terrible programmers: my god, it's full of chars

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > terrible programmers: my god, it's full of chars

«‹›1446 »

bobbilljim: May 29, 2013; this christmas feels like the very first christmas to me

lol if you cant program by flipping a series of switches on a computer from memory, then you just cant program at all

# ? Nov 19, 2014 23:24

Adbot: ADBOT LOVES YOU

# ? May 25, 2024 23:46

orphean: Apr 27, 2007; beep boop bitches
my monads are fully functional

bobbilljim posted:

lol if you cant program by flipping a series of switches on a computer from memory, then you just cant program at all

If you aren't toggling the os into the supercomputer you built by hand from memory you ain't poo poo.

We are all scrubs in comparison

# ? Nov 19, 2014 23:37

qntm: Jun 17, 2009

CPColin posted:

In interviews I've been in on, I've always been prepared to count an answer as correct if the person says, "I don't know, but I'd look here to find out."

Nobody ever does.

this is a relief because annoyingly the only sorting algorithms I can remember are the joke ones (bogosort, quantum bogosort, sleepsort, fonzsort)

# ? Nov 20, 2014 00:05

Captain Foo: May 11, 2004; we vibin'
we slidin'
we breathin'
we dyin'

qntm posted:

this is a relief because annoyingly the only sorting algorithms I can remember are the joke ones (bogosort, quantum bogosort, sleepsort, fonzsort)

remind me what quantum bogosort is

# ? Nov 20, 2014 00:10

pseudorandom name: May 6, 2007

an O(n) implementation of bogosort on a quantum computer

# ? Nov 20, 2014 00:15

Valeyard: Mar 30, 2012; Grimey Drawer

I must be missing something really obvious here, it's making GBS threads the bed because its expecting an Expr. but the function signature says to except a VarTable, which is defined as H.map String Expr - which is apparetnyl what the function is trying to return, if I am reading that right

code:

type VarTable = H.Map String [s]Expr[/s]

parse_par_decls :: [String] -> VarTable    
parse_par_decls par_lines = 
    do
        let k = [run_parser f95_par_decl_parser j | j <- par_lines]
        let kplus = [ pd_parval g | g <- k]
        let knames = [show (pd_parname t) | t <- k]
        let zipk = zip knames kplus
        let ma = H.fromList zipk
        return ma :: VarTable

which results in:

code:

Main.hs:63:16:
    Couldn't match expected type `Expr'
                with actual type `H.Map String Expr'
    In the first argument of `return', namely `ma'
    In a stmt of a 'do' block: return ma :: VarTable

edit:

ok I am 99% sure that I am actually somehow ending up trying to return

code:

H.Map String (H.Map String Expr)

which is why it is expecting an Expr

Valeyard fucked around with this message at 00:33 on Nov 20, 2014

# ? Nov 20, 2014 00:26

Captain Foo: May 11, 2004; we vibin'
we slidin'
we breathin'
we dyin'

pseudorandom name posted:

an O(n) implementation of bogosort on a quantum computer

hahahaha

# ? Nov 20, 2014 00:27

gonadic io: Feb 16, 2011; >>=

Valeyard posted:

this is a classic mistake and a good reason why 'return' should be called something different. to directly answer your question, you shouldn't use do-notation here, instead structure it like (you could also use "= let ... in ..." if you prefer)

code:

parse_par_decls par_lines = ma :: VarTable
    where
    k = ...
    kplus = ...
    ...

as for why, well do-notation is only to be used for monads and in it you don't deal with the objects themselves but the values inside of them*. for your code now, you're actually constructing the container yourself so you shouldn't use do-notation (or return). for now, probably the only monads that you'll use do-notation for will be IO and possibly list, maybe, state and stuff.

