Re: Semantic Web User Agent Conformance

Sean B. Palmer wrote:
> You have the same confusion as Alan, which I think was brought on by
> my use of the term "User Agent". This is part of a User Agent only,
> the very low level access part. Sort of like an HTTP module in
> Firefox--it's just a single and very primitive part of the User Agent
> I'm describing. Nothing to do with actually using the RDF at all, it's
> the step before that: getting the RDF.

Ok, got the picture. Reading your first email again with a fresh mind I 
have a few questions:


First, you say: "...if there were some well defined heuristics for 
telling which possible transformations might apply, that would greatly 
reduce the burden on the suite of parsers required to handle all this 
[testing all possible mechanisms]."

I agree with you that "<link rel="stylesheet" type="application/rdf+xml" 
href="style.rdf" />" and similar cases is the best solution for embedded 
RDF, as it would have a content-type. No need for heuristics in this case.


Second, if you want to build a generic semantic engine (to handle live 
data such as embedded RDF in webpages the browser is opening right now), 
you will need to, at least, read the data.

If you use the browser's DOM tree you probably already have it stored 
and semi-parsed and if you had the content-type in the first place you 
need no heuristics to parse it.

Of course, not always you'll have the right content-type or some at all, 
so following the most probable to least, throwing exceptions as you fail 
to parse and trying again with the next most probable seems much simpler 
and easier than trying to build some heuristics about the data, storing 
and keeping it up-to-date in order to save a few exceptions to be thrown.



Unless, of course, I got it all wrong again and have to read your email 
once more... ;)

cheers,
--renato

Received on Thursday, 22 November 2007 22:11:17 UTC