|
As a toaster programmer who loves to deal with 16k of memory, I think that your script to read a weather database for a desktop widget is... As a server programmer who has 64GB of memory and 18 cores for a single app, I don't see why you couldn't use these 7 frameworks for different layers when writing a phone app because... and so on
|
#
¿
Oct 27, 2013 04:36
|
|
|
|
| # ¿ May 16, 2024 08:05 |
|
|
Way to get the point.
|
|
#
¿
Oct 27, 2013 05:31
|
|
|
A few years ago I've used php to write a minimalist website on the cheapest hosting possible for a friend's relative, then spent more time listening to people telling how bad of a decision it was to use PHP than it took me to write the website. Nobody has had to maintain it and it was in retrospect 100% the right decision. v  v
|
|
#
¿
Oct 27, 2013 16:29
|
|
|
I don't disagree with that. But it's a p-lang that did the job at what it was targeted for and likely came much cheaper for hosting than asp / jsp for a few hours' project for a friend. Aka using PHP was the right decision and context is king.
|
|
#
¿
Oct 27, 2013 16:38
|
|
|
What's your general feeling about Forth? Most criticism I've heard of it for very small projects was that it loving owns, but it may be very close to write-only.
|
|
#
¿
Oct 27, 2013 23:11
|
|
|
Forums Terrorist posted:would it be a bad idea to learn erlang for fucks 
|
|
#
¿
Nov 5, 2013 00:06
|
|
|
Welcome to the world of Rubber Ducking.
|
|
#
¿
Nov 5, 2013 19:01
|
|
|
Talking things out loud (or writing them down) -- and generally thinking about them in a way that forces you to boil down abstract thought to something explicit and tangible -- is often a good way to figure out where your mental abstract model is disconnected from the real world expectations you have about it. Also see subvocalization as an interesting related topic.
|
|
#
¿
Nov 5, 2013 20:55
|
|
|
The only DBA I remember dealing with shrugged it off when we found our DB password leaked on some Russian site and said "who cares it was a MySQL 4 password anyway, it is poo poo hashing" and then never changed the password. He lost his job a few weeks later.
|
|
#
¿
Nov 13, 2013 04:35
|
|
|
rsjr posted:i got a job offer doing erlang backend for a telephone company Learn Erlang dude. Come 'round here, I've got good stuff. Erlang is weird looking, but fun.
|
|
#
¿
Nov 14, 2013 00:18
|
|
|
gucci void main posted:monocqc you should do one of those video series things that are popular these days, but for erlang I wouldn't really know how to go about it, and hate my voice on recordings too much to edit it. Even when I do a talk at a conference or something, I'll tend to transcribe it in text form and redirect people there instead.
|
|
#
¿
Nov 14, 2013 01:00
|
|
|
Socracheese posted:im setting up an amazon ec2 linux vm just for fuckin around on erlang, seems like u can put it on osx but its clearly for a linux It works fine on OSX and you can grab pre-built packages at https://www.erlang-solutions.com/downloads or you can build it from source with Wx and use some graphical apps like observer or whatever (the packages should work with it too, and it works on all platforms in general).
|
|
#
¿
Nov 14, 2013 01:26
|
|
|
Socracheese posted:ok i have dived into this for a lil' bit and erlang is some voodoo poo poo. currently trying to get this thing listening on port 80 instead of 8080 If you get Recon you can start tracking your most active connections over time and see where all the precious bandwidth goes  (use recon:port_info/1 to get details about each connection then)
|
|
#
¿
Nov 14, 2013 16:58
|
|
|
Nomnom Cookie posted:i ran erl today you use '.' to end an expression and get it to run. code:
|
|
#
¿
Nov 15, 2013 04:56
|
|
|
Also I gotta try this thing soon:
|
|
#
¿
Nov 15, 2013 04:59
|
|
|
coffeetable posted:thing is, i have no damned clue whether this is a good design, or if the specific way i've implemented the pipeline (with each generator accepting a generator of the previous type as a constructor argument, so a dependency injection framework can pull it all together) is a good implementation. In my (relatively short) experience, prototyping is where you test ideas real quick and see what kind of hurdles you're gonna hit with your design. It's purposefully lovely code where you throw stuff at the wall to see what will stick and that will highlight flaws in how you thought things out. It's a rough draft you make to gently caress up and make your mistakes early. Then the big risk is having that prototype making it to production.
