- From: Dave Reynolds <der@hplb.hpl.hp.com>
- Date: Sun, 05 Nov 2006 20:48:59 +0000
- To: public-rif-wg@w3.org
Transcribed slide summary of extensibility/dialect breakout group.
Dave
(1) Rule Components
o Core:
- variable declarations
- antecedent
- consequent
o some dialects (e.g. ECA) might add further elements
o other dialects just vary the content, e.g. production rules have
actions in consequent
o Christian's UML diagram example
(2) Dialects
o Starter set: LP, PR, FOL
o Discussion on FOL in scope
o Discussion on LP purity, e.g. is real prolog really a sub-dialect
of LP led to discussion on ordering ...
(3) Ordering discussion
o By default syntactic ordering will not be preserved.
o For dialects with order-dependent semantics then an ordering
construct will be provided.
[Some discussion on ruleset v.s. within rule differences, dropped
after further discussion]
(4) Syntax extensions
o All RIF standard dialects should use a common RIF syntax, extending
only where necessary.
o Syntactic extension mechanism?
- Substitution groups?
- RDF
Significant disagreement on whether RDF is an admissible
mechanism ...
(5) Syntax mechanism issues
o charter says "normative XML syntax"
o Christian - this means XML Schema
- disagreement
- Dave comments on RDF as metadata already opens door to use of RDF
in other situations within RIF
- so this is an ISSUE
Note: It is hard to decide this in isolation, specific extension
requirements, degree of (syntactic) validation would need to be
defined. Examine this issue with test cases and concrete
proposals. Not in abstract.
(6) Extensions
o common libraries to draw from
e.g. mathematical operators not in core but should provide these so
dialects reuse the same set
[we need a name, other than core, that designates the shared stuff that
we provide for all dialects to use]
Received on Sunday, 5 November 2006 20:49:40 UTC