Comments on: Naïve Text Parsing in Scala http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala (permanently in beta) Mon, 09 Jan 2012 20:21:24 -0800 http://wordpress.org/?v=2.9.2 hourly 1 By: Mark http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala/comment-page-1#comment-3684 Mark Sun, 01 Jun 2008 23:51:55 +0000 http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala#comment-3684 Nice post, I'm thinking there is a smaller, more svelte implementation hiding in there somewhere... :P Nice post, I’m thinking there is a smaller, more svelte implementation hiding in there somewhere… :P

]]> By: David Linsin http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala/comment-page-1#comment-3682 David Linsin Sun, 01 Jun 2008 10:39:49 +0000 http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala#comment-3682 Awesome blog post! Keep it coming, it's always fun to follow you on these little journeys! Awesome blog post! Keep it coming, it’s always fun to follow you on these little journeys!

]]> By: Daniel Spiewak http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala/comment-page-1#comment-3663 Daniel Spiewak Mon, 26 May 2008 16:29:24 +0000 http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala#comment-3663 I considered using parser combinators, however they would be overkill for a number of reasons. The entire syntax for C-style printf can be written as a regular expression, thus there is no need for a full pushdown automaton (as mentioned in the post). DFAs are far more efficient than PDAs (because there is no need for a supplementary stack) and also tend to be easier to work with when written by hand. It's actually *harder* to properly express printf as a CFG than it is to just bull ahead with a DFA (I had to formulate the printf grammar specifically for this post, I didn't use it at all in my implementation). It is true that parser combinators would probably be easier to read than my hand-coded approach, but I think the reduced efficiency would be too much of a tradeoff. I haven't spent much time optimizing my parser (read: none at all), but according to my test results, it's almost as fast as the well-vetted JDK implementation. If that's the case, then theoretically I can squeeze at least as much performance out of it, maybe even better. I don't really have this opportunity if I use a pre-baked parser generator. I considered using parser combinators, however they would be overkill for a number of reasons. The entire syntax for C-style printf can be written as a regular expression, thus there is no need for a full pushdown automaton (as mentioned in the post). DFAs are far more efficient than PDAs (because there is no need for a supplementary stack) and also tend to be easier to work with when written by hand. It’s actually *harder* to properly express printf as a CFG than it is to just bull ahead with a DFA (I had to formulate the printf grammar specifically for this post, I didn’t use it at all in my implementation).

It is true that parser combinators would probably be easier to read than my hand-coded approach, but I think the reduced efficiency would be too much of a tradeoff. I haven’t spent much time optimizing my parser (read: none at all), but according to my test results, it’s almost as fast as the well-vetted JDK implementation. If that’s the case, then theoretically I can squeeze at least as much performance out of it, maybe even better. I don’t really have this opportunity if I use a pre-baked parser generator.

]]> By: Justin Camerer http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala/comment-page-1#comment-3662 Justin Camerer Mon, 26 May 2008 16:15:47 +0000 http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala#comment-3662 Great post! I really enjoy every post you do on Scala. I am learning Scala and (re)learning many FP ideas, and your blog has been a great help. I don't think I'd be of much help, but I'd still love to see the code and specs if you had time to ship them my way. Thanks, Justin Great post! I really enjoy every post you do on Scala. I am learning Scala and (re)learning many FP ideas, and your blog has been a great help.

I don’t think I’d be of much help, but I’d still love to see the code and specs if you had time to ship them my way.

Thanks,
Justin

]]> By: Eric Torreborre http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala/comment-page-1#comment-3657 Eric Torreborre Mon, 26 May 2008 08:51:04 +0000 http://www.codecommit.com/blog/scala/naive-text-parsing-in-scala#comment-3657 Hi Daniel, I am a real fan of your blog, you write very frequent and interesting posts Could you send me your spec please? I'm interested to see how you and other users are using the library, to refine and improve it. Back to the subject, wouldn't it be a good idea anyway to use parser combinators, in terms of: -readability of what the implementation does? -speed of implementation? Thanks, Eric. Hi Daniel,

I am a real fan of your blog, you write very frequent and interesting posts

Could you send me your spec please? I’m interested to see how you and other users are using the library, to refine and improve it.

Back to the subject, wouldn’t it be a good idea anyway to use parser combinators, in terms of:

-readability of what the implementation does?
-speed of implementation?

Thanks,

Eric.

]]>