W3C home > Mailing lists > Public > public-rdf-wg@w3.org > November 2012

Re: Disimproved definition of literals in Concepts; close ISSUE-94?

From: Richard Cyganiak <richard@cyganiak.de>
Date: Wed, 14 Nov 2012 17:17:08 +0000
Cc: <public-rdf-wg@w3.org>
Message-Id: <542051FC-18E8-4421-B097-2E753B82F2FD@cyganiak.de>
To: Markus Lanthaler <markus.lanthaler@gmx.net>
On 14 Nov 2012, at 15:04, Markus Lanthaler wrote:
>> This leaves the option of adding language tags to only a limited set
>> of datatypes, like rdf:HTML. The objection there is that this would add
>> *more* exceptions to the design of RDF literals (where our goal was to
>> make the handling of literals more uniform), and also rdf:HTML doesn't
>> need it because it already has a mechanism for language annotation
>> (<span lang="xx">).
> 
> I see, but I still don't understand why that single exception exists then!?

Because otherwise it wouldn't be possible to tag strings with language information. Proper I18n support is a pretty hard design requirement for most things that come out of W3C.

> For historic reasons? To avoid blank nodes? 

Probably because any other approach would be worse. Language-tagged literals are *very* common, so requiring an extra blank node in the general case doesn't sound appealing.

> Yes, it's true that rdf:HTML has a mechanisms for language annotation. The
> problem is that if you want to filter such values based on language you have
> to parse each value which is a quite costly operation.

That's sort of a fair point, but you have to peek inside the rdf:HTML literal anyway because the language may be overridden in the content. If you know that's not the case because you have control over the content, then you may just as well publish the data using the rdf:value design pattern like Andy suggests.

> There are certainly a lot of other data types that could profit from
> language tags, just think of Markdown, Textile, all the Wiki syntaxes, etc.
> yet there's no way language-tag them. That's exactly what triggered my
> question.

Well, back in 2004, W3C still believed that in the future everything would be XML and therefore would already support language annotation.

And today there's no way of retrofitting language tags onto typed literals without breaking all existing parsers and stores, therefore it's more than unlikely to happen.

Best,
Richard


> 
> 
> Thanks,
> Markus
> 
> 
> 
> --
> Markus Lanthaler
> @markuslanthaler
> 
> 
Received on Wednesday, 14 November 2012 17:17:42 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:52 GMT