Tony Albrecht
David Amador
Worthless Bums@djlink the first time as a junior dev you accidentally do a f == f comparison
Will PalmerGetting burned with floating points early, having a strict policy of using decimal arithmetic only for all decimal numbers for decades, eventually forgetting why, become convinced that you'll never encounter it in the wild using the types of values you're actually using...
Then you assert(m >= ceil(n * 1.1)) and it all comes flooding back.
0xDE@djlink I'm stealing this for my course lecture notes
Ehproque has moved
LambdaDuck
Cegorach@djlink …unless you're writing #JavaScript or #TypeScript
But hey! then you're fucked from the start!
Ángela Stella Matutina
ari@angelastella 💯
iba a decir que usarlo para representar género sería aceptable, pero creo que para eso prefiero una matriz de n-dimensiones /hj
also le pifiaste a la cuenta, esa la vuelo hoy (y si, es 100% culpa mia por estar haciendo cagadas con subdominios ajajaj)
iba a decir que usarlo para representar género sería aceptable, pero creo que para eso prefiero una matriz de n-dimensiones /hj
also le pifiaste a la cuenta, esa la vuelo hoy (y si, es 100% culpa mia por estar haciendo cagadas con subdominios ajajaj)
Las matrices n-dimensionales sirven para todo, hoy en día la memoria es barata (?)
Mencioné la cuenta que usaste para el boost, jajaja.
@angelastella wait what, pero en la otra cuenta no me aparece que haya boosteado ese post
oh god oh fuck 
oh god oh fuck
Con calma. Si precisás que mire algo desde "afuera" (desde mi instancia, o sea), acá estoy.
Harry (Hyeonggon) Yoo
Chris Gioran 💔
Ivy [_gay] Mae (moved, see bio)@djlink errm I would actually trust a < or >
@djlink obviously not an equality but if I needed to for one I would check for it within a range
if equalities are important I would probably just use a fixed point
if equalities are important I would probably just use a fixed point
Scott Michaud@ivy @djlink Also depends on the context. If you're dealing with things that must be a power-of-two fraction (ex: 0.5, 0.25, 0.125) then it might make sense to compare... although might be time to question why floating points are being used.
But, yes, most scenarios is less than a delta from the expected value.
Happosai 八宝斎@djlink I trust a>b because the floating point finite field for all IEEE formats is well ordered.
I also trust a==b when performing non-transformative confirmation of pedigree.
</spoiling-the-joke>
I also trust a==b when performing non-transformative confirmation of pedigree.
</spoiling-the-joke>
maybeanerd@djlink Cries in JavaScript.
Everything is floating point here🫠
don@djlink Nice!!!
moved to => @
El Gordo@djlink Tell that to 99% of Excel users.
Izzy@djlink I like to live dangerously
Joel Falcou, Ph. D@djlink you wouldnt download a NaN !
Jonathan ‘theJPster’ Pallant
Josh TriplettIf not for IEEE 754 floating-point compliance, I'd argue we should have just made NaN == NaN and had that be the end of it. (Or, alternatively, have NaN produce an exception like 1/0 by default.) That's a more understandable, less confusing, and in my opinion less error-prone system.
(Whether and how to *sort* NaN so that floats can be Ord is a harder and more debatable question.)
(Whether and how to *sort* NaN so that floats can be Ord is a harder and more debatable question.)
Josh Simmons@djlink Don't Fear The Float!
Samuel
Raven Luni@djlink Reminds me of when I made that mistake in my very first ray tracer (that lump of noise in the middle was supposed to be refraction) :P