Yesterday, I started to work on a small library to parse ID3v2 tags (For those of you that don't know what it is; it's the information embedded in mp3 files, that tells you about the artist/title/recorded year/genre/bpm etc). I normally throw up some mock-up code, and refine it afterwards. Since this is a operation that I would possibly run on *a lot* of files, I want this operation to be fast. So I allready thought about rewriting it with /unsafe, and use pointer arithmetics to parse the tags).
As I started to hack on the library, and tried to comprehend the specification (found at http://www.id3.org/) - I came across "problems" - something that we programmers normally does when we do this kind of development, like "what is the best way to check a 3 byte large array if it contains "ID3" and other minor things. Should we just run throught them with nested if-statements, to check each element - and bail out, if it doesn't match. Or should we create an array containing the match criteria - and check the arrays against each other.
There really should be a wiki that could be used to discuss these small day-to-day problems - that we all run accross.
--larsw - http://www.sral.org/
I actually read that article a week ago - quite interesting...