandard elements. (You can create internal subsets following the rules of ZFC. E.g. consider a nonstandard natural n. The natural numbers less than n form a set by replacement (clamping N by n-1). But this isn't a (sub)model of PA, and it contains nonstandard numbers (e.g. n-1).)
That's why we construct W to let us build more interesting subsets.

2026-05-06 11:25:40 GMT

- reply

yeah i think IST is an example of what I said. The standard set of natural numbers in IST contains nonstandard elements, and the set of standard natural numbers is nonstandard.

2026-05-06 01:54:19 GMT

- reply

losing my mind over the highlighted sentence. in other words. in a nonstandard model of peano arithmetic, you can have "infinite" natural numbers. in some nonstandard model of set theory, the standard model of peano arithmetic has infinite natural numbers, and you need a nonstandard model of peano arithmetic to exclude them.
https://193eb646-6077-4968-9f13-271a9b475591.svc.edge.scw.cloud/media_attachments/files/116/525/122/283/924/262/original/e6c35bfe09b405ba.png

2026-05-06 01:47:52 GMT

infinite natural number glitch

2026-01-12 01:54:04 GMT

a manager can never be held accountable
therefore a manager must never make a management decision

2025-12-09 16:22:28 GMT

- reply

None of these things are monads, however, and's shortcircuiting is very similar to the sequencing you get with the option/maybe monad, which executes a sequence of commands but shortcircuits as soon as you get a none value.

2025-11-23 17:46:23 GMT

oh you downloaded your game? oh yeah? have fun. but me, i'm a ludophile, i only play games on vinyl. it's all about that game quality you know. there's just no way to replace this -- *breaks out the 25 vinyl disk collection box for Riven: The Sequel to Myst.*

2025-11-13 04:27:18 GMT

plato's allegory of the cave is really just a shadow on the wall cast by the platonic ideal of plato's allegory of the cave

2025-11-12 17:59:11 GMT

tfw someone asks someone else a question about one of your favorite subjects and you know you shouldn't jump in, so you don't, but you really want to.

2025-11-10 16:05:30 GMT

here in metric system no guns land we don't measure bad language design in footguns, we measure it in metersticks

2025-11-03 14:09:50 GMT

everyone in PL ignoring forth is such an embarrassment tbh

2025-10-20 15:13:35 GMT

- reply

lambda calculus is easy if you ignore substitution and evaluation strategies and . . .
that said, it's referentially transparent which is nice. maybe that's what they mean by "easier to reason about", relative to some other models of computation.

2025-10-04 03:55:00 GMT

beware of the rabbit hole to pipeline pipeline

2025-09-24 01:33:39 GMT

it's only a proton if it comes from the steam deck region of valve corporation, otherwise it's just a sparkling wine

2025-09-22 15:35:01 GMT

- reply

mad respect for y'all, but i prefer to stay on land.
i think i have a mixture of fear of the sea, and also, i don't really think i could enjoy the sensory experience of sailing. the wind, the waves, sea water in general, ...

2025-09-22 15:26:29 GMT

i would like it to always be autumn
(which it is now in the northern hemisphere)

2025-09-17 20:37:51 GMT

- reply

the ordinals seem like good candidates:
first, secondst, thirdst, fourthst, ...

2025-09-17 20:36:48 GMT

there shouldst be a double english where they add "st" to more wordst, not just to whilst

2025-09-15 14:29:58 GMT

december implies the existence of hexadecember

2025-09-15 05:10:58 GMT

it's not a crisis until a straight white cis man dies
it's not an epidemic until a straight white man gets lonely
it's not a recession until a rich white man cries
it's not political until a straight white cis man notices

2025-09-07 05:06:08 GMT

- reply

the ones who drive away from omelas but come back every week to fill their tank

2025-09-07 05:03:07 GMT

- reply

the ones who walk away from the ones who walk away from omelas

2025-09-07 05:01:56 GMT

- reply

the ones who pretend they walked away from omelas

2025-09-07 04:59:43 GMT

the ones who stay in omelas because there isn't anywhere else

2025-08-28 17:11:40 GMT

- reply

respectfully ... no thanks

2025-08-21 18:31:10 GMT

- reply

for the record, i think return stack manipulation in forth is unequivocally a bad idea. not just to emulate expression languages, but in general. it gets a pass only because forth is so spartan & has so few places to put data.
but a modern forth that isn't trying to be so minimal ought to give you access to many data stacks imo.
e.g. in mirth you have as many stacks as you want, with different names. >x pushes on the x stack, >foo pushes on the foo stack, bar> pops from the bar stack, etc

2025-08-21 18:21:01 GMT

- reply

the mixing of local variables, parameters, and return addresses all on the same stack causes so many security problems. e.g. https://en.wikipedia.org/wiki/Return-oriented_programming
i can only hope that in time we'll view mixing return addresses and data on the same stack with the same distaste as many other flawed practices ... like zero terminated strings, or storing passwords in plaintext.

