[tei-council] num

[Lou Burnard]

This was indeed one of the options I proposed in my email. But I 
wondered whether expanding the meaning (or reducing the constraints) 
associated with "data.numeric" might not be considered to be flouting 
the Birnbaum doctrine...

Sebastian Rahtz wrote:
> I am inclined to agree with Lou's argument
> that we should not re-invent datatypes with
> paired attributes.  but
> why not simply extend our data.numeric to also
> support rational numbers? then Gabriel
> can just use @value="22/7" as he needs.
>
> there is prececedent - several of our data.*
> types are the W3C ones plus additions. data.numeric
> itself is already
>
> <rng:choice>
>      <rng:data type="double"/>
>      <rng:data type="decimal"/>
>    </rng:choice>
>
> so why not change it to
>
> <rng:choice>
>      <rng:data type="double"/>
>      <rng:data type="decimal"/>
>      <rng:data type="token">
>         <rng:param name="pattern">([\d\-]/[\d\-])</rng:param>
>      </rng:data>
>    </rng:choice>
>
> (or does \d already allow for - prefix, ie "-1"?)
>