Valeyard posted:this reminds me that pandas is still on version like 0.2 and still heavily in production use lol on atomic level, core pandas functionality - the data storage - is based primary on numpy arrays, which are like 25 years old and were originally written by dude who later made both jython and ironpython (same single guy). now pandas api though, well, uhh, yeah, it is what it is and i dont envy you too much
|
|
# ? Dec 24, 2017 21:06 |
|
|
# ? Jun 1, 2024 22:55 |
i mean its a bit fallacious to call whatever major version of numpy arrays we are on rn (3 or 4, maybe even more) to be the same old good thing, but you get the point. numpy is a know good quality, has C api, etc
|
|
# ? Dec 24, 2017 21:11 |
|
numpy's good as hell in contrast eigen's docs actually have to say "no you cannot easily slice arrays here, sorry"
|
# ? Dec 25, 2017 05:31 |
|
stop doing heavy numerical computational work in python. just stop. it's wrong, you're wrong, everything you're doing is wrong WRONG WRONG
|
# ? Dec 25, 2017 05:47 |
|
i know people that use pandas just because it has a to_json() function lol
|
# ? Dec 25, 2017 05:49 |
|
DELETE CASCADE posted:stop doing heavy numerical computational work in python. just stop. it's wrong, you're wrong, everything you're doing is wrong WRONG WRONG lol this guy
|
# ? Dec 25, 2017 05:54 |
|
Glorgnole posted:numpy's good as hell true as hell, good as hell the non-tensorial matrix poo poo can go kill itself tho .dot all the way
|
# ? Dec 25, 2017 05:55 |
|
blah blah blah "it's a C library so really the actual work isn't done in python!!!" i don't loving care, if you are writing a complicated enough program that you need these libraries, then you should get as far away from python and other garbage p-langs as you can, as fast as you can. it's only going to get worse, not better
|
# ? Dec 25, 2017 05:56 |
|
DELETE CASCADE posted:blah blah blah "it's a C library so really the actual work isn't done in python!!!" i don't loving care, if you are writing a complicated enough program that you need these libraries, then you should get as far away from python and other garbage p-langs as you can, as fast as you can. it's only going to get worse, not better Acting like the numerical stuff in non p langs is worth a poo poo in any lang except Fortran Lol working in Fortran
|
# ? Dec 25, 2017 06:05 |
|
so I got gambit installed since it was like the only lisp/scheme/whatever i could find that actually works on WSL (SBCL crashes due to something in /proc missing, the rest just crash on startup) and i've got most of a wavefront obj -> binary data script together i bet there's a more idiomatic way to do this lol code:
|
# ? Dec 25, 2017 06:24 |
|
I actually want to do some numeric stuff soon, and am seriously considering Fortran. Technically it was the first programming language I used, but uhh I was like 10 and mostly my dad was telling me what to type. What could possibly go wrong?
|
# ? Dec 25, 2017 06:45 |
|
If not J which I know passably well enough....
|
# ? Dec 25, 2017 06:47 |
|
Mao Zedong Thot posted:I actually want to do some numeric stuff soon, and am seriously considering Fortran. Technically it was the first programming language I used, but uhh I was like 10 and mostly my dad was telling me what to type. What could possibly go wrong? Lol working in Fortran (What specific numeric poo poo are ya doing)
|
# ? Dec 25, 2017 06:50 |
|
Luigi Thirty posted:so I got gambit installed since it was like the only lisp/scheme/whatever i could find that actually works on WSL (SBCL crashes due to something in /proc missing, the rest just crash on startup) and i've got most of a wavefront obj -> binary data script together i dont know much about what you are doing, but i know that it is bad!!!
