Perhaps one of the most anticipated panels at next week’s Semantic Technology & Business Conference in San Francisco is the Wednesday morning session on Schema.org. Since the announcement of Schema.org just prior to last year’s SemTech Business Conference on the west coast, using the Schema.org shared vocabularies along with the microdata format to mark up web pages has been much debated, and created questions in the minds of webmasters and web search marketers along the lines of, “Which way should we go? Microdata or RDFa?”
Concern over the complexity of RDFa for the average webmaster, coupled with the hopes of realizing greater ROI from a structured markup route using microdata that has the support of the major search engines, have played their role in the debate. RDFa 1.1 Lite, which is one of the three W3C RDFa specifications now proposed as recommendations and which is on the same complexity level as microdata, along with new resources such as the launch of RDFa.info (covered here) to help web developers understand how and why to integrate RDFa, should help address the first issue. Additionally, webmasters can begin to implement the schema.org vocabulary in RDFa 1.1 Lite – which, unlike microdata, will soon become an official standard – and examples of its use for types are in-progress, so that SEO ROI needn’t be limited to those publishers who stick with microdata for schema.org.
“You can almost actually make a mechanical exchange between some properties defined in microdata to their RDFa Lite counterpart and vice verse. That’s not a coincidence,” says Ivan Herman, Semantic Web activity lead at the W3C, who will be moderating the SemTech panel. It was the goal when developing RDFa Lite to define a layer as simple and close to microdata as possible, while still enabling it to be a stepping stone for developers that ultimately want to move to full RDFa to do more complex things, he says, like mixing vocabularies.
So, says Herman, “what I hope for and expect, and we’ll see, is that when RDFa becomes a standard, that schema.org will accept RDFa on the same level as microdata.” That is, in addition to the backing schema.org has already expressed for parsing RDFa data, he’s hoping for some more visible and explicit announcements at next week’s SemTechBiz event.
All that said, the two worlds are sure to co-exist for the forseeable future – BuiltWith Trends reports 82,013 sites using microdata. And Yahoo! has said that 7 percent of all web sites contain microdata, in most cases used exclusively with schema.org, Herman says, so “it is here to stay.” Those that are still pondering choices or open to change may want to check out this post by Jeni Tennison of about when it makes more sense to go with microdata and when to choose RDFa.
Vocabulary Over Syntax
Enough about syntax, then? For Herman, the more interesting question he hopes next week’s panel will ponder is about the evolution of the schema.org vocabulary. “How they want to use the expertise of the community in various areas, what areas do they want to cover and where do they want to say this is a special expertise and let another community develop things – that whole mechanism is really interesting,” he says.
rNews, the IPTC-developed standard for using semantic markup to annotate news-specific metadata, is one example of a community whose work was adopted by schema.org, and now the GoodRelations web vocabulary for e-commerce is on its way to be integrated, too. Mapping Dublin Core terms to schema.org elements is also a Task Group project now, and there’s also discussions on connecting the health/medical extension to schema.org with work being done by the Semantic Web Health Care and Life Sciences Interest Group at the W3C.
“I think structured data is good for the web in whatever format,” says Jay Myers, web architect at Best Buy. Schema.org has helped make structured data markup real to people outside the semantic web community, he thinks, but he does question whether there are new objects to be explored on the web that haven’t already been defined by current vocabularies. “The great thing about RDFa is it’s just a syntax and you are using vocabularies that have been worked on for years,” he says.