|
|
#
¿
Nov 23, 2013 04:49
|
|
|
I wanted to write a commit hook that would add in whitespace dicks at the end of files so that people who don't pay attention or diff without whitespace end up not seeing all the poo poo being added. Never found the time or motivation to actually do it though.
|
|
#
¿
Jan 8, 2014 04:17
|
|
|
Elixir's spaces for function application, combined with its eager evaluation, and its use of different namespaces for closures and literal functions can be confusing. I wouldn't consider it to have a neat syntax, or at the very least, not the best semantics around so far. So for example, the pipe operator (|>) allows to transform something like f(g(a,1),2,3) into a |> g(1) |> f(2,3), or with spaces: a |> g 1 |> f 2,3 . That's fine. Now if g is instead a closure, then it works a bit like lisps-2 and you need a different syntax: f(g.(a,1),2,3), and a |> g.(1) |> f 2,3 (closures do not allow the parens-less calling syntax). Similarly to standard lisp-2 shenanigans, you can have both g() and g.() calls happening and referring to two different functions within the same scope. The former refers to the literal 'g' function, while the latter refers to the closure within the 'g' variable. Here's where it gets fun with spaces for function application. Let's say I use this g combo and use it in an expression such as g |> something(x). It isn't clear whether I just called g() and passed its result to something(_, x), or if I passed in the variable itself. Then the |> macro has problems of its own, namely that it's impossible to call a function and use its result in a pipe sequence: g |> (f()) isn't valid as far as I know. Elixir has nice-looking syntax, but warts hiding around it, and you have to be careful about it. I remember finding a bug with its pattern matching assignment/reassignment where poo poo was left unspecified: x = 3 {x, ^x} = {5,5} It was undefined whether the pattern matching would take place left-to-right or not, and whether the re-assignment of 'x' would take place before its fixed matching on a previous value (^x). In one case the expression works, in the other, it fails and errors out. I think this has now been specified into the language, but you just know there's plenty of that stuff lurking in. I know I would instantly remove the parens-free calls given how confusing it is in Elixir. E: just asked José Valim. The variable is always prefered in g |> f, and nullary function calls will only happen if there is no scope ambiguity. MononcQc fucked around with this message at 15:57 on Jan 9, 2014 |
|
#
¿
Jan 9, 2014 15:37
|
|
|
I'm a terrible OO programmer, I end up using classes as modules/interfaces and end up writing in a nearly functional style 
|
|
#
¿
Jan 11, 2014 04:39
|
|
|
Valeyard posted:also said that "computer scientists are the gods of the future" and asked if anyone wanted to disagree with him, they did not everyone learned their first lesson: ignore the sperg with a toxic attitude
|
|
#
¿
Jan 14, 2014 15:24
|
|
|
Print statement debugging has been my bread and butter of debugging. Languages that allow tracing make that easier. When all else fails, I get a piece of paper and pencils and I think really hard.
|
|
#
¿
Jan 30, 2014 19:34
|
|
|
Posting Principle posted:is there a way to configure lager to email me on certain log levels, because that would be rad as hell and i would like it very much Probably would have to make a custom error handler and using something like gen_smtp to make it work. I don't know of any known library to do it directly otherwise off the top of my head.
|
|
#
¿
Jan 30, 2014 22:32
|
|
|
Awia posted:does anyone know a good book on algorithms/algorithmics or whatever its called I liked The Algorithm Design Manual by Steven Skiena (PDF). It reads easily and will be a good intro to basic algorithms in many categories, and can be used as a stepping stone if you need to dig in deeper later on.
|
|
#
¿
Feb 7, 2014 13:50
|
|
|
whenever it's big enough to need its own file
|
|
#
¿
Feb 12, 2014 17:30
|
|
|
Dark on light is also the best for presentations and working outdoors. Solarized light looks decent for this too.
