[om] A Proposal for extending OpenMath with structure sharing

[Michael Kohlhase]

Dear all,

If I may summarize the discussion so far, we seem to have two proposals on
the table 

1) my original proposal of adding an OMR element (only) to the XML
   encoding. 

   [this is equivalent to the OMRSYN proposal of James and does
    not change the OpenMath standard, only the encoding] 

2) to additionally add an OMRSEM elemement to the encoding AND change the
   OM standard for objects by allowing DAGs instead of trees explicitly.

It seems that the consensus is that 1) is relatively safe (encoding only)
and that 2) would have additional benefits, but is more disruptive. I would
propose to discuss (and eventually vote on) the two proposals separately.
Obviously, proposal 2) only makes sense if we also accept 1), so I propose
that we move on 1). If we talk about 2), we will need somebody to spec out
the changes to the standard document like I did for 1), and come up with
some convincing examples of the added value for the OpenMath Standard.

The concrete syntax of these elements (OMR or OMRSYM) is relatively
immaterial at the moment. My preference would be to have 
<OMR semantic="yes" ... /> for OMRSEM and <OMR semantic="no" ... /> for OMRSYN
(we could make the default value for the attribute 'semantic' to be "yes"),
so that the element <OMR .../> is retained as in my original proposal, and
we can still add the semantic sharing idea seamlessly afterwards.

     Michael





--
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