[om] DefMP elements

[Andreas Strotmann]

Professor James Davenport wrote:

>On Mon, 8 Dec 2003, Andreas Strotmann wrote:
>  
>
>>Jacques Carette wrote:
>>    
>>
>>>On the specific recommendations:
>>>
>>>I definitely agree with those first 3.
>>>
>>>As to the fourth:
>>>d) use signatures in definitions: this might prevent a definition from 
>>>being misapplied in an application that doesn't know any better. This 
>>>may well be done in the simple way of using universal quantifiers and 
>>>set memberships inside the defining FMPs, so that this may simply be 
>>>added as a recommendation to CD writers.
>>>      
>>>
>But the symbol being defined has a signature in STS, which is why the sin 
>on matrices is not the sin from transc1.
>  
>
a) In OpenMath, the STS signature file is not normative, but the CD is. 
In this case, the STS file may well need cleaning up. In general, STS 
files may eventually be derived from the definitions in a CD.

b) The STS signature file is allowed to have several signatures; 
different signatures may apply to different definitions.

c) In a previous post, you argued yourself against defining a new CD 
just for sin (etc.) of matrices, since they are not fundamentally 
different from their scalar versions (indeed, it is possible to provide 
definitions that work for both identically, which argues that they are 
fundamentally(!) the same concept, which means that the sine of matrices 
*should* be the one from transc1 if it were not (yet).)

d) Notice that a signature in the sense that I meant it doesn't have to 
conform to the usual syntax of a signature in the STS sense, as the 
examples I posted in a previous message showed. It's just that the 
function of a signature should be fulfilled in a definition.

 -- Andreas
--
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