Re: I guess it's a stupid questions.

Danny Ayers wrote:
> If the metadata is about the document, then you don't have a problem,
> changing rdf:about="" to rdf:about="http://uri-of-the-document" will
> preserve the semantics.

The problem I see is that after loading several documents into a system 
that is not context aware you can no longer connect data about a 
document to data from a document. It's no good to know that data from 
some document is licensed under certain terms, when I no longer know 
what data came from this document.


> Assuming you want something like this, why isn't it practical? (I'm
> sure I've seen discussion somewhere around implementing reification
> efficiently, to minimise the overhead).

Even if reification is implemented efficiently the fact that the number 
of logical triples in the system is multiplied by several is unlikely to 
not have a significant performance impact.

Also, the official serialization syntax doesn't allow you to reify each 
and every statement in a document without making a complete mess.


> "'a something b' in documentX"
> "documentX modified 2004-11-15" 
> => "'a something b'  modified 2004-11-15"

I am indeed making this inference, as a workaround, lacking any other 
usable mechanism for providing context information. Apart from the fact 
that not everyone else may agree with this inference, the solution is 
not very satisfactory as it allows only one context per file.


> In which case, what's the harm in expressing the information slightly
> differently, such as in one of the patterns in the 'N-ary Relations'
> doc [1]?

Not sure how any of the solution shown on this page would be any better 
than reification. Also, I feel I shouldn't have to remodel my data in 
order to be able to attach some metadata to it...


> But it does raise the question - if RDF stores can implement contexts,
> does it matter that the RDF model doesn't as such support them? (I
> dunno, like I said named graphs sounds appealing).

There may indeed not be any need for the data model to be extended. On 
the other hand it would not be a bad idea if the concept of having a 
context was well defined. While most stores meanwhile support some kind 
of context, the big issue for me is that the official serialization 
syntax doesn't!

I am aware that there are a few unofficial alternatives to the official 
syntax, but they have the drawback that they are (1) completely 
ununderstandable to people not familiar with the technology and (2) 
another order of magnitude more bloated than the current syntax.

Received on Monday, 15 November 2004 15:37:34 UTC