Comments on: A Haskell solution to Lars’ Problem

By: John Goerzen

John Goerzen — Tue, 05 Sep 2006 16:10:27 +0000

It is now here: [url]http://changelog.complete.org/posts/536-Another-Haskell-Solution-to-Lars-Problem.html[/ur]

By: The Changelog

The Changelog — Tue, 05 Sep 2006 16:08:53 +0000

Yesterday, I posted an 18-line solution to Lars’ language problem. One problem with it was that it was not very memory-efficient (or time-efficient, for that matter). In other words, it was optimized for elegance.

Here is a 22-line solution that is m

By: Antti-Juhani Kaijanaho

Antti-Juhani Kaijanaho — Tue, 05 Sep 2006 15:59:58 +0000

Maybe using the new ByteString? :)

By: John Goerzen

John Goerzen — Tue, 05 Sep 2006 01:55:09 +0000

You are right — this one is optimized for elegance. I will post one with better memory usage later today.

By: John Goerzen

John Goerzen — Tue, 05 Sep 2006 01:10:07 +0000

You’re right on #1 and #2 — though #2 is harmless. To fix #1, I’ve updated the example to replace isLower with isAlphaNum.

On #3, I still would need to, since we need a descending sort on the count and an ascending sort on the word.

By: Simon Michael

Simon Michael — Mon, 04 Sep 2006 21:32:59 +0000

Very interesting, thanks.

Using ghc 6.4.1 on ubuntu dapper I wasn’t able to complete the huge test within 400M of ram.. how much does it use for you ?

By: Dylan Thurston

Dylan Thurston — Mon, 04 Sep 2006 20:40:42 +0000

I see some bugs and obvious improvements in the code:
[list]
[*] You don’t treat the ASCII digits as word characters, as specified in the problem.
[*] You don’t seem to actually use Data.Map.
[*] If you switched the order of the count and the word, you wouldn’t need to specify your own comparison function.
[/list]

Also, for some reason the ‘<‘ symbol did not show up on the Planet Debian feed.