[om] Status of latest standard?

David Carlisle davidc at nag.co.uk
Tue Jan 20 16:51:24 CET 2004

  It would be nice to state the remaining issues... I have lost the 

Looking at our bugzilla list (which may not have all the issues raised,
and of course anyone may raise more issues) the items appear to be
as follows, together with some quick annotations by me.


 *  update 1.1. history 
 *  be more consistent with initials in the bibliography 
 *  raise cdbase in the examples
     recent drafts of om2 have allowed cdbase to be set on an ancestor
     element and affect all OMS elements in scope, most of the examples
     dont show this usage though


 *  IEEE special values
     See recent discussion on this list on the semantics of NaN, -0 and
     similar IEEE special values. Mainly this is editorial, expanding
     the text to explain the intention of the OMF hex=... syntax in OM1,
     but there is also the technical issue of allowing (or not) the new
     syntax  dec="NaN" and the related issue of making (or not) the dec
     attribute have type xsd:double rather than xsd:string in the

 *  schema for cd improvements
    Bill Naylor suggested on this list some bug fixes, and also some possibilities
    for using Relax NG facilities to make a "tighter" schema for CDs
    than was possible using DTD. I don't think any of this is
    controversial and it is done in an internal draft (I think).

 *  restricting FMP to one OMOBJ 
    Should FMP be restricted to just one OMOBJ (consensus on this list
    appeared to say yes)

 *  indexed variables 
    The current drafts allow <OMV name="x"><OMI>1</OMI></OMV> but there
    appear to be some real issues with the detailed semantics of the
    proposal, especially as it allows arbitrary OM expressions in there
    which causes problems with binding operators. This has been
    discussed on this list without any firm descison yet. My own
    preference would be to take this out and revert to OM1 syntax here.

 *  allowed name characters 
    The current draft has a prose description of the allowed characters
    in Symbol and Variable names, there is a proposal to explictly bring
    this in to line with XML variable names (and use the xsd:NCName type
    in the schema, a sub-issue is whether to use the XML 1.1 names
    rather than XML 1.0 (These are designed to be future-proof with
    respect to future extensions to Unicode).

 *  add example for packet streaming in binary encoding. 
 *  add cdbase to the binary encoding
    Generally the binary encooding needs to be checked that it is
    consistent with the final version of the abstract mode and the XML

   James Davenport (and others) have suggested a modified version of FMP
   that highlights symbol definitions, there was again quite a long
   discussion on this list, but no final desciosn as far as I'm aware.

*  abstract content dictionaries without symbol definitions
   should abstract CD model force that there has to be a symbol
   definition in the CD (currently it doesn't explictly mandate this)

*  Normatitivity of schema
   Should the compact syntax RNC or XML syntax RNG schema be normative
   and do the non-normative versions need to be explictly in the text of
   the standard, or can they be distributed "with" it, just as machine
   readable files. 


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:
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