Re: datasets mean their default graphs?

Andy Seaborne <andy@apache.org> wrote:

>On 22/06/13 12:02, Sandro Hawke wrote:
>> On 06/22/2013 03:11 AM, Peter Patel-Schneider wrote:
>>> I don't see why this is necessary.  Isn't a semantics that gives
>some
>>> meaning to RDF datasets a semantic extension of one that doesn't
>>> bother to give meaning to RDF datasets?  It's not, after all, that
>the
>>> RDF semantics prohibits RDF datasets from having meaning.
>>>
>>> My worry is that there are people who make the default graph be the
>>> union of the contents of the named graphs, and thus wouldn't want
>even
>>> this minimal semantics.
>>
>> It seems me that everyone who makes the default graph be the union
>would
>> fall into one of two camps:  (1) they are comfortable with these
>> semantics, because they consider all their named graphs asserted
>anyway,
>> or (2) they don't consider their overall dataset to be asserted.    I
>> cringe a little at (2), since it seems like a good practice to only
>> publish on the open web things you assert, but it's how things
>already
>> are with people publishing RDF Graphs -- people do publish RDF graphs
>> they don't mean -- and the situation doesn't seem to be made much
>worse
>> by adding the default graphs of datasets into that set.
>>
>>        -- Sandro
>
>Isn't it more likely that they don't publish the default graph at all? 
>If the dft graph is the union, just publish the NGs.
>

Sure, but if they don't publish the default graph at all, then there's no worry, is there?

    - Sandro

> Andy

-- 
Sent from my Android phone with K-9 Mail. Please excuse my brevity.

Received on Saturday, 22 June 2013 15:02:50 UTC