|
rip thread
|
# ? Dec 7, 2014 17:06 |
|
|
# ? Jun 2, 2024 18:23 |
|
i worked around a re-entrancy bug in the version of fcvtbuf that shipped with arm gcc 4.4.1 last week by rewriting a domain-limited version of it since i didn't have the source handy. saved 7.5k. didn't wind up using the white/steel algorithm since pulling in an arbitrary precision arithmetic library was a no-go. validation was a bitch
|
# ? Dec 7, 2014 17:12 |
|
also, has anybody seen BONUS lately?
|
# ? Dec 7, 2014 17:19 |
|
tef posted:rip thread I often wish they would, but for some algorithms it's worth putting up with all the problems of shared memory concurrency
|
# ? Dec 7, 2014 17:28 |
|
dpweb 1 minute ago | link Couldn't you charge people to look at their pull requests? The contributor's fix may seem very important to them, but in actuality may or may not be of any value at all. They are using free software but asking the maintainers of it to spend their time to review their pull request. Some entitlement there. The justification, my fix is important. Well, we don't know that, and to determine it costs time. I would think a fee system where people pay to have their pull requests prioritized. I know it goes against this misguided idea that everything free is a virtue, but there is a good reason why some things have an associated cost involved. reply
|
# ? Dec 7, 2014 17:30 |
|
triple sulk posted:dpweb 1 minute ago | link lol
|
# ? Dec 7, 2014 17:57 |
|
charge people a fee to consider their ideas, but please consider mine for free
|
# ? Dec 7, 2014 20:51 |
|
ugh, thanks PL thread for getting monads stuck in my head now i just accidentally built a loving monad in Java. i think. i'm doing some layer-izing of Java code. i have a factory type T that constructs implementations of various interfaces A, B, C so then i made a factory class, say, T' that still constructs implementations of A, B, C using the instance of T it's given and then wraps each of them in adapter classes A' B' C' which also implement the A, B, and C interfaces by calling down to their wrapped A, B, C-implementing objects but then decorating them in various ways. it got to the point where i was factorising each of those wrapping operations into an instance of a helper class that took a "construct" lambda and a "wrap" lambda and i was like "..............wait a minute. PL THREAD! " what i really learned was that i'm overdesigning the gently caress out of this and it's going to be completely incomprehensible and i should just bake all this poo poo into the implementations of A, B, C to begin with.
|
# ? Dec 8, 2014 00:11 |
|
Mr Dog posted:ugh, thanks PL thread for getting monads stuck in my head now i just accidentally built a loving monad in Java. i think. A monad is a generic type with 3 operations: map, join and return does your type implement these?
|
# ? Dec 8, 2014 00:14 |
|
Malcolm XML posted:A monad is a generic type with 3 operations: merely wrapping and unwrapping does not a monad make
|
# ? Dec 8, 2014 00:15 |
|
akadajet posted:lmao I see a Gandalf, Legolas, and a Gimli, but which one's the Aragorn?
|
# ? Dec 8, 2014 00:17 |
|
sarehu posted:I see a Gandalf, Legolas, and a Gimli, but which one's the Aragorn? red sweater
|
# ? Dec 8, 2014 00:22 |
|
Malcolm XML posted:
Wouldn't it be bind and return? Join plays a bigger part in free monads but isn't super necessary and map I assume you're talking about a morphism as a functor?
|
# ? Dec 8, 2014 01:35 |
|
UncleBlazer posted:Wouldn't it be bind and return? Join plays a bigger part in free monads but isn't super necessary and map I assume you're talking about a morphism as a functor? bind f m = join (map f m) join = bind id most math stuff i've seen uses join + map but in actual programming bind is used way more than join and ya, the map here is the functor one
|
# ? Dec 8, 2014 01:52 |
|
AlsoD posted:bind f m = join (map f m) can you define return in terms of join+map? i think you need return and bind or join+map.
|
# ? Dec 8, 2014 02:20 |
|
could you not
|
# ? Dec 8, 2014 03:07 |
|
hackbunny posted:could you not no they will not
|
# ? Dec 8, 2014 03:08 |
|
what is the point of join at all if it's just bind id? e: this probably belongs in the safezone my homie dhall fucked around with this message at 03:38 on Dec 8, 2014 |
# ? Dec 8, 2014 03:34 |
|
theadder posted:no they will not
|
# ? Dec 8, 2014 04:16 |
|
Ploft-shell crab posted:what is the point of join at all if it's just bind id? because "you can define monads in terms of map and bind id" is kind of a silly characterization
|
# ? Dec 8, 2014 04:45 |
|
akadajet posted:Nah, there are other ways to achieve "the look" i got 4/10
|
# ? Dec 9, 2014 11:16 |
|
Mr Dog posted:ugh, thanks PL thread for getting monads stuck in my head now i just accidentally built a loving monad in Java. i think. gently caress you Malcolm XML posted:A monad is a generic type with 3 operations: gently caress you UncleBlazer posted:Wouldn't it be bind and return? Join plays a bigger part in free monads but isn't super necessary and map I assume you're talking about a morphism as a functor? gently caress you AlsoD posted:bind f m = join (map f m) gently caress you FamDav posted:can you define return in terms of join+map? i think you need return and bind or join+map. gently caress you Ploft-shell crab posted:what is the point of join at all if it's just bind id? gently caress you GrumpyDoctor posted:because "you can define monads in terms of map and bind id" is kind of a silly characterization gently caress you
|
# ? Dec 9, 2014 12:28 |
|
Wheany posted:gently caress you gonad
|
# ? Dec 9, 2014 13:13 |
|
Wheany posted:gently caress you
|
# ? Dec 9, 2014 13:27 |
|
Soricidus posted:cjk unification is fine. the only objection to it is "unicode doesn't distinguish between the 9-stroke 草 traditionally used in japan and the 10-stroke 草 traditionally used in china", which sometimes seems valid to westerners until they realise that it's equivalent to complaining that unicode doesn't distinguish between open and closed variants of the glyph 4. specify a font if the precise shape of the glyph matters, that's not what character encodings are for. Stroke count is important https://en.wikipedia.org/wiki/Collation#Radical-and-stroke_sorting And yeah, unicode's handling of alphabets is wildly inconsistent. Take a look at what they consider to be letters in Hungary some time.
