Re: xml:lang und kgV / ggT ;-)

The thinking has been that the cases you describe for gcd 
are actually the most common at the lower level, and that for more 
advanced use you can flag use of the extended definition through 
use of the definitionURL and encoding.  

Also,  MathML is notational and not computational.  
In particular, you need to be able to write 

    "The expression  "gcd(a,b)"   is undefined  whenever a and b 
    from a .... ( pick your favourite algebraic structure ...)
    "

It cannot be illegal to write incorrect mathematics, or we could
not even describe to someone that it was incorrect.  Correctness
only becomes an issue when you try to carry  out such a
computation, but that is done outside of MathML.

The actual rendering of gcd in  different languages  can be 
(or should be) different, and perhaps the spec needs to be 
clearer about such localization issues.

Stan Devitt

----- Original Message ----- 
From: William F. Hammond <hammond@csc.albany.edu>
To: <strotman@nu.cs.fsu.edu>
Cc: <www-math@w3.org>
Sent: Tuesday, April 11, 2000 8:38 PM
Subject: Re: xml:lang und kgV / ggT ;-)


> <gcd/> and <ggT/>:
> 
> Language issues aside, is it really sensible to have "gcd" hard-coded
> at all?
> 
> For integers it's well-defined if one insists, as usual, that it be
> positive, and for the ring of polynomials over a field, it's
> well-defined if one insists that it be monic.  Beyond that it might
> just exist as an ideal, i.e., may not have the same type as its
> arguments.
> 
> I would submit this as an instance where "less is more".
> 
>                                       -- Bill
>

Received on Tuesday, 11 April 2000 23:29:16 UTC