Sorry, I read your message too quickly; you are right -- tries are for character lookup, not strings. Tom Emerson of Basis gave a very nice talk on data structures for string lookup at the last UTC; you should contact him.
Mark __________________________________ http://www.macchiato.com â ààààààààààààààààààààà â ----- Original Message ----- From: "Theodore H. Smith" <[EMAIL PROTECTED]> To: "Mark Davis" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Tue, 2003 Nov 18 08:42 Subject: Re: Ternary search trees for Unicode dictionaries Hi Mark, Your tries are nice, however they are being used for single unicode characters, not a whole string of them, right? Well, sure some of them are being used for whole strings, but for me, ALL of mine will be used for whole strings. Yours are quite rare. Does this make the advantage of tries not so applicable for a string dictionary? When I am going to write a character property dictionary, I will use a trie, though. > We tend to use tries, which have very good performance > characteristics. See > "bits of unicode" on my site: www.macchiato.com. > > Mark > __________________________________ > http://www.macchiato.com > â ààààààààààààààààààààà â