Re: telecon today + RDFa Implementation Report update

On 8 Aug 2008, at 09:14, Hausenblas, Michael wrote:

>> Deliberately fail because I think my output is better/more useful
>> than the output required by the spec: 93, 99, 108, 112.
>
> I'm not sure I can follow. Can you be more specific, please?


An example, not taken from the test case:

   <span about="ex:foo" property="ex:bar" datatype=""
   >The human brain has about 10<sup>11</sup> neurons.</span>

The output demanded by rigourous adherence to the spec is:

   ex:foo ex:bar "The human brain has about 1011 neurons."

Cognition's output is:

   ex:foo ex:bar "The human brain has about 10[11] neurons."

which I feel is slightly more useful. (Of course, "^11" might be even  
better, but that would require some sort of heuristic to determine  
whether the superscript has been used as an exponent, or a footnote,  
or as something else entirely.)

Basically, Cognition will convert from text/html to text/plain, not  
through a simple concatenation of text nodes, but in a manner as  
close as possible to a text browser (like lynx). I know this deviates  
from the RDFa spec, and I'm not suggesting that the spec change - I'm  
just saying that I'm happy to be non-conformant in this regard.

-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Friday, 8 August 2008 08:39:30 UTC