W3C home > Mailing lists > Public > public-rif-wg@w3.org > April 2010

Re: another question about imports

From: Stella Mitchell <stellamit@gmail.com>
Date: Sun, 18 Apr 2010 10:37:14 -0400
Message-ID: <m2hd64b0f2c1004180737w5521e3av167a70be71ab7432@mail.gmail.com>
To: Sandro Hawke <sandro@w3.org>, RIF <public-rif-wg@w3.org>
 Ok, and how about if in the test document we continue to say, as we
currently do, that there is one normative version (and the other versions
are presentation) of an InputDocument, PremiseDocument and
ImportedDocument?  For an imported RDF document it would be the RDF/XML
version. Or do you want to have both the xml and turtle versions normative
and change the xml schema for the test cases?

Stella


On Sun, Apr 18, 2010 at 9:15 AM, Sandro Hawke <sandro@w3.org> wrote:

> [feel free to quote this in public e-mail]
>
> > (I'm not making any changes to the test cases or documentation yet, just
> > preparing)
> >
> > If the location of the imported file is ...import001, how will the
> > requesting software specify whether it wants the .rdf version or the .ttl
> > version?
>
> We probably need to explain this in the test cases document.
>
> When the access is over the web, they do it via an HTTP "Accept" header,
> using a process known as "Content Negotiation", which shifts the burden
> of translation between formats over to the server.
>
> Specifically, when a client does an HTTP GET operation, there is a
> parameter in which one may specify the acceptable content types to be
> returned.   For example:
>
>  GET -H 'Accept: application/rdf+xml'
> http://www.w3.org/ns/entailment/Simple
>  GET -H 'Accept: text/turtle' http://www.w3.org/ns/entailment/Simple
>  GET -H 'Accept: text/html' http://www.w3.org/ns/entailment/Simple
>
> each one works, and returns different bytes representing what is
> conceptually the same information in different formats.  With a little
> configuration, Apache does this automatically if you just provide files
> with the same base name and different suffixes.
>
> I don't think we need to make a big point of this -- any RDF library
> which can fetch from the web (which is probably all of them) almost
> certainly has this already implemented.  But yes, we should state this
> somewhere in the Test Cases document.
>
> Also: when running locally, without an HTTP server, there has to be some
> similar step.  When the URI is turned into a filename, some logic is
> needed to related the desired format with some filename suffix.
>
>     -- Sandro
>
Received on Sunday, 18 April 2010 14:37:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 18 April 2010 14:37:47 GMT