[ISSUE-51] Formulae wrappers (2/2: required or optional)

All,

Beside the question of enabling different dialects to distinguish 
different kinds of groups of FORMUALE without bloating FLD (see [1]), 
and still leaving aside the orthogonal issue of where to hang metadata, 
there might be another issue in the Group/Rule discussion that I did not 
see explicited in emails nor telecon mminutes (but I missed some 
telecons recently).

In a solution like the one in the first part of this message [1] 
(similar to the one Chris proposed [2]), RULE formulae can be wrapped 
directly in a Group (without each being wrapped in a Rule WRAPPER):

   Document       ::= 'Document' '(' Group? ')' /* leaving meta etc out
   WRAPPER        ::= Group | Rule
   Group		 ::= 'Group' '(' (RULE | WRAPPER)* ')'
   Rule		 ::= 'Rule' '(' RULE ')'

Is that a issue for the people who want a specific Rule wrapper or isn't 
it one?

As I mentionned already, a problem I can see with that solution is that 
there are many different ways to serialise the same set of RULE formulae 
(e.g. as a Group of RULEs or as a Group of Rules). Or isn't that a problem?

If it is, a solution could be to disallow Groups to wrap RULEs directly:

   Document       ::= 'Document' '(' Group? ')' /* leaving meta etc out
   WRAPPER        ::= Group | Rule
   Group		 ::= 'Group' '(' WRAPPER* ')'
   Rule		 ::= 'Rule' '(' RULE ')'

But that would make wrapping each RULE in a Rule mandatory, instead of 
optional: or is that desirable to some people in the group?

Another question regards the Document production: would it be a problem 
to allow a Document to contain several Groups at the same level? If not, 
we could have:

   Document	::= 'Document' '(' WRAPPER* ')'

instead (same in FLD),which would allow a document to contain a set of 
Rule WRAPPERs directly, where an additional Group is not needed.

Opinions? Comments?

Christian

[1] http://lists.w3.org/Archives/Public/public-rif-wg/2008Apr/0186.html
[2] http://lists.w3.org/Archives/Public/public-rif-wg/2008Apr/0148.html

Received on Tuesday, 29 April 2008 10:48:49 UTC