W3C home > Mailing lists > Public > public-rdfa-wg@w3.org > June 2014

Re: Keeping XMLLiterals unchanged when going through RDFa

From: Reto Gmür <reto@apache.org>
Date: Wed, 4 Jun 2014 22:46:21 +0200
Message-ID: <CALvhUEUFjaeNhTW2kZ6dp-=o4M_w-cUvvk2c8dOgKawU_doaRA@mail.gmail.com>
To: Shane McCarron <shane@aptest.com>
Cc: Gregg Kellogg <gregg@greggkellogg.net>, W3C <public-rdfa-wg@w3.org>
Thanks for your replies.

Are SPARQL endpoint expected to check the identity of xml-literals based on
their canonical forms?

There is still the practical issue with GreenTurtle that it returns
whatever the Serializer of the browser returns, this is different from
browser to browser and definitively not canonical XML.

When the HTML is serialized as XML things are a bit more consistent,
however the sent back XMLLiteral has now all the namespaces from the
document it was embedded into, as according to http://www.w3.org/TR/xml-c14n
all namespace declarations are retains this effectively yields to a
different XmlLiteral even for the canonical form.

Cheers,
Reto

On Wed, Jun 4, 2014 at 8:43 PM, Shane McCarron <shane@aptest.com> wrote:

> Agreed - the canonical version is guaranteed to be same as the lexical
> form as far as the parser is concerned.  That's the point of
> canonicalization.
>
>
> On Wed, Jun 4, 2014 at 12:35 PM, Gregg Kellogg <gregg@greggkellogg.net>
> wrote:
>
>> On Jun 4, 2014, at 3:25 AM, Reto Gmür <reto@apache.org> wrote:
>>
>> > Hi,
>> >
>> > I'm writing a browser based editor to edit graphs rendered as RDFa/HTML.
>> >
>> > I encountered the problem that the round-trip via browser changes the
>> value of the XMLLiterals, e.g. a <br/> becomes <br /> (in firefox) or a
>> <br> (in chrome). There seems to be no way to the client library to get the
>> exact original version as the js has only access to the DOM not to the
>> literal source. This of course makes it impossible to update the content
>> using e.g. a SPARQL DELETE/INSERT.
>> >
>> > I though one work around would be to pass the original version in a
>> content attribute. This way the client could read out the exact literal
>> before the editing starts and use this version when sending back the
>> DELETE/INSERT request.
>> >
>> > Unfortunately I noticed that the content node is ignored if the
>> datatype is set to rdf:XMLLiteral.
>> >
>> > Is there any way to encode an arbitrary XMLLiteral in RDFa so that it
>> remains identical character-by-character when encoding as RDFa and decoding
>> again?
>>
>> The point of XMLLiteral (and rdf:HTML, which may be more appropriate for
>> you) is that it does not preserve the lexical representation of the values,
>> but that the values are the same, after some form of canonicalization (XML
>> C14-XL for XMLLiteral and coercion to DOM for rdf::HTML). If you really
>> must preserve the lexical representation, then there’s really know way to
>> do this in RDFa without encoding it using HTML escapes, so that it’s
>> treated just as a string literal.
>>
>> Probably better if your tool didn’t expect the exact lexical form, but
>> allowed for some canonicalization, or always emit canonical XML/HTML.
>>
>> Gregg
>>
>> > Cheers,
>> > Reto
>>
>>
>>
>
Received on Wednesday, 4 June 2014 20:46:48 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:05:35 UTC