I've seen some blogs etc. about this new technology: RFDa, but is there any benefits in using it now. From what I understand, it's mainly used by search engines and other crawlers, but have Google, Bing and others implemented it yet? Will give a better rank on their pages if I use it, or will it just make my results more data enriched?
Looking at the overview, one thing that stands out is that RDFA is only applicable to XHTML 1.1 as far as validation goes, whereas there is an enormous push to finalize HTML5/CSS3 and get it out in production.
Quoting the overview:
We can assume that HTML5 will be similar in that RDFA will not validate, but probably won't break the document in known and popular browsers. That scares me a little when considering mobile devices.
I don't think its going to hurt you, but its no substitute for ensuring that the semantics of your document are well structured (i.e. using heading tags appropriately), which usually results in the proper text/headings showing up in the results.
The area where I see it most useful is describing the relation of a link to its concept, such as the 'license' example given in the overview. It is quite conceivable that someone wants results only offered under a CC-BY-SA license, or perhaps another variant.
I don't think I'd want to assign a heading to be anything but that, especially when meta tags exist to state the author / etc. The typical:
... offers search engines a very easy way to understand the outline of any document. So, as shown in the example, re-mapping 'h3' to indicate the author might be silly, but associating a link relation as 'author-of-name-of-article' might be interesting, similar to the 'license' concept.
I don't know anyone who has actually implemented it, so I'm not sure of success in either scenario. However, I don't think its going to (significantly) help your rank any more than a well structured document. The novelty, at least for me is the ability to find very specific things, such as all documents authored by "Joe Q Smith", for instance.
What I would not do is go re-defining the weight any given tag should carry when determining the structure of a document.
And, I can't resist - this is yet one more reason why people should avoid parsing HTML with regex :)
The benefits are numerous; but might not fit the same problems you are solving.
Take for instance the simple use case of an online Store which sells Music. It'll generate a bunch of pages for its products, typically one feature page per product.
Using RDFa and a few vocabularies you can describe that your page is selling a Track for $x amount; that the track is the same thing as an entry on musicbrainz or last.fm; or it's by the same artist as a particular myspace band page.
At the first level; search engines understand the good relations vocabulary - products, prices, etc. The next level (and this is where the value is; but doesn't yet have enough implementors) - you have an instant RESTful API for anyone with an RDFa parser and a SPARQL database.
There are certainly benefits to using RDFa and other semantic markup tools on certain types of website.
Google has recently introduced something called "Rich Snippets", which allows webmasters to mark up things like recipes, contacts, events, reviews and more, and have them show up with special formatting in the SERPs. Rich Snippets supports RDFa and two other methods of marking up HTML.
While it does not support RDFa, Facebook is also using semantic markup in its Open Graph Protocol.
Although the semantic web is not yet fully-baked, there are definitely situations where it makes sense to take advantage of it.