[om] OMBVAR and related question

Paul Libbrecht paul at activemath.org
Fri Apr 26 09:46:21 CEST 2002


As we have some folks working on statistics content for ActiveMath, we 
have been thinking about it since very very long, including the ways or 
hacks to write it in QMath.

Indexed variables of course occur as a general utility (that is without 
vectors) quite often already. For example, in a sentence like "Let a_1 
and a_2 be this or that.".
For this, it seems <OMV name="a_1"/> is the only approach.

In general however, we needed often such statements as "Let A_1, ..., 
A_n be events." For this, it appears clearly that lists are appropriate 
and that one wants to consider A as a variable of type sequence and the 
object "A_1, ..., A_n" as the application of A to an interval.
	<OMV name="A">
     		<OMS cd="interval1" name="integer_interval"/>
     		<OMI> 1 </OMI>
     		<OMI> n </OMI>

In particular, the semantic of the notations "A_1, ..., A_n" compared to 
that of "A([1,n])" may have a subtle difference (it's all about the 
meaning of the comma...). And if not semantic, we shall need a way to 
write these differently (e.g. with the OMDoc style attribute of OMS, or 
OMA?) even supporting authors that want to write "A_1", A_2, ..., A_n" !.


On Mercredi, avril 24, 2002, at 07:57 , Andreas Strotmann wrote:
> a)  Did you notice that, semantically, you're using a variable "a" here
> (and not a_i) and that the subscripting function is actually the
> projection in tuple space \pi(a,i)  (I think it's usually written 
> \pi_i(a)).
> b)  If you would like to have "subscripted variables", it might be 
> better
> to have something like <omv ...> ... <omv>, with omv becoming a 
> general-purpose constructor for variables that has a variable 
> descriptor (with subscripts etc.) as sub-nodes.
om at openmath.org  -  general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems

More information about the Om mailing list