Re: WOFF and extended metadata

Chris, Liam, James,

I had been assuming that most of us wanted a metadata scheme that was NOT arbitrarily extensible, in terms of data structure. Is there any restriction on what future markup might be accepted for Japanese, Arabic, etc.?

To be clear, I assumed that string data would be the only data type allowed as a value (whether attribute or element content), implying encoding of any privileged XML characters. Just because XML was chosen (rather than JSON or Windows INI) as the transport mechanism for these key-value pairs doesn't mean that *any* XML should be valid in the schema. On the other hand if we allow only plain strings, then we have the great benefit that metadata validation becomes trivial.

What's the problem with decoding a string, then applying XSLT, if a UA knows the format of a given chunk of data?

What I've just written applies equally whether deep hierachies in key-values are allowed or not.

- L

On 20 Jun 2010, at 11:00, Chris Lilley wrote:

> On Sunday, June 20, 2010, 12:48:04 AM, Laurence wrote:
> 
> LP> Liam,
> 
> LP> Are you saying you support, in principle, the idea of a vast and
> LP> complex raw XML document (ok, maybe without the <?xml ... ?>) as the value of a key-value pair?
> 
> No, he is saying he does *not* support the idea of a vast and complex and obfuscated XML document as the value of an attribute :)
> 
> Liam and James are correct, and such a proposal would rapidly be criticized by the internationalization working group at last call.
> 
> Human-readable strings should be element content, not attributes; this is because in some languages they may require markup for things like Ruby (in Japanese and Chinese) or bidirectional overrides (in Hebrew and Arabic).
> 
> Obfuscated xml stuffed into attributes means that xml tools such as xslt cannot get at it. I note that one obvious implementation path for a 'font properties' metadata-displayer would be xslt.
> 
> -- 
> Chris Lilley                    mailto:chris@w3.org
> Technical Director, Interaction Domain
> W3C Graphics Activity Lead
> Co-Chair, W3C Hypertext CG
>

Received on Monday, 21 June 2010 15:09:25 UTC