↓ ↑
Found Gerbil just the other day, pretty promising-looking imho: https://cons.io/
>>2
I now understand why Scheme is unpopular.
Scheme is full of cons.
↓ ↑
There's so many of them: Chez, Chicken, Gambit, Gerbil (Gambit-based), Racket, Guile.
Compile time matters, runtime perf matters more, std lib matters quite a bit I'd say.
Anyone tried them all — or multiple of the above? Any others I've been missing?
Oh, and "not abandoned and bitrotted" also matters.
Apparently Racket has used the Chez incremental compiler since v8.0. Racket's profiler can be imported with (require profile) and Guile's is (use-modules (statprof)), haven't found something similar for Gambit so far besides `gsc` taking a `-track-scheme` parameter.
>>4
Nice, it's interesting how s7 jumps up for tests like cat and tail
↓ ↑
hello I am trying to do graphics on windows and I would very much like to do it by writing to the video buffer directly rather than with the gdi or by learning how to use openGL or what have you, ken silverman's build 2 engine is a win32 application and he says it uses self-modifying code so surely it must be accessing memory directly right?? he also says it triggers antiviruses a lot so maybe it does it with crazy hacks I wouldn't want to deal with but if there's a relatively simple way of doing it please tell me! taskete onegaishimasu!
windows
sell my graphics program for money
That smells like Microshit.
Embrace free and open source software!
>>6
Money.
>>5
Enjoy your anti-virus shutdown
Device Manager lists the physical address of the framebuffer under
Display Adapters > Resources, which you can pass to MmMapIoSpace to get
the nonpaged virtual address. You can draw directly on the framebuffer
using that address like /dev/fb0
If you want to create a window-relative framebuffer then you necessarily
have to go through the Window's API to request it before drawing, the
simplest way seems to be using BitBlt to draw a bitmap after requesting
a device context handle: https://solhsa.com/wintut/ch3.html
For self modifying code you need mmap or its windows analog.
Both support remapping memory as executable.
Only iOS doesn't allow self modifying code.
Because Apple wants to completely control what is allowed to execute.
They even ban virtual machines, including Lisp systems.
↓ ↑
There are people using that to store data?
Yes.
↓ ↑
https://www.infosecurity-magazine.com/news/us-plans-software-security-burden/
Because normgroids are too dumb to write good C99 code, everything have to be dumbed down to accommodate their mental handicaps.
Imagine the number of monkey jobs rewriting everything in Rust!!!
Unemployment crisis averted?
good I hate zoomers
https://ntrs.nasa.gov/citations/19890006940
https://nap.nationalacademies.org/read/5463/chapter/3
>>13
I doubt Rust makes you more productive. *Maybe* it helps you catch certain classes of errors ahead of time but it doesn't do it for every kind of error (even memory errors as any language with support for dynamic allocation can have what are practically equivalent to a memory leak.)
A good compiler really isn't a substitute for careful thought no matter how good it is and Rusts compiler is far from perfect.
>>16
The increased productiveness of Rust is due to the compiler catching a certain class of buggy programming at compile time. When there is a bigger cognitive burden, there is also a correlated cost in programmer time; the opposite also applies.
>>17
Logical fallacy. The cognitive burden is higher when writing Rust because programmers do not know Rust and are therefore even more likely to produce buggy code.
↓ ↑
I am thinking of using Dash because it is less bloated than other shells in the Bourne shell family (e.g. Bash and zsh).
>>36
How does csh compare to ksh or bash? I have only heard that it shouldn't be used for scripting: https://harmful.cat-v.org/software/csh
I have also heard other bad things about it: https://www.grymoire.com/unix/CshTop10.txt
Maybe tcsh has fixed some of these problems?
Right now it's zsh for interactive use and whatever /bin/sh happens to point to for scripting. Only feature zsh has that I actually use and no other shell has is the rc_quotes options, helps when writing commit messages inline so you can type "it''s" instead of "it'\''s". I actually used rc as a login shell for a little while and it works pretty good except for retarded programs like vim using shell syntax to do shit like redirection (lol) and lacking job control which I don't use much anyway, tmux does the "run multiple programs and switch between them" thing easily and better.
I just switched to ZSH after using Bash for around 2 years. I've also messed with fish, but I see a lot of people shit on it because it was written in Rust, which is fair.
I'm testing out zsh right now, but it's retarded that you need to add all this garbage to your .zshrc file to get normal keybindings to work https://wiki.archlinux.org/title/Zsh#Key_bindings
For what possible reason would home/end not be bound by default...
zsh. i was recommended it by a friend when i first switched to linux and i like it well enough and see no reason to switch currently
i spend more time in emacs than i do in the terminal though so eshell gets an honorable mention
↓ ↑
It had an entire an OS made for it http://users.atw.hu/gerigeri/DawnOS/index.html
Runs agonizingly slow and wastes more memory than GNU Emacs.
Similar project, but even more retarded:
https://github.com/Jellonator/Nandlang
>>3
Japanese proposed a 2 instructions extension
https://janders.eecg.utoronto.ca/pdfs/esl.pdf
It speeds SUBLEQ up like 5 times so it becomes practical.
