Archive of UserLand's first discussion group, started October 5, 1998.

JIT-SE (Just In Time - Search Engine)

Author:David Galbraith
Posted:5/18/2000; 6:38:19 AM
Topic:JIT-SE (Just In Time - Search Engine)
Msg #:17375
Prev/Next:17374 / 17376

>9/7/96: Floating Ideas. "Search engines take too long. By the time >they index a story, I'm not interested it anymore. I want a 'Just >In Time' search engine. One which indexes the top fifty news sites >every thirty minutes instead of every sixty days."

The aim at Moreover has always been to move towards this. We have implemented a search feature which returns xml from the 1500 sources that we aggregate near realtime data from. At the moment this searches just headlines, but in about a week this will switch to full-text searching. We've installed XML-RPC and SOAP 1.0 modules with some sample methods to grab our content, based upon search terms. We have tested the XML-RPC with a call from within Java and PERL at the remote end, and the SOAP with a call from within PERL. This should go live tomorrow, so I'll post a URL. I'd be interested in any feedback, since some of the SOAP stuff seems embryonic (or we've been going to the wrong places). For example, IBM's SOAP implementation doesn't seem to like talking to Ken MacLeod's, and the PERL version we are using doesn't return an array which limits its usefulness.

Hopefully this is a first step towards a JIT-SE

Dave

David Galbraith Moreover.com


There are responses to this message:


This page was archived on 6/13/2001; 4:55:13 PM.

© Copyright 1998-2001 UserLand Software, Inc.