Outcomes of RIF telecon 16 Dec

Although we did not pass any formal resolution today, we feel that the
WG is close to consensus on a couple of key issues:

=Terminology=

We first discussed some useful definitions related to RIF conformance
and implementation, that were proposed by Michael and summarized by
Dave.

First, "implementing RIF" is centered on the idea of *translation*.

Second, translation has a direction: TO and FROM RIF.

Third, a translation may be partial or complete in a particular direction. It is at 
this level that conformance is "measured".  This leads to the "profiles" discussion 
for CORE.

We tried to assign terms to these particular things, but I did not have the sense 
that we have consensus there.

=What is the CORE=

We then went on discussing the nature of the CORE. The discussion centered
on whether or not all languages were required to be able to translate FROM "all" of 
the CORE to be conformant.  Some continue to feel this is unrealistic, however we 
lack examples that demonstrate it.  Several expressed support for a very limited 
notion of profiles for the CORE.  Profiles would specify features that we may 
consider "optional" or that may determine the degree of conformance of a translation. 
  Examples of features in a possible CORE profile were: recursion, decidability, 
complexity bounds, functions.

There seemed to be consensus that there is one core dialect with the expressivity of 
about Horn and that we should move forward with the specification of that dialect,
independently of other considerations.  If there is a notion of profiles it should be 
extremely restricted so that the "CORE is still a core".  At the moment, we do not 
have any specific "features" of the CORE that anyone has objected to, except possibly 
recursive rules, so it is still not clear that we need profiles for the CORE.

We discussed whether RIF dialects must include and extend the CORE.  The possibility 
of profiles opens the door for some dialects to eliminate certain features (again, 
from a very restricted set).  In other words, profiles may allow some dialects to 
extend a subset of the CORE.

There seemed to be consensus that the definition of the mandatory CORE must be 
motivated by practical considerations, such as the difficulty to translate a 
particular feature to a particular rule language of practical importance.

=Constraints=

We did not have time for much more discussion, and Hassan was not present. Michael 
presented his position that now that he has tried to work constraints into the CORE, 
he does not see how it can be done while still maintaining the design principles that 
the CORE be a core.  No one on the call claimed to both "understand Hassan's 
proposal" and "support it", and in fact there was only one of the latter.  No one 
seems to be objecting to it yet, but there seemed to be consensus that the pros and 
cons of constraints in the CORE were not well-understood enough at this point and 
need further discussion; and that constraints should be left out of the first public 
working draft as a consequence (without prejudice towards future decisions wrt later 
WDs).

-C&C

Received on Tuesday, 16 January 2007 20:23:14 UTC