W3C home > Mailing lists > Public > public-lod@w3.org > February 2013

Re: Content negotiation for Turtle files

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Wed, 6 Feb 2013 10:54:02 +0100
Message-ID: <CAK4ZFVFxDOg266kFtePux+ehws74xk+2de-yxy=78HuYeMwMhg@mail.gmail.com>
To: public-lod@w3.org
Thanks all for your precious help!

... which takes me back to my first options, the ones I had set before
looking at Vapour results which misled me - more below.

AddType  text/turtle;charset=utf-8   .ttl
AddType  application/rdf+xml    .rdf

Plus Rewrite for html etc.

I now get this on cURL

curl -IL http://www.lingvoj.org/ontology
HTTP/1.1 303 See Other
Date: Wed, 06 Feb 2013 09:28:45 GMT
Server: Apache
Location: http://www.lingvoj.org/ontology_v2.0.ttl
Content-Type: text/html; charset=iso-8859-1

HTTP/1.1 200 OK
Date: Wed, 06 Feb 2013 09:28:45 GMT
Server: Apache
Last-Modified: Wed, 06 Feb 2013 09:19:34 GMT
ETag: "60172428-5258-4d50ad316b5b2"
Accept-Ranges: bytes
Content-Length: 21080
Content-Type: text/turtle; charset=utf-8

... to which Kingsley should not frown anymore (hopefully)

But what I still don't understand is the answer of Vapour when requesting
RDF/XML :

   - 1st request while dereferencing resource URI without specifying the
   desired content type (HTTP response code should be 303 (redirect)):
   Passed
   - 2nd request while dereferencing resource URI without specifying the
   desired content type (Content type should be 'application/rdf+xml'):
   Failed
   - 2nd request while dereferencing resource URI without specifying the
   desired content type (HTTP response code should be 200): Passed

Of course this request is bound to fail somewhere since there is no RDF/XML
file, but the second bullet point is confusing : why should the content
type be 'application/rdf+xml' when the desired content type is not
specified?

And should not a "Linked Data validator" handle the case where there is no
RDF/XML file, but only Turtle or n3?

The not-so-savvy linked data publisher (me), as long as he sees something
flashinf RED in the results, thinks he has not made things right, and is
led to made blind tricks just to have everything green (such as
contradictory mime type declarations).

At least if the validator does not handle this case it should say so. The
current answer does not help adoption of Turtle, to say the least!

Hoping someone behind Vapour is lurking here and will answer :)

Thanks again for your time

Bernard
Received on Wednesday, 6 February 2013 09:54:50 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 31 March 2013 14:24:46 UTC