Re:The Snow Programming Language
by mahmud on Saturday August 08, 2009 @03:30PM
(#749698)
Only runs on x86-64 architectures and operating systems.
Only runs on POSIX-compliant operating systems.
Snow currently requires the experimental GCC version 4.4
Snow also could have benefited from being prototyped in Awk and Sed first. How can you be influenced by "Smalltalk, Python, Ruby and JavaScript", and still end up writing your compiler in C++?
Good luck to the author, if this is a first attempt (but compiler writing != language design.)
Re:The Snow Programming Language
by jcw on Saturday August 08, 2009 @03:32PM
(#749701)
What exactly is interesting about this language? The two strongest qualities I see (easy binding with dlopen and parallelization) are unimplemented, and already exist in similar languages.
Re:The Snow Programming Language
by judofyr on Saturday August 08, 2009 @06:55PM
(#749952)
First of all I should say that it's unfortunate that this is spreading while it's still very much in alpha. But hey, they did put the website there and this is the internett. I also feel that I should clarify some parts since it's not completely up to date:
It's prototype-based at core, but will have a class system built upon the prototype-infrastructure.
The C++ version supports 32bit now, but more interesting, they're rewritig it in pure C. See the mailing list for more details (snow-dev at Google Groups).
And just to make it crystal clear: This isn't a release. They just wanted to create a site to give some information on the project, and some guys decided to post this to Reddit/HN. It's still far away from 0.1.
Re:The Snow Programming Language
by rdtsc on Saturday August 08, 2009 @10:26PM
(#750204)
I like it.
Using '=' for assignment makes less sense than using <- or := or : because traditionally (i.e. mathematically) '=' work like a declaration or a constraint. Then there is the confusion between = and == that has always resulted in countless bugs.
Of course, assignment is a very common operation and I would rather use 1 char (:) for it rather than two (<-, :=).
Re:The Snow Programming Language
by albertcardona on Saturday August 08, 2009 @03:27PM
(#749694)
Snow _could_ be great. Its delivery as a completely unfinished project, though, gives it an air of non-seriousness.
Compare, for instance, with the first public release of Clojure, which was very usable for everyday programming, and came along with detailed explanations on the rational and motivations behind the creation of the new language.
Only runs on POSIX-compliant operating systems.
Snow currently requires the experimental GCC version 4.4
Snow also could have benefited from being prototyped in Awk and Sed first. How can you be influenced by "Smalltalk, Python, Ruby and JavaScript", and still end up writing your compiler in C++?
Good luck to the author, if this is a first attempt (but compiler writing != language design.)
It's prototype-based at core, but will have a class system built upon the prototype-infrastructure.
The C++ version supports 32bit now, but more interesting, they're rewritig it in pure C. See the mailing list for more details (snow-dev at Google Groups).
And just to make it crystal clear: This isn't a release. They just wanted to create a site to give some information on the project, and some guys decided to post this to Reddit/HN. It's still far away from 0.1.
And besides, I'm not sure what this is good for.
Using '=' for assignment makes less sense than using <- or := or : because traditionally (i.e. mathematically) '=' work like a declaration or a constraint. Then there is the confusion between = and == that has always resulted in countless bugs.
Of course, assignment is a very common operation and I would rather use 1 char (:) for it rather than two (<-, :=).
http://en.wikipedia.org/wiki/Maxima_%28software%29
Compare, for instance, with the first public release of Clojure, which was very usable for everyday programming, and came along with detailed explanations on the rational and motivations behind the creation of the new language.
I'm not making any claims as to which is "better", just pointing out that it's a valid way of doing things.