We polled a few experts in the Semantic Web community who have been following the news from Twitter about Annotations and today's news about semantic markup in Facebook's OpenGraph.

Here is what they have to say.

Hank Williams

"Twitter annotations are essentially unstructured data slots associated with a tweet. This is helpful as it adds the ability to send tweets that have metadata. But there are no rules for the metadata. There is no suggested or enforced syntax.

On the other hand, FB OpenGraph is fundamentally their own version of semantic markup for web pages. In fact it is based on the W3C RDFa markup standard. What they have done is added specific typing, so that there are rules for how to indicate that something is a restaurant or a sports team, or band, or whatever. Having someone as big as facebook actually set standards that will, by virtue of their size, have to be supported, is hugely helpful to making the semantic web happen.

While the use cases for annotations and FB open Graph are different, Twitter really needs to set some standards for how to use annotations, while still allowing developers to innovate outside that standard. Without that standardization, I don't see annotations being particularly useful."

Roger Macdonald

"FB's Open Graph is a striking advancement in the early implementation of the dream of a Semantic Web - a ripple in the force.

Twitter will likely stop being coy about delaying metadata standards to their emergent annotation feature and consider Linked Data.

Significant number of developers will become versed in Linked Data standards and advantages.

Linked Data plays will become more attractive to VCs
Search engines and sophisticated social graphing agents will have a field day.

"Socially-validated web" will start making more sense to more people."

Andraz Tori

"Oh, there are so many things to be told about both approaches. And I haven't yet seen real analysis of them, only media hype. The devil is in the details.

Both initiatives are exciting, they open a new layer on top of existing
services, so people don't need to create a separate infrastructure for
exchange of useful metadata (for example, who's influencer in what, more
specific location metadata, friends, etc.)

It is scary that those initiatives are under control of one single
party, but that's how innovation happens. Semantic web/federated way
hasn't worked out for rapid innovation.

Now the main part: The problem with Twitter's apporach is in the details
- they only allow adding of annotations _while_posting_a_tweet_, not
afterwards. Facebook gets that part more right, as it seems they plan to
make it possible to programmatically change the data.

(right now there's a catch: "after the page receives 10 likes, these
properties become fixed to avoid surprising users who have liked the
page already.")

Open Graph API is actually what's much more interesting than the markup
itself. It allows for general access to the data in Facebook. That's
where the real innovation is going to happen - if Facebook is going to
be liberal about the data. Twitter does not seem to offer any kind of
universal API on top of annotations.

All in all, Facebook's data and API seems more directly usable, now the
question is how open will it really be to the developers. Will you be
able to write an independent app that (with user's permission) changes
the data about his music preferences on Pandora? And does exported data
belong to the user or Facebook?

And additional note: Facebook's markup approach is actually very similar
to Google Rich Snippets and Yahoo Searchmonkey's data. Both of them use
the same markup (for example for address information), but Facebook
'invented' its own. So you can guess where this is going..."