[foaf-protocols] Debugging results

[Nathan]

Henry Story wrote:
> Social Web Architect
> http://bblfish.net/
> 
> On 1 Aug 2010, at 23:50, Reto Bachmann-Gmür wrote:
> 
>> On Sun, Aug 1, 2010 at 11:45 PM, Henry Story <henry.story at gmail.com> wrote:
>>> On 1 Aug 2010, at 23:42, Reto Bachmann-Gmür wrote:
>>>
>>>> On Sun, Aug 1, 2010 at 11:32 PM, Henry Story <henry.story at gmail.com> wrote:
>>>>
>>>>>> https://farewellutopia.com/user/me/profile#me
>>>>>> returns: doesn't return any RDF, rather it returns standard XHTML -
>>>>>> accept sent is: application/rdf+xml, text/rdf+n3,
>>>>>> application/rdf+turtle, application/x-turtle, application/turtle,
>>>>>> application/xml, */*
>>>>>> note: can you check it out reto, if you're not already aware
>>>>> yes there does not seem to be any rdfa in there
>>>>>
>>>>> $ rapper -i rdfa  https://farewellutopia.com/user/me/profile  -o turtle
>>>>>
>>>>> returns an empty doc
>>>> which should be ok as long as content-negotiation works correctly
>>>> (which i'm fixing now, currently it chooses of the most concrete
>>>> acceptetd types with the highest q-value the one who comes first
>>>> alphabetically, in this case application/turtle but as this isn't
>>>> supported it falls back to xhtml).
>>>>
>>> Well I checked the html, and I don't think it contains rdfa either.
>> no the html doesn't contain rdfa, but rdf/xml, turtle, rdf/json and
>> -n-triples representations are available.
> 
> that's an intersting case. It suggests a few questions:
> 
> Should the html at least contain a link to the rdf/xml in that case? Which could then be followed?

can't see any reason why 'should' - that would presume that machines can 
read + understand HTML and get the link then follow it - certainly in 
this case there would be exactly zero benefit. Better to just conneg 
through to the alternative formats if they are available.

> Should every representation for a WebID be semanticizable? (Does not quite seem like it should, and in any case they can't all be known)

yes, if it isn't then it's not a webid, id's just a URI, or worse, an 
URL - debatable?

> What are the good reasons for having html be a lower q value than rdf/xml? Is it simply that rdf/xml can only be rdf, whereas html may or may not be machine processable? (In which case any other machine processable format would have higher value)

yes, almost begs that RDFa have it's own mediatype I guess, because one 
may prefer RDFa, RDF/XML, N3 in that order with no 'pure HTML' at all - 
as in,if you don't have RDFa os RDF something, don't give me anything.

Nathan