Persistent Search and OpenSearch
April 10th, 2006 by DeWitt Clinton

Russell Beattie (nice redesign, btw!) writes about Yahoo! and Persistent Search. He was inspired by Bill Burnham’s ideas about ongoing searches. To use Bill’s definition, “Persistent Search allows users to enter a search query just once and then receive constant, near real-time, automatic updates whenever new content that meets their search criteria is published on the web.”

Russell makes the conceptual leap that RSS-enabled search results + Feed Alerts (SMS, IM, etc.) gets you most of the way there. He then points out that Yahoo’s search APIs already give you everything you need and even provides a working example for us to learn from.

I just wanted to add that OpenSearch, which formally layers search capabilities on top of syndication formats such as RSS and Atom, ties this all together very nicely. In fact, adding full OpenSearch support to the Yahoo! RSS Search feeds would involve little more than creating a trivial OpenSearch Description document and inserting the following OpenSearch Response elements into the RSS results:

<opensearch:totalResults>4230000</opensearch:totalResults>
<opensearch:startIndex>1</opensearch:startIndex>
<opensearch:itemsPerPage>10</opensearch:itemsPerPage>

Around this time last year I outlined how one could use OpenSearch to create “persistent search” with a standard RSS reader. My example uses Amazon Web Services APIs, but you could do this with just about any searchable data source.

My guess is that Yahoo!, or any of the other major search providers, could add full OpenSearch support to their APIs in an hour or two. (I say this somewhat knowingly, as I’ve built personal OpenSearch wrappers around all of them in that amount of time.) And if unmetered access is a concern then an OpenSearch Query extension could be added to reuse the same “application ID” limits that each engine’s public API already supports. And with Internet Explorer 7 supporting OpenSearch natively this would provide syndication support directly to tens of millions of desktop clients. (Firefox 2.0 will reportedly support OpenSearch as well…)

Any bets on which major web search engine will adopt OpenSearch first? Google, Yahoo!, MSN, or Ask? I’ll even include Ebay in the challenge, as Ebay now supports RSS search results as well…

PS: Does anyone know if Ask has a public search API? Their algorithmic search results keep getting better and I’d love to be able to code to them the way I can to Google, Yahoo!, and MSN. (By way of arbitrary comparison, check out a search for “how to dress a fish” on Google, Yahoo!, MSN, and Ask. More proof that the search battles have just begun.)

2 Responses to “Persistent Search and OpenSearch”

  1. MikeD Says:

    There was also an article on these concepts posted on XML.com in 2004 - http://www.xml.com/pub/a/2004/12/01/pubsub.html

    I guess it just takes a while for a good idea to take off…

  2. DeWitt Clinton Says:

    Mike — great point about the considerable prior thought that has gone into the idea of persistent search. And I bet if we dig deeper we will see these themes reoccurring for years. It does take a while for a good idea to take off, and it is both frustrating and strangely satisfying to see someone post about something that one has already written about, thought about, or started a company to do years before… I seem to recall a mentor of mine calling that “validation.” : )