|
If you don't use Apache you'll find yourself reimplementing it in your app. See also: mod_wsgi.
|
#
?
Feb 23, 2013 16:30
|
|
|
|
| # ? Jun 6, 2024 19:46 |
|
|
Nomnom Cookie posted:its pretty sick and tight like akka would be if akka's delivery guarantee was better than gently caress YOU I BURNT YOUR MAIL Delivery guarantees don't mean that much. It's delivered, but doesn't say if it was read, or acted upon. You end up needing an application-level ack'ing anyway, and all the related mechanisms.
|
|
#
?
Feb 23, 2013 16:30
|
|
|
Socracheese posted:my theory of computation class prof threatened to make us program stuff for one of these: it's not a true universal turing machine because it doesn't have an infinite tape  xf86enodev posted:the only reason php is as popular as it is is apache mod_php ftfy
|
|
#
?
Feb 23, 2013 18:40
|
|
|
don't be a pussy, program a CA TM https://www.youtube.com/watch?v=My8AsV7bA94
|
|
#
?
Feb 23, 2013 18:56
|
|
|
a bit of a tangent, but I've been looking over the first couple chapters or so of Odersky's Scala book just out of curiosity and it almost seems like the language is trying to do too much at once also the compiler really is slow as balls for even tiny programs
|
|
#
?
Feb 23, 2013 22:44
|
|
|
JVM and the Java Standard Library just seem like a crappy fit for a LISP to me It's a pity Parrot was basically vapourware because it would have been nice to have Ein Managed VM to rule them all, but if you're going to have a managed VM then that means you've got a type system in the VM. Then you're going to have some awful hack to get another language's type system to play ball, other than the language the VM was designed for in the first place, I mean. I know the CLR has F# and all that, but it's mostly either C# and VB.NET, and VB.NET is semantically identical to C#, it's just a different syntax skin and I mean lol seriously wtf is the point of interopping two gratuitously different syntaxes that are otherwise identical.
|
|
#
?
Feb 24, 2013 00:14
|
|
|
Mr Dog posted:JVM and the Java Standard Library just seem like a crappy fit for a LISP to me scala is not a lisp. also if you want f# without all the .net faggotry, try ocaml! 
|
|
#
?
Feb 24, 2013 00:29
|
|
|
JewKiller 3000 posted:scala is not a lisp. also if you want f# without all the .net faggotry, try ocaml! just make my boss stop thinking "dot net" is a programming language
|
|
#
?
Feb 24, 2013 00:33
|
|
|
Mr Dog posted:It's a pity Parrot was basically vapourware because it would have been nice to have Ein Managed VM to rule them all, but if you're going to have a managed VM then that means you've got a type system in the VM. Then you're going to have some awful hack to get another language's type system to play ball, other than the language the VM was designed for in the first place, I mean. So, having farted around in #parrot-dev for a while, IMO this (and a few other instances of over-generalization) is what made parrot vaporware for most purposes. They repeatedly made the object model more general to support every existing or conceivable object system in any language, and as a result it sucked for their only actual "customer" and several other interested parties (eg. there was a pretty successful port of Lua to Parrot back when it had some momentum, but it was on the order of 100x slower than LuaJIT). The decision to make Parrot a VM for everything from the get-go meant it was never the VM for anything; even Rakudo (the closest thing there is to a Perl6 implementation) is moving away from Parrot now as even the long-time contributors are getting frustrated and leaving. The normal argument for this from the Powers that Blotto Skorzany fucked around with this message at 00:44 on Feb 24, 2013 |
|
#
?
Feb 24, 2013 00:41
|
|
|
^^^^ yea it kinda seemed like a doomed project from the start. Interesting to see an inside-ish experience, though (Parrot was fresh in my mind because LWN reposted a core Parrot contributor's ragequit recently) I can highly recommend LWN in general, it's very thorough on the technical stuff and very light on the dumb slashdot/reddit editorialising and circlejerking. To the extent that I only lurk there because I don't want to look like a dumbass lol JewKiller 3000 posted:scala is not a lisp. also if you want f# without all the .net faggotry, try ocaml! My bad, I was thinking of Clojure. Point still stands tho Sapozhnik fucked around with this message at 01:04 on Feb 24, 2013 |
|
#
?
