- From: Robert Muetzelfeldt <r.muetzelfeldt@ed.ac.uk>
- Date: Fri, 14 Jan 2011 11:10:53 +0000
- To: www-math@w3c.org
Hello, I've been involved in the development of Simile, a visual modelling environment for continuous systems modelling (http://www.simulistics.com/). Simile uses Prolog as its native model-representation format, but we have a corresponding XML Schema which has a close, more-or-less one-to-one mapping with the Prolog syntax. This uses Content MathML for representing the model equations. Rather than having two converters (say, Prolog for converting from the Prolog to the XML syntax, and XSLT for going in the other direction), I have explored the possibility of having a single set of conversion rules which can be used in either direction. This has a clear advantage in that there is only one body of code to maintain. I have worked on two quite different technologies for doing this. One is XSugar (http://www.brics.dk/xsugar/), which allows the user to define a grammar in which the right-hand side of each production has two parts - one for a plain-text syntax, the other for the corresponding XML syntax. The XSugar engine then allows the user to provide a file in either syntax, and generates the other. The other is Prolog, and exploits the fact that one argument in a binary Prolog rule can be instantiated, and the Prolog engine will then try to instantiate the other. Both approaches are working, and include the ability to convert fairly complex equations to/from Content MathML. I would appreciate hearing whether this endeavour is of interest to the members of this mailing list, and whether you are aware of other initiatives which are tackling bi-directional conversion between plain-text equations and Content MathML using a single set of mapping rules. Many thanks, Robert Muetzelfeldt -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
Received on Friday, 14 January 2011 11:50:29 UTC