[om] overloaded symbols and error symbols (fwd)

[Bill Naylor]

Sorry, that address got spelt wrongly!!

---------- Forwarded message ----------
Date: Sun, 15 Feb 2004 16:12:18 +1300 (NZDT)
From: Bill Naylor <Bill.Naylor at mcs.vuw.ac.nz>
To: David Carlisle <davidc at nag.co.uk>
Cc: monet-project at nag.co.uk, om at openamth.org
Subject: overloaded symbols and error symbols

>
> Maybe the existing evaluate could be extended by including a third
> signature in poly.sts
>
> evaluate : polynomial_ring[@cd="polysts],Vector(equations) -> NumericalValue
>
> possibly even a fourth
>
> evaluate : polynomial_ring[@cd="polysts],Vector(equations) -> polynomial_ring[@cd="polysts]
>
> This last may be considered partial_evaluation? I'm not really sure
> whether there is anything so fundamentally different about this type of
> operation that it would require a different symbol.
> Would an application which said that it supported a symbol have to
> support all versions of that symbol. Does the spec. say any thing about
> this? It seems to me that if a symbol is overloaded, and therefore has a
> number of different interpretations. An application should not be required
> to either not support the symbol or support all interpretations. The
> evaluate I have just described is a case in point.
>
> If the spec says nothing about this. I think it should! (somewhere in
> section 6.2 I believe). This also brings a further point, which is about
> the error CD, maybe unsupported_symbol needs to be overloaded to specify
> which version of the symbol is not supported.
>
> Any thoughts about these thoughts?
>
> Bill
>

 Thinking a bit more about the error symbols, specifically, the ones from
error.ocd. at present they are expected to only return as an argument the
symbol, where the error occured. This would mean that the application
which sent the erroneous OpenMath, would not be able to differentiate
between two different versions of one symbol (in the case that it was
overloaded).
 I suggest that the error symbols from the error CD should be overloaded,
to accept the children of the surrounding context. This would allow the
receiver of the error symbol to work out what the source of the error had
been even in the case of an overloaded symbol.

Bill


________________________________________________________________________
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
________________________________________________________________________
--
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