|
|
#
¿
Mar 25, 2014 20:11
|
|
|
Also modal editing is a good thing. And vim commands are multiple but are structured in a logical way to be composable together.
|
|
#
¿
Mar 26, 2014 00:08
|
|
|
Shaggar posted:another rule of thumb is to avoid anything popular amongst academics or idiot college kids because neither group understands anything.  http://redmonk.com/sogrady/2013/04/04/academia-and-programming-languages/
|
|
#
¿
Apr 14, 2014 14:39
|
|
|
dang them professors always teaching in languages they hate
|
|
#
¿
Apr 14, 2014 14:45
|
|
|
it's computer science vs. software engineering debate time!
|
|
#
¿
Apr 14, 2014 17:18
|
|
|
MeruFM posted:can't newer routers just slowly seed in alternative transport layers with fallback to tcp/udp maybe once IPv6 is there
|
|
#
¿
Apr 30, 2014 00:52
|
|
|
Valeyard posted:Erlang eh?
|
|
#
¿
May 2, 2014 00:48
|
|
|
Currying also reduces the need for parens. In strict languages I tend to still prefer them to properly delimit everything, and argument-free parens-free calls are the worst thing.
|
|
#
¿
May 9, 2014 16:04
|
|
|
MeramJert posted:are those like properties in python? They're like properties in python (given you can do the __get and __set dance) except imagine that all methods that take no argument (other than self) in python are now properties also. object.prop == object.prop() in terms of what it does or gives. Also functions so that if you do 'x = f', it isn't obvious if 'f' represents the variable 'f' or the function call 'f' (equivalent to 'f()').
|
|
#
¿
May 9, 2014 16:48
|
|
|
I don't think functional is intrinsically superior, but it's a model that meshes a bit better with how I think than OO does, which just never really meshed well with me.
|
|
#
¿
May 9, 2014 21:41
|
|
|
raruler posted:did they replace erlang with php for facebook chat or are people just lying to me? They wrote the first version in Erlang then went "wait we got tons of C++ people and like 2-3 Erlang devs" so they rewrote it in C++. Then they bought Whatsapp and now they have 30something Erlang guys.
|
|
#
¿
May 9, 2014 21:45
|
|
|
Bloody posted:sounds bad It's not necessarily that bad. I couldn't do it, but that thing's one way to do it directly in ASM, with lots of docs, and that's 2300 LOCs
|
|
#
¿
May 10, 2014 00:49
|
|
|
object1 (ˈɒbdʒɪkt) 1. to express or feel disapproval, dislike, or distaste; be averse. 2. informal a ridiculous or pitiable person, spectacle, etc
|
|
#
¿
May 10, 2014 01:10
|
|
|
suffix posted:the right proportion of ingredients depends not just on serving size, but also on the composition of the flours, which can vary between countries don't allow recipes with flour, problem solved
|
|
#
¿
May 19, 2014 05:15
|
|
|
'(lisp is a lot simpler)
|
|
#
¿
May 20, 2014 16:12
|
|
|
|
| # ¿ May 16, 2024 08:05 |
|
|
apt gangbang posted:why isnt ocaml's map tail recursive for most map operations on lists and whatever finite data structure, no matter the language (as long as it's immutable and strict), it may be faster to use a body-recursive version. Basically, the cost on stack will likely be little more than the duplication of pointers for the function you're using to map, and the pointer to the data structure you're mapping over. The data structure you're passing in basically limits how deep the stack goes. So you get two lists: the original one, and the mapped one. You've had an overhead of a couple pointers per list element to go with it, along with wherever the function returns. OTOH, a tail recursive map will need to generate a second list anyway, and do it in a reverse-manner by prepending to the head of the list if immutable, to avoid the cost of rewriting the list on every traversal. The end-result of this is that you then need to reverse the list at the end, creating a total of 3 lists: the original one, the intermediary one, and the final one (reversed intermediary). The tail-recursive map may overall have required more memory than the body-recursive one, potentially triggering GC and whatnot. That's super dependent on the language and its implementation, of course, but there are definite cases where a tail-recursive map can be seen as worse than a body-recursive one. You can benchmark them under different circumstances and see.
|
|
#
¿
May 27, 2014 04:45
|
|