Re: A modest attempt to re-open ISSUE-20

I agree with you all that there is nothing wrong with the *document* that,
according to the current LDP spec, the client has to POST to the LDPC. It
*does* represent a graph, as the LDPC will provide a base URI.

My problem is about the graph it represents, since the client can not know
that graph. So in order to produce that document (serializing said graph),
it has either to :

* produce the serialization directly, without relying on a graph API, which
is a PITA in the general case ;
* produce another graph with a dummy URI, and hope for the best when
serializing it with that dummy URI as the base URI (i.e., hope that the
serializer will indeed replace all mentions of that URI with the null
relative URI - I'm not sure the serializers I'm using guarantee that...) --
this is an ugly hack ;
* use a library that implements more than the RDF spec, allowing relative
URIs in nodes.

The last option is probably the most reasonable, as RDF libraries (at least
two that I know) seem to do that.
But I really don't feel comfortable with this WG producing normative text
relying (even if implicitly) on non-standard extensions of RDF.

  pa

On Wed, Mar 13, 2013 at 12:06 PM, Kingsley Idehen <kidehen@openlinksw.com>wrote:

>  On 3/13/13 6:54 AM, Henry Story wrote:
>
> On 13 Mar 2013, at 10:40, Andy Seaborne <andy.seaborne@epimorphics.com> <andy.seaborne@epimorphics.com> wrote:
>
>
>  On 13/03/13 07:10, Pierre-Antoine Champin wrote:
>
>  Henry,
>
> On Wed, Mar 13, 2013 at 2:15 AM, Henry Story <henry.story@bblfish.net
> <mailto:henry.story@bblfish.net> <henry.story@bblfish.net>> wrote:
>
>      The abstract syntax  specificiation allows for relative URLs:
>    http://www.w3.org/TR/rdf11-concepts/#rdf-documents
>
>
> This section is about serialization; it explicitly says "concrete syntaxes".
> On the other hand, the definition of IRI for the graph model
> http://www.w3.org/TR/rdf11-concepts/#dfn-iri
>
> explicitly says "IRIs in the RDF abstract syntax MUST be absolute".
>
>  Yes.
>
> A syntax may allow a relative URI but that's in a document and a document has a base URI.  The relative URI is relative to some base URI.  Relative URI have a role in syntax
>
> RFC 3986 makes it clear:
>
> [[
> 5.1.  Establishing a Base URI
>
>   The term "relative" implies that a "base URI" exists against which
>   the relative reference is applied.  Aside from fragment-only
>   references (Section 4.4), relative references are only usable when a
>   base URI is known.
> ]]
>
>  And indeed they are: when you POST content the server will know what
> URIs the relative ones are referring to, once he has created the resource.
>
> Just a little further down in the spec one reads
>
> [[
> 5.1.4.  Default Base URI
>
>
>    If none of the conditions described above apply, then the base URI is
>    defined by the context of the application.  As this definition is
>    necessarily application-dependent, failing to define a base URI by
>    using one of the other methods may result in the same content being
>    interpreted differently by different types of applications.
>
>    A sender of a representation containing relative references is
>    responsible for ensuring that a base URI for those references can be
>    established.  Aside from fragment-only references, relative
>    references can only be used reliably in situations where the base URI
>    is well defined.
> ]]
>
> So when the client POSTs to an LDPC the client ensures that the base
> URI for the resources can be established: it is posting to an LDPC, and
> the contract is that the LDPC when POSTed to creates a new resource with a
> Location: header.
>
> This all works fine currently.
>
>
> +1
>



>
> Links:
>
> 1. http://lists.w3.org/Archives/Public/public-lod/2013Mar/0042.html --
> thread from LOD list about ORCID and Linked Data deployment (demonstrates
> the value of relative URLs)
> 2. http://bit.ly/MYLsMu -- DIY Linked Data Deployment via Dropbox .
> 3. http://bit.ly/MgSz3F -- Ditto via Amazon S3 bucket .
>
>   	Andy
>
>
>  Social Web Architecthttp://bblfish.net/
>
>
>
> --
>
> Regards,
>
> Kingsley Idehen	
> Founder & CEO
> OpenLink Software
> 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
>
>
>
>