|
tef posted:the only limit is yourself. 
|
#
?
Oct 17, 2013 18:02
|
|
|
|
| # ? Jun 6, 2024 13:32 |
|
|
Nomnom Cookie posted:i meant scheme. lisp is a miserable pile of kludges scheme is cute. people love the idea of a language whose spec they can casually read in an afternoon. the reality is not so nice. common lisp has a spec so hairy the only thing i can compare it to is iso c++. the abridged version is around 1500 pages. but the reality is quite pleasant:
as far as i am concerned, it is better to have a huge, ugly, well-specified standard than to have a small, comfortable one that i can't use for anything XML >>>> JSON CL >>>> Scheme Java >>>> pretty much everything Notorious b.s.d. fucked around with this message at 18:12 on Oct 17, 2013 |
|
#
?
Oct 17, 2013 18:09
|
|
|
that is not to say that CL isn't still under-specified with its 1500 page standard. there are a ton of de facto standards that should have been added to the spec at some point. (gray streams, CLIM, FFI) but it is much better-standardized than scheme
|
|
#
?
Oct 17, 2013 18:13
|
|
|
SAHChandler posted:civ 5 abandoned python in favor of lua oh that's cool i thought it was still python. do you know if it's stock lua or luajit, or is it something like that havok lua implementation?
|
|
#
?
Oct 17, 2013 18:14
|
|
|
Notorious b.s.d. posted:that is not to say that CL isn't still under-specified with its 1500 page standard. there are a ton of de facto standards that should have been added to the spec at some point. (gray streams, CLIM, FFI) wait isn;t CL old as hell? why on earth is FFI unspecified?
|
|
#
?
Oct 17, 2013 18:20
|
|
|
Hey everyone. I've got a quick public service announcement. If you have an application that humans log into, that application needs to use claims based authentication. That means Kerberos for your desktop apps and SAML for your web zones. If you think your application will only ever be used by users at your company OR users at the company that buys the application, you are wrong! Your sales team has already sold it (and/or the SSO features you were too lazy to include) to outside users! By spending that extra time to do it the right way, you will be saving loads of future headache and hatred. Remember: only you can prevent office fires due to users forgetting their 15 different logins.
|
|
#
?
Oct 17, 2013 18:33
|
|
|
spongeh posted:oh that's cool i thought it was still python. do you know if it's stock lua or luajit, or is it something like that havok lua implementation? i think it is just stock lua, no language modifications either.
|
|
#
?
Oct 17, 2013 18:53
|
|
|
Notorious b.s.d. posted:
Unfortunately the Java standard had Guy Steele of Scheme fame on it, making it a measly 672 pages, whereas the C++11 standard has a superior 1324 pages. I think that by your arguments, C++ is at the very least twice better than Java, given the specification has to be huger and uglier.
|
|
#
?
Oct 17, 2013 18:56
|
|
|
lisp is the worst language since it enables the programmer to be as insufferable as paul graham
|
|
#
?
Oct 17, 2013 19:13
|
|
|
Werthog 95 posted:http://en.wikipedia.org/wiki/Game_Oriented_Assembly_Lisp all the Jak games post-date the sony acquisition. I think they abandoned it because they couldn't get it to work on the Cell.
|
|
#
?
Oct 17, 2013 19:30
|
|
|
MononcQc posted:Unfortunately the Java standard had Guy Steele of Scheme fame on it, making it a measly 672 pages, whereas the C++11 standard has a superior 1324 pages. I think that by your arguments, C++ is at the very least twice better than Java, given the specification has to be huger and uglier. 
|
|
#
?
Oct 17, 2013 19:38
|
|
|
hehe his briefcase changes every panel please post more computoons
|
|
#
?
Oct 17, 2013 19:43
|
|
|
Notorious b.s.d. posted:[*]a wide range of high-performance compilers for every platform -- SBCL, Allegro CL, CCL, LispWorks racket is decently fast, isn't it?
|
|
#
?
Oct 17, 2013 19:49
|
|
|
pseudorandom name posted:all the Jak games post-date the sony acquisition. I think they abandoned it because they couldn't get it to work on the Cell. oh i didn't realize the broken-linked citation came from paul graham lol nevermind
|
|
#
?
Oct 17, 2013 19:57
|
|
|
tef posted:- badly designed. features added with little or no consideration to how they interact with the rest of the language,
|
|
#
?
