[Om] Content-dictionary notations draft

[Professor James Davenport]

Bill wrote (Thu, 22 Mar 2007 11:08:05)
>    ?
>   d
> ----- f(x)  in this example, we might expect ? = x + y
> d  d
>  x  y
In fact ? = #{x,y}. More complicated still, and requiring general 
arithmetic, is 
   ?
  d
----- f(x)  in this example, we might expect ? = x + y
 2  3
d  d
 x  y
where ? = 5 (as 2+3: a rendered which printed 2+3 would not be taken 
seriously), but in 
   ?
  d
----- f(x)  in this example, we might expect ? = x + y
 m  n
d  d 
 x  y
one has no choice but to print m+n.
<begin logician mode>
I suspect that this means that a "perfect rendered" has to solve the 
constant problem (Richardson, 1968), but this doesn't mean that a 
mechanism can't specify addition, and leave the renderer to do the best it 
can. 
<end logician mode>

Paul said (Thu, 22 Mar 2007 12:52:44)
> Your template function idea is nice and useful, for sure, but it would rely
> on one enormous unknown: a specification of evaluation of OpenMath 
> (somewhat like the NMC is approaching).
I hope my point above answers this.

James