Feb 24, 2013 01:01
|
|
|
the java standard library is pretty rigid and depressing when used in any modestly more dynamic language than java lisp failing certainly has nothing to do with a lack of quality implementations at any rate, one of the few languages where one is genuinely spoiled with tons of fine implementations
|
|
#
?
Feb 24, 2013 01:53
|
|
|
gucci void main posted:a bit of a tangent, but I've been looking over the first couple chapters or so of Odersky's Scala book just out of curiosity and it almost seems like the language is trying to do too much at once nah, scala is cool. sbt helps with the slowness of the compiler, but yeah I wish it weren't so slow.
|
|
#
?
Feb 24, 2013 01:58
|
|
|
MononcQc posted:Delivery guarantees don't mean that much. It's delivered, but doesn't say if it was read, or acted upon. You end up needing an application-level ack'ing anyway, and all the related mechanisms. I think I didn't quite mean delivery guarantee. Storm guarantees that a tuple source will be notified eventually if a tuple fails. there's also some machinery to help you with exactly-once or transactional processing, but it's not transparent and doesn't try to promise that you can write Storm code and magic distributed sauce will fix everything. the bookkeeping is done in Zookeeper so there are probably weird failure modes involving an entire Zookeeper cluster falling over at the wrong time but whatever. it still beats the poo poo out of trying to build everything yourself on top of "you didn't actually care whether that message was received" akka
|
|
#
?
Feb 24, 2013 02:11
|
|
|
p.s. my point is that when it fits your problem, actor model is fuckin awesome for distributed systems. also i wish i knew more about the reliability analysis stuff shaggier shits on bc i think its definitely applicable to distributed poo poo where i need a dozen systems to cooperate to handle one request
|
|
#
?
Feb 24, 2013 02:15
|
|
|
Nomnom Cookie posted:p.s. my point is that when it fits your problem, actor model is fuckin awesome for distributed systems. also i wish i knew more about the reliability analysis stuff shaggier shits on bc i think its definitely applicable to distributed poo poo where i need a dozen systems to cooperate to handle one request i learned it first in a class at school called "engineering probability", which was the ECSE department's take on a probability & stats class. basically it meant that the examples for a poisson distribution used packets coming into a router instead of cars going through a tollbooth (as the MANE department's prob&stats class used). then it got reinforced a few more times. it's me, im the ciseducated oppressor ruthlessly othering people
|
|
#
?
Feb 24, 2013 02:42
|
|
|
http://www.theverge.com/2013/2/22/4016582/after-so-many-hacks-why-wont-java-just-go-away
|
|
#
?
Feb 24, 2013 03:21
|
|
|
gucci void main posted:http://www.theverge.com/2013/2/22/4016582/after-so-many-hacks-why-wont-java-just-go-away "Java applets are bad, lets torch the entire jvm."
|
|
#
?
Feb 24, 2013 03:28
|
|
|
Condiv posted:"Java applets are bad, lets torch the entire jvm." 7.0 / 10
|
|
#
?
Feb 24, 2013 03:36
|
|
|
i came back to my browser after doing some planging and i had the new post screen open but i forgot what i was going to write so for now lets have a "lol java"
|
|
#
?
Feb 24, 2013 03:39
|
|
|
Mr Dog posted:JVM and the Java Standard Library just seem like a crappy fit for a LISP to me yes, but your opinion doesn't matter, because java is the titan. the jvm is so important that writing a new lisp language tailored to fit the jvm's peculiarities got more traction than any lisp has had in my lifetime this is especially funny because clojure has a tiny standard library and wears a horrifying functional programming straitjacket. using mutable state or trying to implement oo in clojure will make you cry, whereas those things are trivial in CL parrot was started long enough that the jvm was not the obvious winner-take-all, and that is the only reason it ever existed
|
|
#
?
Feb 24, 2013 04:03
|
|
|
Notorious b.s.d. posted:yes, but your opinion doesn't matter, because java is the titan. the jvm is so important that writing a new lisp language tailored to fit the jvm's peculiarities got more traction than any lisp has had in my lifetime idgi when people say state is hard in purely functional languages. The State monad in haskell is pretty easy to understand and the ST monad, aside from its type is as well.
|
|
#
?
Feb 24, 2013 04:14
|
|
|
[C++11] template template parameter template parameter pack that depends on another parameter pack
|
|
#
?
