4 ms·

I've been wondering about distributed solutions to torrent indexing for a while - maybe some of you know some good discussions on this topic? What I noticed is

by pudo 17y ago

I've been wondering about distributed solutions to torrent indexing for a while - maybe some of you know some good discussions on this topic? What I noticed is that .torrent files are essentially weird, JSON-like dictionaries. It would be trivial to place some metadata in there (ID3 tag info, DublinCore or even just a link to imdb) and this could be indexed very easily. Even easier: turn the parts protocol into a microformat and place some invisible part listings in your HTML files. I feel like such approaches would reduce the need for centralised torrent search sites like tpb/mininova/etc.

extension 17y ago
BitTorrent already uses a distributed hash table for tracking. A search index can be built on a DHT by using normalized search terms as keys, as with e.g. http://www.faroo.com http://www.faroo.com This effectively results in individuals being responsible for specific words in the index. I would assume that this becomes a weakness if particular search terms are staggeringly popular or carry liability, though I'm no expert and there may be solutions to this using redundancy and flux in the allocation.