|
# ? Dec 25, 2017 06:53 |
|
DELETE CASCADE posted:stop doing heavy numerical computational work in python. just stop. it's wrong, you're wrong, everything you're doing is wrong WRONG WRONG i will not desist, in fact i will continue to do it
|
# ? Dec 25, 2017 06:56 |
|
Mao Zedong Thot posted:I actually want to do some numeric stuff soon, and am seriously considering Fortran. Technically it was the first programming language I used, but uhh I was like 10 and mostly my dad was telling me what to type. What could possibly go wrong? it kinda depends what you mean by numeric stuff. if you're dealing with something where the hardware costs become a serious issue, then sure maybe doing some fortran and c sprinkled with some delicious assembler might make sense and save you some time, and thus cash, in the end. but even then most numerical stuff reduces to well known algorithms, and optimized versions already exist for most of those. you'll probably also want to make sure whatever you use has gpu support, because it's a no-brainer for a lot of numerical code these days
|
# ? Dec 25, 2017 06:59 |
|
Just hobby loving around with finance. Zero need for fast processing or efficiency. Just looking for a 'fun' change of pace, otherwise I'd do it in go Anybody using modern Fortran? Is it a horrorshow?
|
# ? Dec 25, 2017 07:36 |
DELETE CASCADE posted:stop doing heavy numerical computational work in python. just stop. it's wrong, you're wrong, everything you're doing is wrong WRONG WRONG DELETE CASCADE posted:blah blah blah "it's a C library so really the actual work isn't done in python!!!" i don't loving care, if you are writing a complicated enough program that you need these libraries, then you should get as far away from python and other garbage p-langs as you can, as fast as you can. it's only going to get worse, not better u mad dog Valeyard posted:i know people that use pandas just because it has a to_json() function lol i use it in our homegrown "database flusher" because it was fastest to shitcode a table update mindful of autoincrementing p-keys
|
|
# ? Dec 25, 2017 08:35 |
Mao Zedong Thot posted:Just hobby loving around with finance. Zero need for fast processing or efficiency. Just looking for a 'fun' change of pace, otherwise I'd do it in go i wouldnt be surprised if less people did modern fortran than modern cobol. there's very little reason to use it.
|
|
# ? Dec 25, 2017 08:36 |
|
fortran is basically used in either legacy code or on supercomputers doing physics simulations
|
# ? Dec 25, 2017 08:54 |
|
Luigi Thirty posted:so I got gambit installed since it was like the only lisp/scheme/whatever i could find that actually works on WSL (SBCL crashes due to something in /proc missing, the rest just crash on startup) and i've got most of a wavefront obj -> binary data script together lisp is so loving unreadable
|
# ? Dec 25, 2017 08:55 |
|
Luigi Thirty posted:so I got gambit installed since it was like the only lisp/scheme/whatever i could find that actually works on WSL (SBCL crashes due to something in /proc missing, the rest just crash on startup) and i've got most of a wavefront obj -> binary data script together you don’t need the outermost lambda, you can define a function directly—there, one level of indentation saved, ship it honestly though it looks pretty idiomatic to me, in that it represents a bunch of stream processing; maybe pull the (lambda (arg) up to the same line as the (map to maybe match more common coding style the Lisp (though not necessarily Scheme) way might be to define a few macros that take a stream and a body to apply to it, which applies it repeatedly to make a new stream; it reduces to the same thing but makes the code look more like this: code:
the advantage of the Scheme you wrote over this is that there’s no implicit accumulators so you could make it work on a lazy (pull) stream as easily as a complete list
|
# ? Dec 25, 2017 09:10 |
|
brap posted:lisp is so loving unreadable nah, though formatting can make a difference code:
|
# ? Dec 25, 2017 09:17 |
|
a real Lisper would do it all via the loop macro which is its own entire semi-infix procedural sublanguage full of mystery and wonder, that gets its own entire chapter in CLtL
|
# ? Dec 25, 2017 09:25 |
|
the loop macro is an abomination upon the lord prefer chaining higher order functions together. lisp's syntax makes this harder to read than it should be though because you have to read inside out
|
# ? Dec 25, 2017 09:28 |
|
well here's what I came up anyway with as an occasional lisper and a first-time schemer. it does the needful, converting the mesh from an .obj file from my modeler into an include file for my 68k assembler. it has some warts but it works! https://pastebin.com/MQSr8DRk
|
# ? Dec 25, 2017 09:45 |
|
I wrote a decent amount of racket code when I was first learning programming and wouldn’t discourage others from doing the same. sooner or later, though, you realize that a language where the first thing that happens in the program is always at the center of an onion of parentheses is just painful.