Feb 24, 2013 04:42
|
|
|
Notorious b.s.d. posted:parrot was started long enough that the jvm was not the obvious winner-take-all, and that is the only reason it ever existed well, no. parrot existed due to concerns that the jvm is a lovely fit for dynamic languages in general and the type system of perl 6 in particular (still true - the jvm's support for dynamic languages is a joke; additionally, perl 6's support for continuations would be somethine like impossible to do sanely on the jvm ), and because it was thought that a register-based virtual machine could perform significantly better than a virtual stack machine. while the idea of a register based vm is still attractive, parrot started out as a second-system perl 5.6 vm and turned into an overgeneralized behemoth and organizational clusterfuck so the performance never came close to materializing and it never gained any traction. if parrot the project had been run as well as pypy the project, we wouldn't be having this discussion. heck, in 3 years pypy might be a more popular target than the jvm for dynamic languages
|
|
#
?
Feb 24, 2013 05:21
|
|
|
Mr Dog posted:I know the CLR has F# and all that, but it's mostly either C# and VB.NET, and VB.NET is semantically identical to C#, it's just a different syntax skin and I mean lol seriously wtf is the point of interopping two gratuitously different syntaxes that are otherwise identical. at the time programmers of c-like languages were outnumbered 10:1 by monkeys writing tons of poo poo in vb6 and office macros (vba).
|
|
#
?
Feb 24, 2013 05:42
|
|
|
Notorious b.s.d. posted:yes, but your opinion doesn't matter, because java is the titan. the jvm is so important that writing a new lisp language tailored to fit the jvm's peculiarities got more traction than any lisp has had in my lifetime this is perhaps more a testament to how disliked java is, rather than the ubiquity of the jvm. technically I think elisp has more users than clojure. going by local biases, scala seems to be the more popular notjava on the jvm. that said, making a lisp that lisp weenies can smuggle into middleware environments is possibly the first lisp ever to take adoption into account. i had a soft spot for hickey's languages until he gave the lisp away and charged for the datalog. quote:parrot was started long enough that the jvm was not the obvious winner-take-all, and that is the only reason it ever existed the jvm isn't the winner takes all though, it's all been but abandoned on the desktop, and if it wasn't for goog, it would have disappeared from mobile altogether (but dalvik is not the jvm, until a court shows otherwise). the jvm is basically the current defacto platform for business middleware, although c# has been eating away at it. i.e. it's the current generation's cobol. oracle are doing their best to stagnate it though, so there is hope for the future at least. Otto Skorzeny posted:well, no. parrot existed due to concerns that the jvm is a lovely fit for dynamic languages in general and the type system of perl 6 in particular (still true - the jvm's support for dynamic languages is a joke; additionally, perl 6's support for continuations would be somethine like impossible to do sanely on the jvm ), parrot started as an april fools joke that some perl people took seriously, and then, in a similar vein to perl 6, they tried to make it do everything before making it do anything. then, after years of pissing around, even the current champions of perl 6 abandoned it. quote:and because it was thought that a register-based virtual machine could perform significantly better than a virtual stack machine. this is generally true, so i'm lead to believe by lua 4 -> lua 5, but stack vms are easier to implement and target. quote:while the idea of a register based vm is still attractive, parrot started out as a second-system perl 5.6 vm and turned into an overgeneralized behemoth and organizational clusterfuck so the performance never came close to materializing and it never gained any traction. c.f. perl 6 quote:if parrot the project had been run as well as pypy the project, we wouldn't be having this discussion. heck, in 3 years pypy might be a more popular target than the jvm for dynamic languages well, in the sense that it had a specific goal in mind, targeted a language that didn't change week in, week out, and aimed to be a replacement for a specific part of the system, rather than a total replacement for the ecosystem. so yeah if it had different goals and design for the outset, it might have been more successful. i.e having nothing to do with the catastrophe that is moving on from perl 5 ps. I don't think you 'target pypy', you write an interpreter in rpython and compile it into a faster interpreter under pypy.
|
|
#
?