*: the 'a' inside of 'Parser a' for example, you never deconstruct Parsers but can only pass them around. you're free to gently caress around with the values of type 'a' as much as you want though.

gonadic io fucked around with this message at 00:41 on Nov 20, 2014

# ? Nov 20, 2014 00:34

gonadic io: Feb 16, 2011; >>=

as for why you get that message, which Map are you using? it looks like it might have a monad instance, so the compiler is assuming that you're using do-notation correctly and inferring the type from that

# ? Nov 20, 2014 00:39

Moist von Lipwig: Oct 28, 2006; by FactsAreUseless; Tortured By Flan

qntm posted:

this is a relief because annoyingly the only sorting algorithms I can remember are the joke ones (bogosort, quantum bogosort, sleepsort, fonzsort)

Hahaha what the hell is fonzsort? Does it involve jukebox punching?

# ? Nov 20, 2014 00:42

Valeyard: Mar 30, 2012; Grimey Drawer

AlsoD posted:

this is a classic mistake and a good reason why 'return' should be called something different. to directly answer your question, you shouldn't use do-notation here, instead structure it like (you could also use "_lines = let ... in ma :: VarTable" if you prefer)
code:
parse_par_decls par_lines = ma :: VarTable
    where
    k = ...
    kplus = ...
as for why, well do-notation is only to be used for monads and you generally don't deal with the objects themselves but the values inside of them*. you're actually constructing the container yourself so you shouldn't use do-notation (or return).

*: the 'a' inside of 'Parser a' for example, you never deconstruct Parsers but can only pass them around.

arghh, I was tricked! the dummy skeleton code for this function had a dummy return statement, so I just assumed it would work that way without thinking about it

it is a Data.Map, which also came predefined.

if i had used the do notation correctly, would it have involved making the types explicit when doing "let"s?

# ? Nov 20, 2014 00:43

my homie dhall: Dec 9, 2010; honey, oh please, it's just a machine

imo those list comprehensions are weird dude.

# ? Nov 20, 2014 00:47

bobbilljim: May 29, 2013; this christmas feels like the very first christmas to me

VLADIMIR GLUTEN posted:

Hahaha what the hell is fonzsort? Does it involve jukebox punching?

tongue punching in teh jukebox

# ? Nov 20, 2014 00:48

gonadic io: Feb 16, 2011; >>=

i couldn't help myself, i did a little refactoring. you can replace zipk with a list comprehension and then inline kplus and knames to give:

code:

parse_par_decls par_lines = H.fromList [(show (pd_parname t), pd_parval t) | t <- k]
    where k = map (run_parser f95_par_decl_parser) par_lines

i mean you could easily inline k too but that line's already starting to look a little long

# ? Nov 20, 2014 00:59

gonadic io: Feb 16, 2011; >>=

Valeyard posted:

if i had used the do notation correctly, would it have involved making the types explicit when doing "let"s?

replace "shouldn't" with "can't" in my previous statement. there is no correct way to use do-notation in this case. do-notation is sugar that makes it easier to work with monads, you're not working inside a monad here so you can't use do-notation. (but no, when using let statements inside do-notation you aren't forced to state all your types explicitly except in edge cases)

gonadic io fucked around with this message at 01:14 on Nov 20, 2014

# ? Nov 20, 2014 01:00

Valeyard: Mar 30, 2012; Grimey Drawer

AlsoD posted:

there is no correct way to use do-notation in this case. do-notation is sugar that makes it easier to work with monads, you're not working inside a monad here so you can't use do-notation. (but no, when using let statements inside do-notation you aren't forced to state all your types explicitly except in edge cases)

i went back and looked at the monad section of learnyouahaskell and i see what you are saying there about monads (although most of that chapter is still daunting)

AlsoD posted:

i couldn't help myself, i did a little refactoring. you can replace zipk with a list comprehension and then inline kplus and knames to give:
code:
parse_par_decls par_lines = H.fromList [(show (pd_parname t), pd_parval t) | t <- k]
    where k = map (run_parser f95_par_decl_parser) par_lines
i mean you could easily inline k too but that line's already starting to look a little long

the question about this is; is it any more efficient? space wise, sure

# ? Nov 20, 2014 01:14

gonadic io: Feb 16, 2011; >>=

Valeyard posted:

the question about this is; is it any more efficient? space wise, sure

probably not any faster (or less space) depending on how aggressively the compiler inlines/fuses. anybody else want to weigh in?

