namark @namark@qoto.org

so you look up a-law and mu-law, and they are upselling them so freakin hard, with all the terminology and the sophistication of the study of the human hearing, and nature of audio signals, and speech in particular, and then follow up with formulas so backwards and mathy you have to wonder if logs are base 2 or base e... and of cource you have classic "you don't need to understand this stupid programmer, just use this lookup table" scaremongering

then you read up on the actual standard codec spec
https://en.wikipedia.org/wiki/G.711
and it turns out it's basically just

float compress(long long value)
{ return value; }

except smaller and simpler

what if the operand names for addition and subtraction were consistent with multiplication and division though?

addend + adder = sum
subtrand - subtractor = difference

programmified?

to + adder = sum
from - subtractor = difference

but then

to + added = sum
from - subtracted = difference

and it's not at all consistent anymore, unless

by * multiplied = product
by / divided = quotent

which seems bleh

#programming

how it seems:
- gimme human mod twiyoufacetubeterbook!
- no, i control the means of production, you don't get to demand nothin, algo will do for you, keep working, pleb!
- yesh milord

how it actually is:
- maybe gimme human mod sometimes faceyoutwibooktertube, huh?
- omg, i can't, am so sorry, i really can't, plz don't leave me, i improve algo, i achieve sentient AI soon, promise, plz, so sorry
- mkay, as long as you sorry, i see no problem with this, i actually hate dealing with humans if we're being honest here...
- frens? ^_^
- frens ^_^
- i protec your brand ^-^
- thonk ^-^

also noticed that it loops infinitely if the needle is empty, but the only meaningful thing to do here is to assert it being not empty as a precondition. You could maybe split character by character in that case, however that would be inconsistent with the overall behaviour of this split, as it does not ignore consecutive separators (it's a feature not a bug) and there are infinity consecutive empty strings between any two characters, or at the beginning of any string. You could also detect it and return some sentinel value, but to me that's an optional accessory that should go outside the main algorithm, because it can.

the recovery http://ix.io/3n94/cpp

why the hell is set_difference not allowed to be in-place?

#cpp

abcdxyz123 bah!

augend + addend = sum
minuend - subtrahend = difference
multiplier * multiplicand = product
dividend / divisor = quotient
dividend % divisor = reminder

#arithmetic

Well, at least std::ranges::search returns a subrange, I guess it might not be the worst thing that happened to standard library since iostreams after all.

But why doesn't it return the end iterators for both sentinels used huh, you niebloid?! Oh, perhaps because it doesn't freaking matter? Is is possible that every algorithm doesn't also need to be a find_if? OMG could it be that if I want/need an end iterator, I just wouldn't use/setup a sentinel in the first place??

#cpp

Sad c++ programmers are so afraid to write new algorithms, that they have to resort to generators.

Instead of writing
transform(begin, end, dest, f);
ima write
copy(begin, end, tranformer(dest, f));
or
copy(transformed(begin, end, f), dest);

or instead of
copy_until(begin, dest, condition);
ima write
copy(begin, sentinel(condition), dest);
where the opertor != of begin and sentinel is overloaded to be the condition

look ma, I reuse copy, am galaxy brain, let me also overload another stupid operator so that this whole mess looks like this
copy( range | trasnform(f) | filter(f2), dest );
this is so readable, ma role model is haskell, the most bestest readable and not at all cryptic language.

oh no, wait, because of filter, transform is done twice for some elements... I mean, oh yeah, this is a feature!

#cpp

trolleybus

GPU - glue programmer's utopia

if you can't write an optimal algorithm, write a suboptimal one that runs in GPU, problem solved

gee, am unoox, everutin ees uh foile, hurr durr

opinion on vector/container type deduction
https://git.sr.ht/~namark/cpp_musings/tree/master/item/vector_of_vector.cpp
#cpp

galaxy brain move
https://github.com/telegramdesktop/tdesktop/commit/69b70cda542c631c5413ae1c1fb799ed9b604a6b
I thought you can't be this smart and code in c++, but guess that's what Qt does to you.

Robots can only replace those who act like robots. Ingenuity, accountability, intuition, empathy - any one of those will keep you relevant. Employ your humanity.

Put the phone in ma pocket and it switches to airplain mode for stealth and turns on the flashlight to melt itself, while I walk around in the sun unaware. An NSA backdoor? A haxor root virums?? no! It's my arch nemesis - the touchscreen

the gibberish generator of

the happy c++98er without a supporting library: http://ix.io/3inZ/cpp

the troubled c++17er with a supporting library: http://ix.io/3ivL/cpp

#cpp

random number generators take seeding object by lvalue reference, brilliant. I get it, the library was written for c++98, but someone fix it already -_-

#cpp

A: how to such and such in c++?
B: have you heard about our lord and savior llvm rust?
A: sorry, no time.
B: oh, you'll regret this... the judgement day is nigh! the unbelievers are doomed! mark my words!

#cpp #rust

Lemon seeds carry bubbles out.