Feb 24, 2013 06:31
|
|
|
Mr Dog posted:JVM and the Java Standard Library just seem like a crappy fit for a LISP to me the easiest way to make a lisp is to throw away the front end of the compiler and let your users write in the data structures your compiler uses. (p.s. Clojure isn't really much like the classical lisps, like scheme or CL imho. it has some syntax and doesn't reject the outside world) quote:It's a pity Parrot was basically vapourware because it would have been nice to have Ein Managed VM to rule them all, but if you're going to have a managed VM then that means you've got a type system in the VM. Then you're going to have some awful hack to get another language's type system to play ball, other than the language the VM was designed for in the first place, I mean. i'd rather have an unmanaged but safe vm personally. turns out lots of people would too. managed vms basically dictate far too much about the host languages. quote:I know the CLR has F# and all that, but it's mostly either C# and VB.NET, and VB.NET is semantically identical to C#, it's just a different syntax skin and I mean lol seriously wtf is the point of interopping two gratuitously different syntaxes that are otherwise identical. vb has mildly different semantics but was created as a migration path from vb6. if you've used vb6 you would be thankful for vb.net. i mean, what's the point of supporting a large userbase on your new technology platform, and making it awkward to change from c#. doing the opposite of what you said sounds far more stupid.
|
|
#
?
Feb 24, 2013 06:36
|
|
|
Malcolm XML posted:idgi when people say state is hard in purely functional languages. The State monad in haskell is pretty easy to understand and the ST monad, aside from its type is as well. they mean 'i have to be precise about state and work out how to smuggle it throughout my codebase' and this is harder than just 'pfft global implicit state'
|
|
#
?
Feb 24, 2013 06:38
|
|
|
tef posted:the jvm isn't the winner takes all though, it's all been but abandoned on the desktop, and if it wasn't for goog, it would have disappeared from mobile altogether (but dalvik is not the jvm, until a court shows otherwise). you'd have typed "was abandoned years ago" where it not for goddamned minecraft...
|
|
#
?
Feb 24, 2013 06:50
|
|
|
Base Emitter posted:you'd have typed "was abandoned years ago" where it not for goddamned minecraft... eclipse
|
|
#
?
Feb 24, 2013 06:51
|
|
|
gucci void main posted:eclipse well, for normal human people
|
|
#
?
Feb 24, 2013 06:59
|
|
|
Base Emitter posted:well, for normal human people Can you recommend a cross-platform alternative for C++? Code::Blocks? e: This one is tough if you consider developing for Android something that normal human people would do. Even if you were to omit eclipse, there's still the build system based on ant, the fact that the GUI for managing virtual devices and SDKs is available only through Eclipse, so on. Bream fucked around with this message at 07:22 on Feb 24, 2013 |
|
#
?
Feb 24, 2013 07:10
|
|
|
tef if you wanna, my forrst is mnnd im a lil drunk now but my application is for real. seems like a p good site to get feedback on my bullshit code e: i woulda sent this in some message but i guess you aint got those
|
|
#
?
Feb 24, 2013 07:34
|
|
|
Bream posted:Can you recommend a cross-platform alternative for C++? Code::Blocks? vim/subl
|
|
#
?
Feb 24, 2013 07:35
|
|
|
Bream posted:e: This one is tough if you consider developing for Android something that normal human people would do. normal human people don't develop that's the point
|
|
#
?
Feb 24, 2013 07:41
|
|
|
gucci void main posted:vim/subl I hadn't ever seen Sublime Text, that's actually quite nice! Thanks!
|
|
#
?
Feb 24, 2013 07:43
|
|
|
i can't sleep so i'm being really anal about parsing https://github.com/mojombo/toml/issues/42
|
|
#
?
Feb 24, 2013 08:10
|
|
|
tef posted:i can't sleep so i'm being really anal about parsing laughing at how many of those questions would have been answered if he had just provided a grammar for his lovely "language"
|
|
#
?
Feb 24, 2013 08:35
|
|
|
tef posted:i can't sleep so i'm being really anal about parsing i read it, and i agree, and like an idiot i dont know how else to express it except by quoting and typing +1
|
|
#
?
Feb 24, 2013 08:36
|
|
|
tef posted:i can't sleep so i'm being really anal about parsing haha holy poo poo
|
|
#
?
Feb 24, 2013 08:42
|
|
|
|
| # ? Jun 6, 2024 19:46 |
|
|
Carthag posted:i read it, and i agree, and like an idiot i dont know how else to express it except by quoting and typing +1 github has an emoticon for that 👍
|
|
#
?
Feb 24, 2013 08:48
|
|
