W3C home > Mailing lists > Public > public-rdf-comments@w3.org > May 2012

Re: Use of xml:base in rdf:HTML Datatype

From: Gregg Kellogg <gregg@greggkellogg.net>
Date: Mon, 21 May 2012 10:45:42 -0400
To: Richard Cyganiak <richard@cyganiak.de>
CC: "public-rdf-comments@w3.org" <public-rdf-comments@w3.org>
Message-ID: <D5DD55F9-92CE-4A04-9B32-8A44E77D67B0@greggkellogg.net>
On May 21, 2012, at 3:49 AM, "Richard Cyganiak" <richard@cyganiak.de> wrote:

> Hi Gregg,
> On 18 May 2012, at 23:25, Gregg Kellogg wrote:
>> The editor's draft of RDF Concepts section on rdf:HTML Datatype [1] makes a specific callout about @lang needing to be included explicitly in the HTML literal. I presume that this reasoning also applies to any in-scope @xml:base (valid in XHTML5). So, if we consider the following:
>> <div property="rdf:value" datatype="rdf:HTML" xml:base="http://example.com/foo">
>> Interesting topic located <a href="bar">here</a>.
>> </div>
>> The xml:base context in effect during processing would _not_ be retained in the literal. This is probably worth a similar note in the concepts document.
> I've updated the note, it now reads:
> [[
> Any language annotation (lang="…") or XML namespaces (xmlns) desired in the HTML content must be included explicitly in the HTML literal. Relative URLs in attributes such as href do not have a well-defined base URL and are best avoided.
> ]]
> http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html#note-html-context

> Is this better?

Yes thanks, I think making the bit about relative URLs in @href makes this clear.

> (As Gavin noted, xml:base is only allowed in XHTML5 document but not in HTML5 documents, therefore mentioning it explicitly would be confusing IMO. I don't think there's anything wrong as such with using @datatype="rdf:HTML" in XHTML5+RDFa — it basically means that the fragment has to be valid XHTML5, but will be treated as HTML5.)
> Note that RDFa *could* define additional rules when parsing the contents of elements that have @datatype="rdf:HTML", such as absolutizing all URLs, or copying XML namespaces from the context onto the elements that need them, before generating the HTML literal. I don't have a particular opinion on whether that would be a good idea or not.

Jut my opinion, but I don't believe the RDfa group will want to do _any_ processing of the literal. My Ruby RDFa implementation supports a simple innerHTML-like content model..


> Best,
> Richard
Received on Monday, 21 May 2012 14:50:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:29:53 UTC