[Om3] [OpenMath] #150: CD-level properties (CMP and FMP)

[OpenMath]

#150: CD-level properties (CMP and FMP)
-------------------------------------+-------------------------------------
             Reporter:  lars_h       |         Owner:  kohlhase
                 Type:  SEP          |        Status:  new
  (Standard Enhancement Proposal)    |     Milestone:  OM2+OCD/MD
             Priority:  major        |       Version:
            Component:  CD3 Format   |    Blocked By:
             Keywords:               |  Dependencies:
             Blocking:               |  Due to close:  YYYY/MM/DD
        Due to assign:  YYYY/MM/DD   |
Include in GanttChart:  0            |
-------------------------------------+-------------------------------------
 The current CD format requires that every CMP or FMP is attached to a
 specific symbol, but it is not uncommon that a property one wishes to
 state is hard to classify as having more to do with one symbol than with
 another.

 One example would be the identity
 {{{
   (a \circ b) \otimes (c \circ d) = (a \otimes c) \circ (b \otimes d)
 }}}
 (which appears on page 15 of http://arxiv.org/abs/1204.2421, if you want
 some context): is this an identity for \otimes or an identity for \circ?
 Neither choice has any precedence over the other.

 A solution to this would be to allow CMPs and FMPs also as a direct child
 of the CD element, since this lets you state the basic claim that this is
 a property of the symbols in this dictionary, without attributing it to
 one specific symbol. Come to think of it, that applies equally well to
 Examples too.

 By contrast, I expect it to be less common to have properties not specific
 to one content dictionary. A property may certainly mix symbols from
 several dictionaries, but it is usually introduced as part of the creation
 of one specific dictionary, so that will be where it is natural to state
 it. If one is simultaneously creating two content dictionaries and have to
 state a property that mixes symbols from both, then the question must
 arise why the two dictionaries should be separate. (There could be
 reasons, but I would not expect that to be usual.)

 As per Michael's remark on #149, I'm adding the om3 list as Cc here.

-- 
Ticket URL: <https://trac.mathweb.org/OM3/ticket/150>
OpenMath <http://www.openmath.org>
The development of the OpenMath Standard and Content Dictionaries.