Web Development

1. I want to write the program myself. I don't want to download, modify, and contribute.
2. I think I'll go with the interpreter.
3. I have not written a program more complex than a Fibonacci number calculator before.

Well, it is a project. You have to have an acquisition engine to find target documents. You need a repository friendly to search. You need a user interface to submit searches and present finds. An admin interface for submitting new target areas for the acquisition engine. You need a computer, usually on a network and usually lots of storage.

You want a data structure that expands, updates, deletes in a other-user-invisible way, like leaf-to-root modification. Users coming down the old tree are not bothered by new trees you build to replace, or new subtrees.

A lot of code and though goes into dealing with kill-words, words and phrases that happen so often you never want to index them. You can discover them as they hit a threshold, or just trim them as needed for space.

JAVA using persistent objects may work well for this. You might want to make your own persistent objects out of map'd flat files.