Fwd: Parsing Microdata into RDF Graphs: URI Comparison

(Forwarding to the HTML Data TF list.)

Martin raises some good points about URI modification required in the Microdata spec. I'll need to note this behavior in the Microdata to RDF spec, and it also needs to be considered as advice for people choosing between RDFa and Microdata.

The Microdata spec [1] references the HTML spec when talking about resolving a URL. However, it's only referenced with regards to resolving a global identifier (@itemid).

This does have some implications for HTML+RDFa, as resolving includes the use of xml:base, which is not used for XHTML+RDFa or HTML+RDFa which rely solely on the base element. One implication is that this should be considered for HTML+RDFa.

The HTML spec touches broadly on resolving URLs as part of DOM access. This has implications for both Microdata and HTML+RDFa, if access to these attributes is intended to be compatible with DOM.

RDFa uses a different URI resolution algorithm [3], using just the algorithm in RFC-3987 more broadly compatible with other RDF serializations, basically relying on the document location or base URI (XML1 and SVG do use xml:base).

The ramification of this is that the same graph serialized in RDFa and Microdata could result in different object references, due differences in resolution. Moreover, @href and @src can be used in RDFa chaining, so that these URIs because the subject of other statements. In Microdata, this would be done with @itemid, which is subject to HTML URL resolution, so at least the graph would be self-consistent, but different from the equivalent HTML+RDFa version.

My own processor doesn't perform HTML-style URL resolution, and there should probably be a test suite that verifies appropriate URL resolution.

Gregg

[1] http://www.w3.org/TR/microdata/#resolve-a-url
[2] http://lists.w3.org/Archives/Public/public-vocabs/2011Oct/0156.html
[3] http://www.w3.org/2010/02/rdfa/drafts/2011/ED-rdfa-core-20111020/#s_curieprocessing

Begin forwarded message:

> From: Martin Hepp <martin.hepp@ebusiness-unibw.org>
> Subject: Parsing Microdata into RDF Graphs: URI Comparison
> Date: October 29, 2011 2:26:03 AM PDT
> To: "semantic-web@w3.org" <semantic-web@w3.org>, "public-lod-request@w3.org" <public-lod-request@w3.org>
> 
> Dear all:
> 
> We had a longer discussion on URI comparison in RDF settings earlier this year. Note that when it comes to writing Microdata parsers, which will be a common thing for LOD architectures, there is a new spec to observe:
> 
>   http://www.whatwg.org/specs/web-apps/current-work/multipage/urls.html#resolve-a-url
> 
> AFAIK, itemtype URIs are never resolved this way, but itemprop and href URIs will.
> 
> This may result to different behaviors of queries in RDF settings if the same markup is used in Microdata vs. RDFa.
> 
> Best
> 
> Martin Hepp
> 
> 
> --------------------------------------------------------
> martin hepp
> e-business & web science research group
> universitaet der bundeswehr muenchen
> 
> e-mail:  hepp@ebusiness-unibw.org
> phone:   +49-(0)89-6004-4217
> fax:     +49-(0)89-6004-4620
> www:     http://www.unibw.de/ebusiness/ (group)
>         http://www.heppnetz.de/ (personal)
> skype:   mfhepp 
> twitter: mfhepp
> 
> Check out GoodRelations for E-Commerce on the Web of Linked Data!
> =================================================================
> * Project Main Page: http://purl.org/goodrelations/

Received on Saturday, 29 October 2011 20:34:04 UTC