Yahoo! was the first major search engine to wholeheartedly embrace the Semantic Web with SearchMonkey, which can make use of metadata embedded inside web pages using either microformats, eRDF or RDFa. The future of SearchMonkey is a little cloudy in the wake of Microsoft’s search deal with Yahoo! last year â€“ cleared in February by the Justice Department â€“ but one can’t ignore the footprint Yahoo has made on the Semantic Web. So the SemanticWeb Blog caught up with Peter Mika of Yahoo! Research â€“ who was instrumental in SearchMonkey’s development as well as that of the AB Meta RDFa-based format for annotating pages about things â€“ to get his perspective about the impact of Facebook’s Open Graph Markup, which lets publishers provide structured data about their web pages.
SemanticWeb Blog: Since you had a hand both in Search Monkey and AB Meta, you’re in the sweet spot to tell us what you think Open Graph’s impact around semantic markup will be compared to what the impact of other efforts have been.
Mika: We welcome the OpenGraph proposal lead by Facebook, which builds on the RDFa standard and therefore can be directly consumed by all tools (including search engines, but also browser extensions) supporting RDFa.
The approach they follow is similar to AB Meta in that it requires publishers to modify only the HEAD of the HTML page. This method is simple, but it has the disadvantage that the same data is duplicated within the page. As a search provider, we also prefer “contextual awareness”, i.e. the ability to tie structured information to the relevant parts of the page. Compared to AB Meta, the Open Graph proposal makes further simplifications for the convenience of publishers.
Some of the simplifications mean that the data is not ideally represented, which can only be amended by additional processing. Despite these issues (which might be addressed in a future version), we value Facebook’s effort in building out the Semantic Web by motivating publishers, not only through a simplified approach, but also by providing additional services of value, such as social recommendations.
Semantic Web Blog: So there are plenty of ways to mark up the web that rely on RDFa, including what Google is doing. Can these efforts all co-exist happily together and leverage each other in any way, or is there the potential that content providers would have to support each in isolation if at all?
Mika: With Facebook, Google and Yahoo all supporting RDFa as a syntax, there seems to be a convergence on at least the syntax of encoding metadata in HTML. The next step of convergence has to be forging agreements around vocabularies (ontologies), which unfortunately only happened in limited instances until now (a positive example is Google’s support of our video markup). Unless this happens, there is still the danger the publishers have to add the same information using multiple vocabularies in order to support different consumers of the information.
Semantic Web Blog: Is there opportunity for Yahoo! to innovate on what Facebook is doing?
Mika: As Yahoo! is already indexing RDFa data, we can easily leverage Open Graph data, for example, in displaying enhanced search results in Yahoo Search. However, we are not ready to share concrete plans at this point.
Semantic Web Blog: Last one, since we have you. Any new information on how the Microsoft-Yahoo search agreement will impact SearchMonkey?
Mika: At this point we cannot discuss the details of the implementation of the MS-Yahoo search agreement.