[Matroska-devel] Tags & RDF

Steve Lhomme steve.lhomme at free.fr
Tue Jun 8 13:52:02 CEST 2004


As some one you may have noticed in the ML, I want to work on a project 
called CoreSync/CoreLib/MDDLib. For more info on CoreLib/CoreSync please 
have a look at my page : http://mukoli.free.fr/corelib/

The idea is to be able to sync folders/devices and also have a local 
DataBase of your multimedia content (first target is audio). And talking 
with the people working on MDDLib it seems that storing this DB in and 
working with RDF is the way to go.

RDF is a very simpla way to describe relation between informations. It 
always follow this logic : URI->relation->value. A way to store RDF is 
to use XML (even though that's not the easiest way to read it). The only 
important thing, apart from the relations you build, is the semantic, as 
for EBML and Matroska. And the idea behind RDF is the Semantic Web. 
Which is the idea that says that the content of the web and most 
specifically the meta informations on this content should be 
standardised. So that everyone uses the same tags, so that a computer 
could process it.

For CoreLib/MDDLib we'll want to store meta informations on various 
files in a DB. And these info will come from tags (ID3, Matroska, APE, 
whatever). So we'll probably need a mapping at some point to know which 
tag correspond to which RDF element to output (either done when storing 
in the DB or when getting data from the DB). So I suggest that in the 
tags mapping we already have we add the corresponding RDF elements (or 
actually the semantic element we want to use, because anyone is free to 
redefine his own semantic)... Most probably some tags will match the 
DublinCore semantic or the RSS semantic. When no known elements is found 
  we can create our own semantic.

Pamel can you do that ? I can help on understanding RDF, the known 
semantics and such.

More information about the Matroska-devel mailing list