> Would it be useful to have a trie that would index on the tail ends of > strings in reverse, for supporting quick searches for stuff 'ending in'? > > Are there any situations where it would be valuable to index on all > n-to-last-character subsequences of entries, or would the overhead just > be too tremendous for that to be useful? (I mean, can a trie be used to > implement fast lookups of any occurrence of a subsequence?)
I think you're talking about a suffix tree (which I've just discovered recently). The Trie interface and implementation I've submitted could be used i a suffix tree. Here are two links that show the relationship: http://www.csse.monash.edu.au/~lloyd/tildeAlgDS/Tree/Suffix/ http://www.cs.mcgill.ca/~cs251/OldCourses/1997/topic7/ Rich -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>