2025-08-21 18:14:55 GMT

- reply

the advantage of using separate return and data stacks is that we can take any code that uses just the data stack and factor it out, e.g.
: add-thirty 30 + ;
is a forth word that adds 30 to the value on top of the data stack. but the corresponding "return stack oriented" version is not 30 >r plus, because calling a word puts the return address on the return stack. you'd have to wrap the definition with a return stack pop (r>) and a push (>r) as so,
: add-thirty r> 30 >r plus >r ;

2025-08-21 18:02:39 GMT

- reply

this isn't a joke btw. we could write words in forth that take parameters from the return stack, and doing so turns the forth into an expression language, because you have to invoke everything with the right number of parameters or else it crashes.
e.g.
: plus r> 2r> + >r >r ;
to invoke it i need to pass exactly two parameters on the return stack, e.g.
10 >r 20 >r plus
puts 30 on the return stack. but i can still "nest" these expressions (postfix style) e.g.
10 >r 20 >r plus 30 >r plus

2025-08-21 17:57:32 GMT

a conventional programming language is just a forth where the return stack is also used for parameter passing

2025-08-16 01:05:22 GMT

"functional programming doesn't have state or class"
except all its practicioners work in academia or finance

2025-08-10 23:15:09 GMT

what i want for software: bring back scripting. give users the steering wheel and the levers. make all the functionality discoverable & make it fun to automate.
what tech companies want: hide the steering wheel. hide the levers. hide anything that looks like code. replace everything with a chat interface.

2025-07-26 16:48:05 GMT

- reply

uxn also has finite memory so it's a finite state machine, so uxn is a regular expression;

2025-07-26 16:42:04 GMT

syntax is when syntax error; semantics is when segmentation fault (core dumped)

2025-07-20 21:33:24 GMT

Cool monoid alert!
The Kuratowski monoid represents all possible combinations of complement (c) and closure (k) operations on subsets of a topological space. (Note that the interior operation can be given as the composition ckc.) Only 14 possible operations exist.
See: https://en.wikipedia.org/wiki/Kuratowski%27s_closure-complement_problem
This diagram is from the paper:
Schwiebert, Ryan (2017). "The radical-annihilator monoid of a ring". Communications in Algebra. 45 (4): 1601–1617.
https://193eb646-6077-4968-9f13-271a9b475591.svc.edge.scw.cloud/media_attachments/files/114/887/654/041/866/117/original/b51a3830dbeaf581.png

2025-06-14 04:40:11 GMT

- reply

did you know the reason we write yen is because the first japanese-portuguese dictionaries made by missionaries thought the e syllable was ye. to this day the yen is actually pronounced en in japan. anyway. it's also not called japan. anyway...

2025-06-14 04:17:52 GMT

- reply

https://maps.app.goo.gl/seFYdMvSCBDjLP8p7

2025-06-14 04:02:40 GMT

folks have to stop reading the japanese kyo, ryo, ... as two syllables
it's not kee-oh-toe, it's kyo-to

2025-05-17 17:14:25 GMT

64 bit integers are big integers, right

2025-04-29 21:52:57 GMT

LLM-generated code is shit. don't use it. don't accept it. don't propagate it. definitely don't generate it.
if the person who submitted code couldn't be bothered to write it, don't review it. blanket rejection is not just appropriate & efficient, it's necessary.
code lives in the minds of its developers, as mental abstractions that provide leverage on the domain. generating code without that mental labor is like building a lever without a fulcrum.

