> are we going to have a catalog of indexes? I don't think that would be very useful. My usage model for the indexes has two parts. First, there will be well-known indices for submission. Your client will submit your key plus metadata to any relevant indices (relevance will be determined by the content of the metadata). Second, there will be user-operated indices. They will troll the well-known indices for submissions, categorize, filter, and rate them, and then put them into another index.
The well-known indices will be agreed upon by client writers. The user-operated indices will be found the same way that web sites are found, from people recommending them. All of this will be transparent to the user, who will have an interface to search on various metadata fields or to search a particular index which they've heard about. > BTW, I'd be MORE than happy to implement the hashed-key search code. Unless > the existing Freenet code is pretty hairy, it shouldn't very long. If you want to then go ahead. I'd be interested to see what you come up with. Unless it can't be done totally client-side and requires actual changes to the node, in which case don't waste your time coding things until you convince Oskar that your code should go in the node. _______________________________________________