|
# ? Dec 25, 2017 10:19 |
|
Mao Zedong Thot posted:I was like 10 and mostly my dad was telling me what to type. Pair programming
|
# ? Dec 25, 2017 11:54 |
|
Mao Zedong Thot posted:Just hobby loving around with finance. Zero need for fast processing or efficiency. Just looking for a 'fun' change of pace, otherwise I'd do it in go On the subject of lisps, try julia
|
# ? Dec 25, 2017 12:52 |
|
does that lisp not have a threading macro like Clojure and if not, why?
|
# ? Dec 25, 2017 15:25 |
|
julia (https://julialang.org/) is a lisp?
|
# ? Dec 25, 2017 17:18 |
|
AWWNAW posted:does that lisp not have a threading macro like Clojure and if not, why? You're talking about Clojure's -> and ->>? should be pretty easy* to write, if the answer is no. *not much harder than writing any appropriately hygienic macro for the lisp in question
|
# ? Dec 25, 2017 18:24 |
|
DELETE CASCADE posted:blah blah blah "it's a C library so really the actual work isn't done in python!!!" i don't loving care, if you are writing a complicated enough program that you need these libraries, then you should get as far away from python and other garbage p-langs as you can, as fast as you can. it's only going to get worse, not better eat poo poo and die, fucker numerical code is non-programmer code and dear god i would rather deal with non-programmer python than non-programmer literally any other language if you really gotta make it sanic then that's what cython is for
|
# ? Dec 25, 2017 18:34 |
|
i write c and will go blue in the face defending it but ugh writing huge data shoveling in it, just use pythong especially for quick prototypes? experiments? nothing like ipython notebooks
|
# ? Dec 25, 2017 19:26 |
|
Shinku ABOOKEN posted:julia (https://julialang.org/) is a lisp? It supports macros and you are expected to use the macros lol
|
# ? Dec 25, 2017 19:27 |
|
bob dobbs is dead posted:On the subject of lisps, try julia gross, it has syntax
|
# ? Dec 25, 2017 22:52 |
|
DELETE CASCADE posted:blah blah blah "it's a C library so really the actual work isn't done in python!!!" i don't loving care, if you are writing a complicated enough program that you need these libraries, then you should get as far away from python and other garbage p-langs as you can, as fast as you can. it's only going to get worse, not better
|
# ? Dec 26, 2017 00:35 |
|
lol
|
# ? Dec 26, 2017 05:54 |
|
Mao Zedong Thot posted:I actually want to do some numeric stuff soon, and am seriously considering Fortran. Technically it was the first programming language I used, but uhh I was like 10 and mostly my dad was telling me what to type. What could possibly go wrong? I started my career as a professional terrible programmer with fortran and to be honest it's... fine. I mean Fortran 90, not Fortran 77. dealing with f77 legacy code was a nightmare lol. it's a way more straightforward language than C, if you have a hardon for it I say go for it. just make sure you understand floating point numbers well and the way fortran does arrays. it's been a while since I touched fortran seriously though, because every numeric problem that I stumble upon I just prototype a half-assed solution in python/numpy and it's always been good enough.
|
# ? Dec 26, 2017 07:25 |
|
|
# ? Jun 1, 2024 22:55 |
|
Why load css in .css format? self.javascript Submitted 5 hours ago by vmajsuk Hey folks, I've got some idea and I can't see why it is stupid, so you help me please Basically, why not load css in .js file? Storing css in strings with format like protobuf, maybe just simply use 'p' instead of 'padding', for example, and loading it together with a simple function that restores css based on these strings and appends it to the document.head seems like a huge win for the overall bundle size.
|
# ? Dec 27, 2017 00:52 |