|
# ? Dec 9, 2014 15:11 |
|
Wheany posted:gently caress you
|
# ? Dec 9, 2014 15:29 |
|
monads are not particularly interesting theoretical objects. the only reason anyone cares about them is because of haskell and its lazy evaluation, and the only monad that haskellers really care about is the IO monad. when they observe that lots of other things "are monads" too, they're doing that to make themselves feel better about needing the IO monad. it's a hack, and haskellers need to feel elegant like pretty princesses
|
# ? Dec 9, 2014 20:26 |
|
akadajet posted:Nah, there are other ways to achieve "the look" arguably a lot of today's (relatively) successful programmers might have been social deviants 40+ years ago IT work contains some pretty high-paying high-prestige jobs for some people who would otherwise be sweeping floors, avoiding eye contact, and torturing animals in basements
|
# ? Dec 9, 2014 20:28 |
|
JewKiller 3000 posted:monads are not particularly interesting theoretical objects. the only reason anyone cares about them is because of haskell and its lazy evaluation, and the only monad that haskellers really care about is the IO monad. when they observe that lots of other things "are monads" too, they're doing that to make themselves feel better about needing the IO monad. it's a hack, and haskellers need to feel elegant like pretty princesses yes to the reason people care is haskell part, no to the rest
|
# ? Dec 9, 2014 21:48 |
|
fart simpson posted:i'm a haskell programmer
|
# ? Dec 9, 2014 21:52 |
|
what about computation expressions in f# they look cool
|
# ? Dec 9, 2014 21:54 |
|
pointsofdata posted:what about computation expressions in f# they look cool they are cool, but they're just a syntax extension on top of a standard interface style, which can be and is used in other similar languages with higher-order functions and parametric polymorphism. you could rewrite the standard library of ocaml or standard ML to use monadic style. we don't do that because it sucks. haskell has to do it.
|
# ? Dec 9, 2014 22:03 |
|
I'm doing a little Python demo today, any suggestions for neat things to show ruby and Java devs current flow is gonna be an overall about the lang, nice stuff about expressions, context managers and decorators then an ipython notebook and flask demo
|
# ? Dec 9, 2014 22:18 |
|
ahmeni posted:I'm doing a little Python demo today, any suggestions for neat things to show ruby and Java devs You reminded me about context managers and I immediately used it in my project, ha ha
|
# ? Dec 10, 2014 09:48 |
|
JewKiller 3000 posted:monads are not particularly interesting theoretical objects. the only reason anyone cares about them is because of haskell and its lazy evaluation, and the only monad that haskellers really care about is the IO monad. when they observe that lots of other things "are monads" too, they're doing that to make themselves feel better about needing the IO monad. it's a hack, and haskellers need to feel elegant like pretty princesses what is monad? can any1 explain it 2 me
|
# ? Dec 10, 2014 15:20 |
|
KARMA! posted:what is monad? can any1 explain it 2 me http://www.tmk.com/ftp/humor/macburrito.txt
|
# ? Dec 10, 2014 15:24 |
|
KARMA! posted:what is monad? nobody really knows.
|
# ? Dec 10, 2014 15:25 |
|
would someone please repost the link to that article / blog post that discusses how (most, all?) monad tutorials aren't helpful or useful? sorry i'm short on details or misremembering the specifics, that's why i want to reread that blurb
|
# ? Dec 10, 2014 19:23 |
|
monads are the anarcho-capitalism of functional programmers
|
# ? Dec 10, 2014 20:23 |
|
|
# ? Jun 2, 2024 18:23 |
|
bear with me here, anarcho-capitalists have this vision of an ideal society where things are mostly the way they are now except that there will be itemized fees on everything you take for granted, e.g. roads, law enforcement, food inspection, bankruptcy adjustment etc. and they write all this theory on how to reconstruct everything the way it already is but without any nominal government then you have monads which have all this theoretical literature on how they can be used to do perfectly ordinary tasks like doing one thing after another thing without formal side effects. so the same situation really
|
# ? Dec 10, 2014 20:43 |