Re: Categorisation, database searches and RSS format results

["David Davies" <d.a.davies@bham.ac.uk>]

--- In syndication@egroups.com, Aaron Swartz <aswartz@s...> 
wrote:

> Question: Does your searcher only search the RSS metadata, 
> or does it search the whole content of the item.

Because it's a search engine it's searching the data itself. No 
RSS file exists until the search results are generated.

In my aggregator in Frontier, which Dave kindly (and bravely) 
released the source for a few days ago, I have a customised 
aggregator that doesn't fetch data until asked to do so in 
response to a search query, unlike the convential aggregator that 
fetches data every so often and waits for you to look at it.

My project for tomorrow is to have the aggregator search across 
databases on different servers. I'll post the results of that and 
other things on the http://aggregator.weblogs.com/ site as it's 
probably more relevant there.

Depending upon the nature of the data, you could prepare a full 
RSS file, store that on a server then search just the metadata in 
the acculumated RSS files along the lines you suggested. For 
the applications I have, and I suspect other's too, this would 
miss much of the possibilities of full-text searches (like Ben's 
post earlier http://www.egroups.com/message/syndication/629 - 
BTW, keep it simple, create a site searcher that reads META 
page tags for the page description and use that plus the page 
title and URL to construct an RSS web site searcher. It won't 
work for some but will for others, depends upon how careful you 
are when making your HTML pages - I'll show you how I got that 
working on my site, too, if anyone's interested).

Cheers,

David.