# ? Nov 20, 2014 01:15

gonadic io: Feb 16, 2011; >>=

actually is there a fusion rule for 'zip'? it probably would end up faster then. either way, it won't be an appreciable difference and the refactoring would be for the person reading the code, whether it's future you or a marker

# ? Nov 20, 2014 01:20

Stringent: Dec 22, 2004; image text goes here

Seaside Loafer posted:

Good examples:
1) I worked on an military aircraft simulator project and the brains of the simulator were in an ancient 6 foot tall computer less powefull that your phone probably that parts for were worth a fortune. Its code base was in FORTRAN. Now at first you might think 'fucks sake just convert it to something modern' but the problem is the amount of certification required to sign off a new codebase as functional for a goverment project, it requires massive amounts of testing and checking and certificates, especially with anything airplane related etc etc etc. It was a valid cost benifit analysis to keep this poo poo. So I know for a fact that a certain RAF base in the UK is running its simulator on 30 year old FORTRAN code on a computer that may well be older than some of you. Its not just a case of copying the code to a new language.

I understand needing to keep the software, but couldn't the hardware be virtualized?

# ? Nov 20, 2014 01:20

DimpledChad: May 14, 2002; Rigging elections since '87.

hey guys lets talk about monads

# ? Nov 20, 2014 01:29

Bloody: Mar 3, 2013

DimpledChad posted:

hey guys lets talk about gonads

# ? Nov 20, 2014 01:32

DimpledChad: May 14, 2002; Rigging elections since '87.

that's a monads implementation in golang right

# ? Nov 20, 2014 01:33

Valeyard: Mar 30, 2012; Grimey Drawer

DimpledChad posted:

hey guys lets talk about monads

"please work..."

# ? Nov 20, 2014 01:35

gonadic io: Feb 16, 2011; >>=

as far as anybody who is programming haskell need be concerned, monad means 'can use do-notation on this thing'

# ? Nov 20, 2014 02:04

DimpledChad: May 14, 2002; Rigging elections since '87.

b-b-but higher-kinded types functor transformation monad laws type theory burritos!

# ? Nov 20, 2014 02:07

Seaside Loafer: Feb 7, 2012; Waiting for a train, I needed a shit. You won't bee-lieve what happened next

Stringent posted:

I understand needing to keep the software, but couldn't the hardware be virtualized?

That would be a good idea in theory but since nobody had already done it for the thing (wish I could remember what it was) the company couldnt really embark on a project to emulate it could they, that would have been more difficult than the actual project. If I get a brainwave and remember what model it was i'll post it. it was (is for another 5 years at least according to the contract) a 'mini-computer' haha, of some sort http://en.wikipedia.org/wiki/Minicomputer

Seaside Loafer fucked around with this message at 03:16 on Nov 20, 2014

# ? Nov 20, 2014 03:02

Seaside Loafer: Feb 7, 2012; Waiting for a train, I needed a shit. You won't bee-lieve what happened next

e: quote is not edit

# ? Nov 20, 2014 03:15

kitten emergency: Jan 13, 2008; get meow this wack-ass crystal prison

as an aside that job told me to redo the exercise and "go crazy" with it so I've made some horrific monstrosity in swing (which actually isn't that bad now that I've sit down with it for a few hours) that I'm getting happier with

ui programming is really a pain though.

# ? Nov 20, 2014 03:58

gonadic io: Feb 16, 2011; >>=

"a picture is worth 1000 words" means that it takes 1k loving works to create a picture

# ? Nov 20, 2014 04:02

Bloody: Mar 3, 2013

uncurable mlady posted:

as an aside that job told me to redo the exercise and "go crazy" with it so I've made some horrific monstrosity in swing (which actually isn't that bad now that I've sit down with it for a few hours) that I'm getting happier with

ui programming is really a pain though.

