Re: <ci> with <sep/>

Clare,

    The <sep/> element only comes into play with explicit
representation of numbers - inside of <cn> tags.  It is 
not allowed inside of <ci> tags.

While one might imagine a markup such as

    <ci type="complex-cartesian">x<sep/>y</ci>

the association of x with a type would be implicit in
the construct and the recommendation generally avoids
inferring types from position.  The cn tag use of <sep/>
was a consession to the special needs of actual numeric data.

Your best bet for a symbolic representation of complex-cartesian
coordinates is something like

    <apply><csymbol definitionURL="complex-cartesian"/>
        <ci type="real">x</ci>
        <ci type="real">y</ci>
    </apply>

It is especially helpful if the definitionURL point to some
standard definition such as an entry in an OpenMath dictionary.

Stan Devitt

----- Original Message ----- 
From: "Clare So" <cmso@uwo.ca>
To: <www-math@w3.org>
Sent: Friday, July 26, 2002 2:51 PM
Subject: <ci> with <sep/>


> 
> 
> 
> 
> Hello all,
> 
> I am wondering the followings are valid Content MathML:
> 
> <ci type="complex-cartesian">x<sep/>y</ci>
> 
> <ci type="rational">x<sep/>y</ci>
> 
> <ci type="complex-cartesian">x<sep/>y</ci>
> 
> <ci type="complex-cartesian">3<sep/>y</ci>
> 
> <ci type="rational">x<sep/>3</ci>
> 
> 
> Obviously, x and y are identifiers in the cases above.  How about:
> 
> <ci type="rational" base="16">a<sep/>b</ci>
> 
> In this case, a and b can be a hex number!
> 
> 
> Can anyone here suggest a better way to handle <ci>s using <sep/>?
> 
> thanks!
> 
> 
> Clare
> 
> ============
> Clare M. So
> Summer Undergraduate Assistant
> Ontario Research Centre for Computer Algebra (ORCCA)
> University of Western Ontario
> London, Ontario, CANADA
> 
> clare@scl.csd.uwo.ca
> 
> 
>

Received on Monday, 29 July 2002 09:14:09 UTC