Oct 17, 2013 20:03
|
|
|
PrBacterio posted:hey you posted the definition of the word "c++" c++, like "ain't", isn't a word friendo
|
|
#
?
Oct 17, 2013 20:06
|
|
|
Notorious b.s.d. posted:scheme is cute. people love the idea of a language whose spec they can casually read in an afternoon. the reality is not so nice. yes CL had a lot of work put into it. the spec certainly specifies many things. i hesitate to call it well engineered, however
|
|
#
?
Oct 17, 2013 22:23
|
|
|
decided to check out typescript, so git cloned it and the repo is loving 550 megs because it contains PDF and DOCX specs they're also not even taking PRs on it right now and use loving codeplex microsoft shouldn't be allowed to open source things, it might give the impression that you should actually try contributing to it
|
|
#
?
Oct 17, 2013 23:03
|
|
|
scheme is a nice starting point for a language. if you want to use it for anything, you're committing yourself to one implementation and all their nonstandard extensions, but with those extensions it at least works fine for a lot of projects that arent CPU intensive. good Schemes are smart enough to make C interop really pleasant tho, so it's easy to load off the important parts to C. it would obv be a lot better if cross-implementation portability was something anyone cared about. are people still working on r7rs-large? last i heard it was on hiatus while they worked on small. that sounds like it could help fix a lot of problems with scheme
|
|
#
?
Oct 18, 2013 00:06
|
|
|
MononcQc posted:Unfortunately the Java standard had Guy Steele of Scheme fame on it, making it a measly 672 pages, whereas the C++11 standard has a superior 1324 pages. I think that by your arguments, C++ is at the very least twice better than Java, given the specification has to be huger and uglier. guy steele was also the principal author of the abridged CL standard. cool guy (obviously there was a committee to define the standard, steele was the principal author of the document outlining the standard) Notorious b.s.d. fucked around with this message at 00:20 on Oct 18, 2013 |
|
#
?
Oct 18, 2013 00:14
|
|
|
abraham linksys posted:decided to check out typescript, so git cloned it and the repo is loving 550 megs because it contains PDF and DOCX specs historically, lots of things have been open source but in no way interested in independent contributions. unix and java come to mind as famous examples (open source but non-free java later gave way to openjdk, obviously. and who gives any fucks about att unix in 2013)
|
|
#
?
Oct 18, 2013 00:16
|
|
|
MeramJert posted:racket is decently fast, isn't it? compared to other scheme implementations? yes compared to CL or java? lol
|
|
#
?
Oct 18, 2013 00:16
|
|
|
Shinku ABOOKEN posted:wait isn;t CL old as hell? why on earth is FFI unspecified? because the CL standard never, ever got updated after 1994. and lots of stuff that was well-known in 1994 was omitted because it was too controversial however much you love/hate the C++XX fiasco of updating the standard every few years, common lisp demonstrates that the alternative is worse. please update your standards. think of the children
|
|
#
?
Oct 18, 2013 00:18
|
|
|
MononcQc posted:Unfortunately the Java standard had Guy Steele of Scheme fame on it, making it a measly 672 pages, whereas the C++11 standard has a superior 1324 pages. I think that by your arguments, C++ is at the very least twice better than Java, given the specification has to be huger and uglier. wait wasn't there something likes this... Otto Skorzeny posted:adopt the Chuck Moore Way(tm) and change the problem until forth is an elegant* solution, then call everyone else who has solved the problem in a way that doesn't ignore all the ugly but necessary aspects an idiot pissfucker necessary. sufficient. hurhfuhsehfsjdfsdk
|
|
#
?
Oct 18, 2013 01:54
|
|
|
Notorious b.s.d. posted:guy steele was also the principal author of the abridged CL standard. cool guy Guy Steele is loving cool. https://www.youtube.com/watch?v=_ahvzDzKdB0 Still my favorite talk ever.
|
|
#
?
Oct 18, 2013 01:59
|
|
|
MononcQc posted:Guy Steele is loving cool. guy steele raped and killed a 9 year old girl in 1998. which is why we do not have operator overloading in java.
|
|
#
?
Oct 18, 2013 02:08
|
|
|
everyone is just letting him do computer poo poo when thousands of porn stars each year can't even make up a name as good as his, much less be born with it?
|
|
#
?
Oct 18, 2013 02:12
|
|
|
He also worked on Fortress, which was objectively crazy awesome in its approach of being a language for mathematicians that would be both in a regular notation to execute and could be displayed as generated PDFs or actual equations when editing, and would be auto-parallelized. Guy Steele is the most awesome guy around.
|
|
#
?
