Re: [cme-dev] feature suggestion: predicates based on queries

[Peri Tarr <tarr@xxxxxxxxxx>]

Hi Stan,

Well, good news and bad news.  The
good news is that support for predicates is part of the query language
definition (and was from the start).  The bad news is that it's one
of the last major features that haven't been implemented yet.  :-(

The concern modeling component (conman)
already supports constraints (consistency conditions), though they have
not yet been used (translated: you may find some problems and/or things
that aren't quite implemented completely), and they aren't yet surfaced
in the GUI, along with different evaluation policies (the set of which
is extensible).  I'd love to see you use them and report back what
you think.

     Peri

Stan Sutton/Watson/IBM@IBMUS
Sent by: cme-dev-admin@xxxxxxxxxxx

08/06/2004 10:00 AM

Please respond to
cme-dev

To	cme-dev@xxxxxxxxxxx
cc
Subject	[cme-dev] feature suggestion:  predicates based on queries

Hi All, 

I'd like to suggest that we add a feature to the UI to provide support
for predicates, i.e., boolean expressions over concern models.

There are at least a few forms that a predicate language could take.  To
start with, though, I'd be content with an adaptation of our current query
language.  For example, I'd like to be able to test whether the results
of evaluating one query were the same as, a subset of, or disjoint from
the results of evaluating another query.  (Not that I might not enjoy
being able to evaluate some more complicated expressions.  :-)  )

Also, to start with it would be a big step to have a predicate-oriented
version of the current query and results views, i.e., somewhere a user
could formulate, evaluate, and possibly save predicates.  That would
at least support the manual testing of things like the correctness, consistency,
and completeness of concern models. 

It would also be nice to be able to attach predicates to concerns, as we
do now with queries, but as a way of specifying and possibly enforcing
consistency conditions.  Enforcement is a sizeable and tricky issue,
but just being able to attach predicates to concerns, and perhaps having
a button that would allow a user to manually evaluate the predicates attached
to a concern model, would be a big step. 

Regards, 

Stan