[om] A Proposal for extending OpenMath with structure sharing

Steve Braham warp at polylab.sfu.ca
Fri Mar 29 18:17:06 CET 2002


This is actually a great idea for those of us who are working in
compact semantic transmission (in my case, for spacecraft telemetry
and remote software updates).

It would be easy to implement in the PolyLAB library, as we already
use DAGs. The only issue being that some of our cloning mechanisms
project the result to trees.

Probably the biggest issue would be how to come up with rules for
updating such structures, because you have to differentiate between
updating one of n references, or updating all objects that
reference the object (indeed, that's the main reason for cloning
protection in object-oriented data transmission systems).

But using xlink concepts might help there, as you can differentiate
between updating the link, and updating the value. Only issue would
be the issue in a "strong" OpenMath world, where you'd have to decide
about whether an xlinked object can change type.


			Steve

Michael Kohlhase wrote:
> 
> Dear all,
> 
> In Cannes I proposed to add structure-sharing to the XML encoding of the
> OpenMath standard. The main benefits would be that OM objects can be
> encoded as directed acyclic graphs (DAG), as opposed to trees, resulting in
> more compact representations.
> 
> In the following I will sketch an extension of the OpenMath XML encoding and
> discuss the design decisions and necessary changes to the standard.
> [rest of the cool stuff deleted]

--
Stephen P. Braham			Director, PolyLAB
warp at polylab.sfu.ca			TIME Centre 
(604) 268-7981				Simon Fraser University 
(fax) 268-7980				Harbour Centre Campus
http://team.polylab.sfu.ca/~warp/	Vancouver, BC, Canada

   PolyLAB: From the Classroom to Space, http://polylab.sfu.ca/
--
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