Oct 18, 2013 02:15
|
|
|
Source code can be rendered as ASCII text, in Unicode, or as a prettied image. This will allow for support of mathematical symbols and other symbols in the rendered output for easier reading. mad genius
|
|
#
?
Oct 18, 2013 02:17
|
|
|
the problem with lisp is macros however, the main reason to use lisp is also macros
|
|
#
?
Oct 18, 2013 03:39
|
|
|
Notorious b.s.d. posted:compared to other scheme implementations? yes according to these definitive benchmarks, they're not that far off: http://benchmarksgame.alioth.debian.org/u64q/racket.php
|
|
#
?
Oct 18, 2013 04:03
|
|
|
i love reading about all the awesome technologies that have grown up around java/j2ee and knowing they are forever out of reach because my boss is terrified of anything new and different i mean i understand that you dont go updating the piles and piles and piles of legacy code just for the fun of it but we have a brand new app, completely new development, about to go into uat that uses nothing but raw servlets and jsps filled to the brim with scriplets in tyool 2013 somebody help
|
|
#
?
Oct 18, 2013 05:27
|
|
|
HORATIO HORNBLOWER posted:nothing but raw servlets and jsps filled to the brim with scriplets mods
|
|
#
?
Oct 18, 2013 05:32
|
|
|
MeramJert posted:according to these definitive benchmarks, they're not that far off: http://benchmarksgame.alioth.debian.org/u64q/racket.php they're pretty far apart. but racket is only half the speed and three times as memory hungry. staying inside the same order of magnitude as SBCL or java is impressive. (p.s. it amuses me that clojure is still 2x the speed of racket in these toy benchmarks)
|
|
#
?
Oct 18, 2013 14:44
|
|
|
software f/p implementations that don't implement all four rounding modes correctly  
|
|
#
?
Oct 18, 2013 14:45
|
|
|
fp is just in general
|
|
#
?
Oct 18, 2013 14:48
|
|
|
Notorious b.s.d. posted:(p.s. it amuses me that clojure is still 2x the speed of racket in these toy benchmarks) it doesn't surprise me that the pile of hacks to make clojure sit atop a very good general purpose jitting vm is more performant than racket sitting on the racket vm, which isn't anything like as mature as the jvm (but then again i spectated parrot's self-destruction and that probably left me much less sanguine on new vm projects than the average guy - someone who got involved with pypy or luajit or whatever would probably be more optimistic about racket)
|
|
#
?
Oct 18, 2013 14:50
|
|
|
Notorious b.s.d. posted:fp is we had a long discussion here about whether it would be worth it to do our own fixed-point library to make things less of a mindfuck, but the conclusion we came to was that we needed more dynamic range than we could get with a fixed point implementation that was simple enough to get correct quickly and performant enough to run on the low end freescales we use on cheap stuff, so we continue to use a mishmash of half baked vendor implementations of ieee754 single precision floats. it makes me want to quote passages from kipling in emails.
|
|
#
?
Oct 18, 2013 14:54
|
|
|
i got sorta interested in using lua but i cant think of anything to use it for. i googled about lua and found this thing some nice people made using lua http://en.wikipedia.org/wiki/Flame_(malware) quote:Once a system is infected, Flame can spread to other systems over a local network or via USB stick. It can record audio, screenshots, keyboard activity and network traffic.[6] The program also records Skype conversations and can turn infected computers into Bluetooth beacons which attempt to download contact information from nearby Bluetooth enabled devices.[7] This data, along with locally stored documents, is sent on to one of several command and control servers that are scattered around the world. The program then awaits further instructions from these servers quote:Due to the size and complexity of the program—described as "twenty times" more complicated than Stuxnet—the Lab stated that a full analysis could require as long as ten years.[7] 
|
|
#
?
Oct 18, 2013 15:22
|
|
|
|
| # ? Jun 6, 2024 13:32 |
|
|
Otto Skorzeny posted:it doesn't surprise me that the pile of hacks to make clojure sit atop a very good general purpose jitting vm is more performant than racket sitting on the racket vm, which isn't anything like as mature as the jvm (but then again i spectated parrot's self-destruction and that probably left me much less sanguine on new vm projects than the average guy - someone who got involved with pypy or luajit or whatever would probably be more optimistic about racket) what's the current perl 6 situation? i would love to feel hope 
|
|
#
?
Oct 18, 2013 15:38
|
|