Re: use of literal 'pi' in cn example (section 4.4.1.1.3)

At very least, the example now encourages less than optimal encodings.
It might be less controversial if a different greek letter 
had been chosen, but I also think the points raised in this
dicusssion remain valid.

Stan.

> 
> 
> 
>   1) if the intention is to refer to a symbol called 'pi', then the
>   appropriate mechanism is to use the <csymbol/> element instead of using the
>   <cn/> element
>   2) this constant is already defined as <pi/>
> 
> I think that's probably true (now). The example dates from mathml 1 which
> didn't have either <csymbol> or <pi/>. So as you commented, it's
> definitely still legal but I think almost all the uses for which
>  <cn type="constant" was intended could better be done with the newer
> MathML2 Content elements such as <pi/> or with <csymbol>. so this usage
> probably should be deprecated. (Now is a good time to raise such issues
> as we are working on a draft of mathml2 2nd edition....)
> 
>   Since the grammar does not say what type="constant" means,
> 
> The grammar doesn't, but 4.3.2.9 on the type attribute does say more
> about this (and would need changing if your suggestion to deprecate this
> usage was accepted).
> 
>  
> David
> (personal opinion: not discussed this with the other WG members yet)
> 
> ________________________________________________________________________
> This e-mail has been scanned for all viruses by Star Internet. The
> service is powered by MessageLabs. For more information on a proactive
> anti-virus service working around the clock, around the globe, visit:
> http://www.star.net.uk
> ________________________________________________________________________
> 

Received on Thursday, 23 January 2003 09:29:38 UTC