That is because you are not using c# and wpf

# ? Nov 20, 2014 04:05

Valeyard: Mar 30, 2012; Grimey Drawer

Bloody posted:

That is because you are not using c# and wpf

or more specifically, its because he is using swing

# ? Nov 20, 2014 04:09

Moist von Lipwig: Oct 28, 2006; by FactsAreUseless; Tortured By Flan

bobbilljim posted:

tongue punching in teh jukebox

text me

# ? Nov 20, 2014 04:09

DONT THREAD ON ME: Oct 1, 2002; by Nyc_Tattoo; Floss Finder

Seaside Loafer posted:

Pick something you know a shitload about so you dont have to do any systems analysis and just try and write a program to do it. Every time I need to learn a new language I try to write this really complicated board game I played when I was a little kid, I never finish it but by the time im a quarter of the way through ive learnt enough of the language to be ok at it.

i think he's referring to the difficulty of jumping from one ecosystem to another like, professionally. learning the new ecosystem isn't the issue, it's doing something worthwhile that you can point to and say 'yes i am a professional user of this thing'

# ? Nov 20, 2014 04:12

MeruFM: Jul 27, 2010

there's a heroku issue where a dyno could down and there's no notification so all requests going to that dyno just get stuck until the heroku timeout.

apparently this doesn't happen if you pay for the uber-dyno that costs like 10x as much. wtf

# ? Nov 20, 2014 04:23

DONT THREAD ON ME: Oct 1, 2002; by Nyc_Tattoo; Floss Finder

MeruFM posted:

there's a heroku issue where a dyno could down and there's no notification so all requests going to that dyno just get stuck until the heroku timeout.

apparently this doesn't happen if you pay for the uber-dyno that costs like 10x as much. wtf

didnt that happen like 2 years ago

# ? Nov 20, 2014 04:37

Valeyard: Mar 30, 2012; Grimey Drawer

here is a more general haskell question, for which i am assuming the answer is no based on what ive read, but is there any kind of equivalent functionality of specifying a specific field in a datatype to be what is returned by show? I am trying to test something involving a really nasty data structure, and its reallllly hard to read whats happening, so i just want to get rid of most of the junk its showing me

attempting to install one of the pretty print libraries just now

Valeyard fucked around with this message at 05:08 on Nov 20, 2014

# ? Nov 20, 2014 04:54

kitten emergency: Jan 13, 2008; get meow this wack-ass crystal prison

Bloody posted:

That is because you are not using c# and wpf

im not familiar with c# as well as I am with Java, and the only PC I have access to is really bad and doesn't run VS so hot. Java it is~

# ? Nov 20, 2014 05:07

Bloody: Mar 3, 2013

uncurable mlady posted:

im not familiar with c# as well as I am with Java, and the only PC I have access to is really bad and doesn't run VS so hot. Java it is~

You've chosen poorly

# ? Nov 20, 2014 05:10

kitten emergency: Jan 13, 2008; get meow this wack-ass crystal prison

Bloody posted:

You've chosen poorly

Mac OS X is the worlds best operation system

# ? Nov 20, 2014 05:15

Adbot: ADBOT LOVES YOU

# ? May 25, 2024 23:46

my homie dhall: Dec 9, 2010; honey, oh please, it's just a machine

Valeyard posted:

here is a more general haskell question, for which i am assuming the answer is no based on what ive read, but is there any kind of equivalent functionality of specifying a specific field in a datatype to be what is returned by show? I am trying to test something involving a really nasty data structure, and its reallllly hard to read whats happening, so i just want to get rid of most of the junk its showing me

attempting to install one of the pretty print libraries just now

are you defining the datatype? you can use record syntax to give your fields names.

# ? Nov 20, 2014 05:18

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > terrible programmers: my god, it's full of chars

«‹›1446 »