Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I was rambling on about this in 2001. You use Gnutella to search the 'transient web', which is just a bunch of home webservers serving up people's pages. Gnutella would link them all together and perform searches, as it does for MP3s now. Note: not searching filenames, but - like Google - a list of words in a HTML document.

I tried to get Gnucleus (John Marshall's Gnutella project) to incorporate full-text searches. He said he'd help but my code ended up - disabled - in Morpheus Preview Edition [itself based on Gnucleus] which was downloaded over 100,000,000 times (!). If anyone can be described as blowing a chance, that was me with MPE.

Then I tried Limewire, and wrote the code to interface with Lucene (a Java full text search engine), and they blew me off too.

Since then I've given up, but I did at least get a poster paper at HT03 about it!



I've thought that kademlia or some other DHT would be the ideal solution to this.

Think newsgroup/forum style conversational threading, with a separate data channel for large transfers, and landgrab DNS based on popularity.


That's really neat to know. I really like that idea of defining a static service, similar to DNS, for looking up a user's "transient web" data. Do you have a link to your paper? I'd love to read it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: