Re: HTTP in RDF

Hi Nick

Nick Kew wrote:

> Some comments on the current note:
> 
> 
> 2.1 The Request class
> 
> I don't like enumerating request methods.  Make the method a property?

I think Jim likes it :-)

> 2.1.1 The Request properties
> 
> * Missing Query String

Right, we found that, too. It will be added to the "URI in RDF" if the 
group thinks, we should write that, or the "HTTP in RDF".

> * absPath is IMHO misleading terminology.  Just Path?

Yep

> 2.3 The header property
> (and later discussion of extensions)
> 
> What do we gain by enumerating headers?
> 
> RFC2616 has to enumerate them because it specifies them.
> But it's extensible, as noted later.  Instead of enumerating
> extensions (which falls down on future or private extensions,
> both of which are permitted by HTTP), we should adopt the
> same extensibility as HTTP.

Same as with request methods: I started with your approach, but was told 
that it is far easier to query the RDF for a specific header if there is 
a specific property.

> There are other issues, including headers concerned with
> transmission only (not with a document), and those which
> it might be a security issue to store.  That's no reason to
> exclude them, but it needs some explanation.  I guess I
> could volunteer to write something there.
> 
> 2.4 The body property
> 
> Binary bodies may need to be suitably encoded!

That's why we have the bodyEncoding property.

> Even text bodies may contain "]]>", particularly
> when the text is itself markup!

Do you have a solution?

-- 
Johannes Koch - Competence Center BIKA
Fraunhofer Institute for Applied Information Technology (FIT.LIFE)
Schloss Birlinghoven, D-53757 Sankt Augustin, Germany
Phone: +49-2241-142628

Received on Wednesday, 5 April 2006 07:57:10 UTC