2025-04-10 23:55:29 GMT

consider a binary variable, i.e. a variable which can hold one of two values. what are those values called?
boole: true, false
c programmer: 0, 1
forth programmer: 0, -1
quantum physicist: 1/2, -1/2

2025-03-13 07:25:19 GMT

- reply

levels of forth --
level 1: stack manipulation words and reverse polish notation
level 1.5: using the return stack for temporary storage
level 2 (dark path): control flow via return stack manipulation
level 2 (light path): control flow via control flow words
level 3: metaprogramming with immediate and postpone
level 3.5 (dark path): immediate words that parse the input buffer
level ∞ (dark path): rolling your own string literal word s" that works in both compiler and interpreter mode

2025-03-13 07:14:07 GMT

modality in forth is a bottomless pit of complexity ... it's funny when forth is treated like the pinnacle of simplicity

2025-02-03 17:45:54 GMT

- reply

you need a function to rebuild with a higher bound.
resize :: (m:Nat) -> BVec x n -> BVec x (n+m)
resize m [] = []
resize m (x :: xs) = x :: resize m xs
assuming that + is left-recursive

2025-02-03 16:21:51 GMT

- reply

yes
data MyList : Nat -> Type -> Type where
Nil : MyList n x
(::) : x -> MyList n x -> MyList (S n) x

2025-01-12 21:12:14 GMT

division is the most bullshit arithmetic operation
i don't think most programmers have internalized just how bullshit it is

2025-01-12 13:06:36 GMT

Important question for programmers.

2025-01-11 01:57:51 GMT

you, you come into my home? and try to use my programming language???
that's really brave of you actually, let me know if you have any questions

2025-01-01 16:34:49 GMT

oop is the app-ification of programming languages
(or maybe it's the other way around ... apps are the oop-ification of programs)

2024-12-21 19:43:54 GMT

pretending to be a person who isn't pretending to be another person -- the impostor paradox

2024-12-18 01:57:39 GMT

- reply

hmm i can't agree wið reddit on this one, i can't help pronouncing þ as unvoiced 😔

2024-12-18 01:51:02 GMT

- reply

💭 shouldn't it be "ðe Þinkpad" since the "th" in "the" is voiced

2024-12-16 16:05:57 GMT

- reply

Yep, this is it! And in this case the compiler didn't really need to use the full 64-bit multiplicative inverse, it could have used just a 32 bit one.
https://cdn.masto.host/gamedevlgbt/media_attachments/files/113/663/351/018/729/159/original/ad4166d419fa3a69.png

2024-11-18 04:58:18 GMT

I love how students in Japan who don't join a school club are said to be part of the "going home" club ( 帰宅部 = きたくぶ ). That's my club!

2024-11-07 12:52:41 GMT

it's No-vember, it's time to practice saying No a lot

2024-10-04 18:04:14 GMT

imagine inheriting this mess
https://cdn.masto.host/gamedevlgbt/media_attachments/files/113/250/463/145/928/331/original/d31511d34699b27e.png

2024-09-24 16:18:00 GMT

Thinking about multi-line string literal syntax where each line starts with a quote character ("), and only the last line has an end quote ... like it's a fantasy novel and a character is going off on a multi-paragraph monologue.
"hello this
"is a multiline string literal
"and it ends
"on this line"
The purpose of " at the beginning of each line (after the first) is so you can do indentation without affecting the contents. Also each line can be lexed independently.

2024-09-23 14:22:17 GMT

- reply

if you do care, ask around. ask how other people in your life feel about immigrants. ask about immigration reform. ask about how one could become a full-fledged citizen in your country, and what percentage of immigrants are completely barred from that.
you will be surprised (for the negative) how many "socially liberal" and "left-leaning" folks really don't care about immigrants at all. at least the right is explicit about its hatred for immigrants.

2024-09-23 14:09:45 GMT

- reply

anyway, my personal history has led me to be very distrustful of the idea behind the word "citizen", because its primary function is to discriminate against people who were born elsewhere (and whose ancestors were born elsewhere). we should collectively abolish the idea imo. but what can i do. most people don't give a fuck about immigrants.

2024-09-23 13:53:48 GMT

- reply

and these departments are very intersectional, of course. they aren't limited to one form of discrimination.
personal story time. in early 2013 i had the chance to go to the US to study and I decided not to because it would mean leaving my husband (at the time) behind, because the department of birthplace discrimination did not recognize same-sex marriages. DOMA was struck down later that year, so if I had waited a year I could have made a different choice.

2024-09-23 13:39:06 GMT

- reply

If and when universal healthcare comes to the US, it will come with sharp exemptions for women's reproductive care and trans healthcare.
In fact I'm a little surprised the extremely transphobic states have not already passed some form of universal healthcare. It worked so well in the UK.

2024-09-23 13:35:25 GMT

- reply

New social services are provided only on the condition that they can bolster the types and severity of legitimized discrimination.

2024-09-23 13:28:13 GMT

The primary function of the state is to automate and legitimize discrimination. You have the department of birthplace discrimination, the department of class discrimination, the department of race discrimination, the department of gender discrimination, ...

