Re: WOFF and extended metadata

> Jonathan Kew:
>>  This sounds to me like something that goes way beyond what people would reasonably expect to express 
>> in a generic key-value pair metadata form. Also, to be widely useful, it ought to be given a standard, defined 
>> place in the metadata spec. As such, it could easily be defined (e.g., in WOFF metadata version 2.0) as a new 
>> <sample-xhtml> element that directly includes the relevant XHTML fragment. This is NOT what the key-value 
>> metadata extension mechanism is intended for, and we wouldn't want UAs cluttering their "Show Font Info" 
>> panel with it anyway.

Sylvain:
> Strongly agree. This is meant to support font metadata, not sample pages, demos, advertising and what not. 
> It is entirely possible to link to one or more sample pages from the metadata and this is in fact the kind of
> use-case key-value pair extensibility is well suited for.
> 
> Besides, wouldn't one prefer to keep such samples in a place where they can be accessed and updated 
> independently of each individual font files ? 

I agree too. I do not advocate including them.[*]

But in a freely extensible metadata scheme people will shoehorn complex data in there, so it's beside the point if any of us like it or not. An unpopular use case that some people insist upon still needs to be handled.

My XHTML example was intended to provoke discussion on how we encode complex data in general. Deep nesting of KVPs is unpopular on these lists, yet arbitrarily deep XML seems to be acceptable, preferred over a plain string.

- L

[*] BTW if people wanted to include HTML templates, rather than XHTML, of course that *would* need to be entity-encoded since in general HTML is invalid XML.

Received on Monday, 21 June 2010 21:54:56 UTC