Does anyone here in the Atom world have thoughts or opinions about Yahoo's SearchMonkey and related specs, esp. DataRSS [1]? SearchMonkey is obviously aimed at improving search for producers and consumers, but the approach to adding metadata to resources has wide applicability. Not that it's the be-all and end-all, but rather another interesting point of triangulation on the problem(s) that we are all trying to address.
I keep seeing DataRSS as a more generic MediaRSS (also from Yahoo), although I am not sure that's an accurate take. It also looks like a way to use RDF (by way of RDFa mechanisms) with an Atom feed. I also see it as a pretty clean and easily understandable way to add arbitrary metadata to a link (something which I feel like we need, and which has been discussed here many times before). Essentially, you create an "adjunct" extension that live in an atom:entry. It's repeatable, in that multiple parties can make an assertion about a particular resource. By the way, the docs talk about DataRSS as a "specification for conveying structured data for URLs," which in the context of SearchMonkey mean that when a particular URL is turned up in a search result, this adjunct is ready in the wings to provide the metadata for a nicely presented Google OneBox style summary. Within the adjunct, you can have "meta" elements (key-value pairs describing the adjunct) and/or "items" (think atom:link) which in turn can have meta elements. Interestingly, items can contain other items, which makes for a powerful, if potentially unwieldy structure. Perhaps to deal with some of that potential explosion of complexity and non-standard approaches, they offer both a recommended set of vocabularies to use with DataRSS [2] and a set of sample DataRSS for various common content types [3]. --peter keane [1] http://developer.yahoo.com/searchmonkey/smguide/understand_datarss.html [2] http://developer.yahoo.com/searchmonkey/smguide/profile_vocab.html [3] http://developer.yahoo.com/searchmonkey/smguide/datarss-examples.html