>>4
which defeats the asinine simplicity purpose of atleast dawnos midori neko
im not sure what the purpose of "subleq os" was but it was on some obscure forum that was barely about niche operating systems and it looked similar to dawnos really bad interface but came long before 2017
the author was acting how do you say it like a netizen child and most of the forum didnt take him seriously
imo you basically spelled out the issue with the subleq instruction in >>1 "fixing it" by putting in more instructions fails oisc subleq is not a good instruction for "practical oisc" no matter how obvious of a choice it is this shows some of the vast fallacious contradicting logic in this project you need a extremely complex oisc instruction to make it practical not something that does rebound jmp
this paper calls it tisc vary bait and switchy or maybe it was written as a subtle joke on the unpractical subleq oisc case
i mean its using words like bloat seriously monadic can be meme worthy too but its used properly here disregard subleq is already esoteric operating system meme terminology
also the filename is esl im sure you are doing this on purpose
>>5
Very complex instruction word computer.
Consider ChatGPT. At its core is a tensor-matrix multiplication instruction.
In general, multiplying the machine's state by a matrix can produce arbitrary change.
So why not make tensor-matrix multiplication into an OISC?
Take VLIW to its maximum potential!!!
>>7
they call this meme a single functor supercombinator smd monad but its a processor instruction and it requires a processor that isnt stack backed certain mills maybe or has >99 primary stacks
in a way this is like a pseudo visc with programmable microcode much like lisp processors are but its a different design to get there
unlike niscs oisc isnt vary interesting and if the ai lab was still going strong machines would have both properly programmable visc ciscs and neuron nisc processors
↓ ↑
Does history have any precedents we could learn from of vast numbers of the members of a profession inexplicably insisting on working for free, devastating the opportunities for profit of those remaining? Should we go to their conventions and smash all their laptops like union workers barring the gates of a factory to stop anyone else working while they're on strike? Maybe lynch them wherever we find them like medieval guild members when they found someone charging less than the agreed on rate for their services? We have to do something I need to EAT for god's sake
just don't release your free software under a cuck license
I've come around to the idea of a middle ground between free and proprietary software, non-free but still open source software. E.g. something that allows personal use and modification but not wholesale redistribution. Something like a Creative Commons Non-Commercial license. Seems like a good balance between letting developers profit off their works and allowing users the freedom to use the software how they want.
how can we stop communism
You can’t
>>13
can not stop death
>>14
This is what someone with senescence says.
1% of the animal kingdom already regrows entire limbs and brains, you just need to CRISPR gud.
OP probably already intoxicated theirself to death, because he's been cancelled from gigs.
↓ ↑
cloe doesnt automatically make use of something like lm-2s programmable display for optimizations its useless for something like doom even if it compiles to macrocode
almost a can it run crysis i statement
dont forget >>5-san said scheme even with a compiler written in mlisp its a compiler written in mlisp not scheme the nonargument stops there
i used a lisp cpu as a example right before 286s since scheme processors and their hardware are modeled after lisp machines
My opinion: lazy direct translation, instead of building a custom engine. Author haven't even bothered to convert the graphics into EGA, doing the on fly translation. And EGA was important, since most VGA cards appeared late and required i386 to run their bioses.
>>11
good and fair opinion except the custom engine part but this is rather a philosophical question of what a doom port is and what a custom engine for a doom port is
my opinion is most meme doom ports are fundamentally broken like this without looking xd but this one is rather honest about alot of faults
then theres this site where the quality of programs to not get shit on should be not only what the operator would want (hairsplitting fps gamers about their fps in current times) but beyond that (theres no simple comparison here)
its much different from "can it run doom" here
a "wince smart calculator" "can run" crysis i but a "muslim mlg gamer" will dishonour off you you for suggesting serve gamer haram
you can think of it like this it runs (ceethe) it works to the target operators specification given proper hardware for the task (lipth) its near perfect (scheme)
im sure to fund a scheme machine and get it out there its going need something flashy for l*teral redd*tors corporate hackernews paypigs like a doom port written in optimal nearly purely functional scheme for it but eh youre too sane for me to go on about this and im no longer drunk
↓ ↑
i'm finally turning my shitty imageboard into a superior textboard, but i'm really retarded and can't figure out this issue. i install the MIT Scheme patch, but whenever I run make it outputs an error 2 towards what i think is the end. i'm not sure if this is because of it (probably) but whenever i try to ./init 8080 this pops up:
;Value: retry-post-form
1 ]=> (listen server (string->number (car (command-line))))
;The object #f, passed as an argument to string-length, is not the correct type.pls help thank u
i really want to use schemebbs, but i'm also considering using other scripts, but also i feel like that is somehow more trouble. plus i love scheme
Isn't the car of (command-line) the program name? (or in this case, apparently #f)
You need the server port number as a command argument.
Check this thread: https://textboard.org/prog/327
Your issue seems to be this: https://textboard.org/prog/327/4