Raw

Jon Udell's looking at XMP, has an example as used by Vista Photo Gallery. For background, Mike Linksvayer summarises neatly:

XMP, which is basically a constrained RDF/XML serialization following super-ugly conventions that may be embedded in a number of file formats (most prominently PDF and JPEG, but potentially almost anything)...

There's loads of the stuff around - e.g. every Adobe tool embeds some RDF in their output. 

(Like Mike I don't really understand Jon's references to microformats - I first assumed he meant XMP could be replaced with a uF)

There's already been one metadata problem reported re. Photo Gallery (proprietary stuff may get squidged), it looks to me like there's another one. I don't think it's likely to cause immediate problems, but could prove an irritating microspanner in the works for automated systems later on.

Here's a snip from Jon's example:

<rdf:Description
xmlns:MicrosoftPhoto="http://ns.microsoft.com/photo/1.0">


 
<MicrosoftPhoto:Rating>1</MicrosoftPhoto:Rating>


 </rdf:Description>

The modelling in the full example is questionable (like Mozilla's RDF, XMP appeared fairly early on before anyone had a clue about best practices), but in isolation this is fair enough - " something has a Rating of '1'" or in Turtle (N3):

[  MicrosoftPhoto:Rating "1" ]

or in NTriples (which is also valid Turtle):

_:something  <http://ns.microsoft.com/photo/1.0Rating> "1".

1.0Rating? Oops.

Problem is, RDF/XML goes further than most XML languages, making a URI for each term by concatenating the namespace name and the term. This is seriously elegant IMHO, and works a treat, as long as the namespace name ends in a "/" or "#" (see HashVsSlash). Which Microsoft's doesn't. Their RDF seems a bit cargo cult.

I can't remember offhand what the specs have to say in these circumstances. I'm fairly sure the RDF/XML is legal, and most if not all RDF APIs will be able to read the stuff and use it. But I'm betting that very few will be able to serialise it to anything other than RDF/XML and read it back in again without breakage.

btw, Atom's namespace is http://www.w3.org/2005/Atom - I've no idea how that came about, but as Atom isn't RDF/XML it shouldn't be a problem. It may even make the politics a little smoother for an RDF mapping of Atom's model, as http://www.w3.org/2005/Atom# is a different URI.

 

 

@en