Zeno's Paradox
October 27, 2003 - New Oyez Search Feature

I've been working on the SpokenWordQuery interface to Oyez lately. For the NSF / JISC demo, I created a Flash interface that translates a series of pulldowns and queries into a SpokenWordQuery document that would be shipped to the server and the results displayed in a visually appealing manner. Today, I got around to putting query caching in the implementation, which effectively opens up the searches to a much larger base of applications.

If you do the default search, you are given an RSS document that is a list of the hits. Until today, you had to submit a SpokenWordQuery (SWQ) document to obtain the RSS list. This is great for searching, but limited outside any software that cannot make SWQ requests. If you do a search now, the link element of the channel will contain a URL that points to the cached results. This cached URL opens up all sorts of possibilities.

The first use is that it is now possible to create an arbitrarily complex SWQ document that allows users to specify queries as complex as possible:

Give me all cases between 1964 and 1978 that deal with freedom of speech but not equal protection with Justice Warren as a participant and the descriptions contain the phrase "gobbledygook".

The user would build this query with some tool -- either the Flash tool, or by submitting raw SWQ via the textfield in the link above, or something else and they get back the document with a query URL. If the user is a NetNewsWire (or other RSS aggregator user), he can subscribe to the search query as if it were any other feed. If the user is the maintainer of a website that uses a content-management backend such as Scoop or Slash, the user can make the feed a channel in the web site. This opens up a whole world of content crossover that thus far has been previously unavailable.

When things like search proxies are deployed (these would take a search and execute that search on child archives), the system becomes that much more useful. (Imagine using a search proxy that would search and return hits from all major historical archives.)

In any case, by adopting a standard such as RSS, much of the interoperability work has been done, and there's a readily available base of clients that can immediately use the hits.

For those curious, fire up your RSS aggregator and subscribe to any of the following feeds:

http://www.oyez.org/SpokenWordQuery/cache.jsp?id=10 - All abortion cases that have MP3 resources
http://www.oyez.org/SpokenWordQuery/cache.jsp?id=11 - All copyright cases after 1980
http://www.oyez.org/SpokenWordQuery/cache.jsp?id=12 - All cases from the 2000 term

The idea by now should be apparent.

Posted by br284 at 11:09 PM

Trackbacks

Comments

Post a comment
Name:


Email Address:


URL:


Comments:




Note that by posting a comment on this site using this form or any other software tool, you have agreed to the advertising terms of this site and you agree to submit payment for any advertising in the comment above as outlined in the advertising terms. This applies to ads posted in the URL or Comment fields above.