[foaf-protocols] Problem with certificate on home-grown WebID
kidehen at openlinksw.com
Mon Dec 19 17:27:30 CET 2011
On 12/19/11 10:58 AM, Sebastian Trüg wrote:
> I created a new WebID: http://www.trueg.de/people/sebastian
> It serves rdf+xml, turtle, or html based on the accept header now. The
> linkeddata.org vaidator is happy but the lookup at webid.fcns.eu still
> does not return anything.
> foaf.me login test result attached.
Okay, I'll leave Henry to debug his system :-)
> On 12/18/2011 09:30 PM, Kingsley Idehen wrote:
>> On 12/18/11 2:33 PM, Henry Story wrote:
>>> On 18 Dec 2011, at 19:42, Kingsley Idehen wrote:
>>> I was not quite sure what question you were answering here.
>>>> Simple Linked Data trouble shooting re. Name / Address disambiguation
>>>> laced with bits of content negotiation.
>>>> -- Vapour Linked Data Utility
>>> This one does not seem to understand Turtle as it is not able to
>>> register that the primary topic is a foaf:Agent.
>> That's a Linked Data validator, it disambiguates Names, Addresses, and
>> associated resources. It's a simple tool for determining if a resource
>> is indeed a proper bearer of Linked Data i.e., name and address properly
>>>> -- URI Debugger .
>>>> When serving up a Linked Data resource, you need to consider handling
>>>> QoS algorithms as part of the HTTP request, or at the very least,
>>>> make a best
>>>> effort to guide the user agent to the resource type it seeks. The
>>>> extension ".ttl" isn't really the mime type determinant, as you know :-)
>>> I think Treug got the mime type right:
>>> $ curl -I http://www.trueg.de/clara/foaf.ttl
>>> HTTP/1.1 200 OK
>>> Date: Sun, 18 Dec 2011 19:31:26 GMT
>>> Server: Apache/2
>>> Last-Modified: Sun, 18 Dec 2011 08:56:50 GMT
>>> ETag: "43808c-4c2-4b45a03c2d880"
>>> Accept-Ranges: bytes
>>> Content-Length: 1218
>>> X-Powered-By: PleskLin
>>> Connection: close
>>> Content-Type: text/turtle
>> My point, if you actually play with the URI debugger, you'll notice it
>> requests HTML by default and Sebastian's server returned Turtle. Change
>> the request to RDF/XML and it still returned Turtle. Thus, even if the
>> client used a sophisticated QoS algorithm (i.e., attempted to leverage
>> dynamic content negotiation) it still wouldn't have worked.
>> Sebastian's URI for the Object that represents the foaf:Person Object
>> was: http://www.trueg.de/clara/foaf.ttl#me
>> There's nothing wrong with that URI from a Linked Data perspective
>> since, it resolves to a resource bearing the description of the URI
>> referent. The problem is that when requests are made for non Turtle
>> resources, his server was ignoring the specificity of those requests and
>> serving up a Turtle resource. He could easily have indicated 406 and
>> redirection to the resource types his setup is capable of offering for
>> Here is a simple example using one of my URIs:
>> -- 406 in response to an RDF/XML request for an N3 resource
>> -- What happens when you make a default request (as most user agents do
>> re. HTML, so take note of the response headers and<head/> section of
>> the HTML) .
>> Linked Data is about exploiting HTTP's dexterity. There's ultimately no
>> way around that. My response was about what I've outlined above re. how
>> to work with resources in a format agnostic way.
>>>> Kingsley Idehen
>>>> Founder& CEO
>>>> OpenLink Software
>>>> Company Web: http://www.openlinksw.com
>>>> Personal Weblog: http://www.openlinksw.com/blog/~kidehen
>>> Social Web Architect
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 1625 bytes
Desc: S/MIME Cryptographic Signature
Url : http://lists.foaf-project.org/pipermail/foaf-protocols/attachments/20111219/bdc21359/attachment.bin
More information about the foaf-protocols