First, as to the start with a quad-nested loop. I frequently (especially when writing) start with the most elemental brute force implementation possible. There are a couple of really good reasons to do this.

A brute force solution is usually much easier to prove correct. It is often much easier to understand, especially for someone who is not an experiences programmer or developer. So it provides a really good starting point for a more optimized version of the algorithm. It’s natural for a developer to jump directly into something more optimal, and it certainly makes sense to do that for regular work, but for an article I happen to like the slower approach.

Second, I’m not sold on the notion that a choice of a static or scripted language predisposes you to certain behaviors. I think it more likely that the facilities your language provides are what drives your choices.

As in this example. It’s no secret that C/C++ (the language I use at work all day, and thus my language of choice) is lamentably weak on string handling. Further, associative arrays are a relatively new feature, and hashed associative arrays are still not part of the standard. (Unless TR1 has passed and I haven’t received the news yet.)

So someone who writes Perl code all day is much more likely to have worked on problems analyzing text, using regular expressions, and using strings as keys into other data structures. It’s not the scripting that does this, it would be just as true in a statically typed language that had great string support.

The interesting thing is that an optimal solution to this problem ends up looking similar in all the languages we’ve seen here. Even the unfamiliar Haskell looks vaguely familiar.

I’d really enjoy seeing some cases with language X, Y, and Z where the optimal C++ solution was not so good, and the optimal X/Y/Z solution looked amazingly different. Last time I did any functional programming was maybe 1975? and that seems like one place where you might be able to Think Different.

Check out Anders Pearson’s analysis about your approach to the problem and how he sees that relates to programming languages:

http://thraxil.org/users/anders/posts/2007/10/30/A-Simple-Programming-Puzzle-Seen-Through-Three-Different-Lenses/

An interesting read, as was this article.

In essence the solution consists of the following
Build a list of the unique 2 state pairs and combine the letters in the pair into string. store the list of the original strings in on table on a sql db and store a list of alphabettically sorted strings in another column in that same table. do a select(unsortedstring) where count(sortedstring > 1) order by sortedstring. you will then get the answer to the problem

Cheers! –Tom