Here’s how I see it:

The h-formats (hReview, hCalendar etc) have the advantage that they can go anywhere HTML goes — perhaps on a web page, or inside an RSS ‘description’ element.

Their disadvantage is that if you don’t know about the specific format used, you won’t even be able to realise that there is metadata attached to the review (or whatever).

This seems more like the problem with REST interfaces – you need to write a new parser each time.

My favourite thing about XML-RPC is that you can use the same code to send and receive calls every time, and the only thing that changes is what the arguments mean.

Perhaps an XML-RPC style approach to reviews would be more like Russell Beattie’s recent suggestion, or like RSS-Data, which we talked about last year sometime, I think.

That way an aggregator could just look for certain standard tags, and then pass them off to a separate parser for review, rather than having to recognise certain “hidden” elements (class attributes etc).

XML-RPC : REST :: Microformats : RDF/XML

On the one hand, there’s REST– a nice, clean, standards pure way to do things. And there’s RDF, a nice, clean, spec pure way to do things.

On the other hand, there’s XML-RPC– a workable hack that shoehorns remote procedure calls across a channel not quite meant for that purpose. And there’re Microformats, workable hacks shoehorning data into a channel not quite meant for that purpose.

Except that, well, microformats are kinda in the middle, since there is a lot of effort made to express things in semantic HTML, rather than outright subverting the carrier wave.

Re. Phillip’s point: “Their disadvantage is that if you don’t know about the specific format used, you won’t even be able to realise that there is metadata attached to the review (or whatever).”.

It’s not as clear as it could be in the microformat docs, but these things can be decribed as profiles of XHTML, which can be indicated by a profile attribute (containing one or more URIs) in the head element of the doc. This is a huge plus. On the one hand it means the microformat data is globally identified as such, and potentially discoverable using existing tools (like Google). It also opens the way to use GRDDL (XSLT) to extract unambiguous RDF/XML from the XHTML. Taken as a whole, this means you don’t need to know anything about the format in advance for it to be useful (in say a queryable RDF store).

btw, I’m in the process of a little proof-of-concept with microformat+GRDDL, working with DOAP (Description of a Project) data:
http://purl.org/stuff/hdoap/profile
so far, so good…