2024-09-17 14:09:29 GMT

explaining the difficulties of implementing an assembler by pointing out how difficult it is for a youtuber to say an exact timestamp for a video during audio recording of the same video

2024-09-10 14:33:24 GMT

- reply

"Have you ever engaged in any other activities which might indicate that you may not be considered to be a person of good character?"
Good lord. What a ridiculous question. Pretty sure we get a bunch of questions like this when we go to the US, too, but I don't remember such a ridiculously vague question on those forms.

2024-09-09 13:08:52 GMT

all software development is language development 🤔

2024-09-07 08:56:09 GMT

"If any requested extension is not supported, vkCreateDevice must return VK_ERROR_EXTENSION_NOT_PRESENT. If any requested feature is not supported, vkCreateDevice must return VK_ERROR_FEATURE_NOT_PRESENT."
So, uh, what happens if an extension and a feature are both not supported?
Nice specification you got there...

2024-09-02 12:33:18 GMT

why are QTs still not a thing :/

2024-08-27 13:48:21 GMT

- reply

is the font supposed to look like this ?
https://cdn.masto.host/gamedevlgbt/media_attachments/files/113/034/292/420/305/918/original/4c80388dfd3e67fe.png

2024-08-18 03:34:21 GMT

here's some tips for working with people for whom English is not a first language:
1. do not assume their inability to communicate in a form of English that makes you comfortable has any bearing on their intelligence or ability to do the job.
2. do not condescend or get angry if they ask questions.
3. do not use a lot of slang and idioms on purpose to trip them up.
4. do not correct them, you're not here to teach English, they're not here to learn English. stay focused on the job.

2024-08-12 02:17:34 GMT

- reply

not saying that this is what you're looking for, but, this reminds me of why i like OTT (observational type theory). it's an intensional type theory with function extensionality (actually, the type (f = g : A -> B) reduces to (x: A) -> (f x = g x : B) depending on the version of OTT, iirc) and equality is proof irrelevant (and that simplification is part of why OTT works. type coercion is free because equality is based on observational equality and is proof irrelevant).

2024-08-11 01:25:05 GMT

ok you know what's wrong with C.
` *foo = v; ` means "store the value v, which is an integer, at the address pointed to by foo".
` int *foo = v; ` means "declare the pointer foo and give it the initial value v, which is a pointer to an integer".
these may look almost the same but the meaning is completely different.

2024-08-03 00:42:22 GMT

- reply

other idealistic takes from me:
- if you buy a multiplatform game on one platform, you should be able to play it on other platforms.
- if you buy a "live service" game, it should still be playable after the publisher withdraws support.

2024-08-03 00:37:27 GMT

call me idealistic but
i believe a video game should include its soundtrack
in the base price.

2024-06-24 00:00:57 GMT

the british isles are actually much further north than the maps show. that's why they get no sun for large parts of the year.

2024-06-06 22:13:14 GMT

- reply

Lawyers love it when you say two contracts are the same if they contain 95% of the same words, ignoring word order.
https://github.com/licensee/licensee
https://cdn.masto.host/gamedevlgbt/media_attachments/files/112/571/951/851/071/420/original/5f3141278f55cda6.png

2024-06-06 22:02:52 GMT

Today I learned that github's license detection algorithm doesn't require an exact match -- you could add or remove a word, or scramble the words, and it might still claim your repo has an MIT or BSD-3 or whatever license.

2024-05-23 00:47:05 GMT

C library functions are always like: "SYNOPSIS. This function converts foos into bars depending on the user locale. ARGUMENTS. src and dest pointers must be distinct; it is undefined behavior if they are not QPU-aligned. RETURN VALUE. Returns the number of foos converted. A zero value indicates failure, or that zero foos were converted. A negative value indicates that the final foo was only partially converted (function got tired). Check this global variable to find out why."

2024-05-19 10:18:12 GMT

user: hello company, i would like to access your main product, you know, the one people come to you for, the main thing
company: HAVE YOU HEARD OF A I

2024-05-16 23:55:53 GMT

wine is symbolic of blood, it is hemoiconic

2024-05-09 05:39:35 GMT

TIL that on Windows, you can have a drive name in front of a relative path, e.g.
C:Hello\World
D:Foo\Bar
and it's different from an absolute path,
C:\Hello\World
D:\Foo\Bar
Because it will use your current working directory as a base, even if it's in a different drive. E.g. if your cwd is C:\Here then:
C:Hello\World resolves to C:\Here\Hello\World
D:Foo\Bar resolves to D:\Here\Foo\Bar
I have to wonder if this is really a good system . . .