Ian noted that DAML+OIL requires that both
subclass
and
subproperty
can be reflexive.
DAML+OIL does use metaclasses (rdfs:Class
and rdfs:Property
are essentially
both classes and metaclasses) as a syntactic device to define our
classes (e.g. DatatypeProperty), but metaclasses are not addressed in
the semantics.
We were unsure what "fit in the middle" referred to.
This list doesn't address XML syntax.
We've received considerable feedback that
the current syntax is good for machines, but not for humans.
We may want to also relate our observations about
ID
/about
and the
inability to use namespaces in attribute values.
We may want to separate the "data types" issue into 2 issues: XML Schema builtin datatypes (now in DAML+OIL) and derived and complex data types (not directly addressed by DAML+OIL).
In addition to "RDF gives meaning to all syntax" (although rdf:parseType provides some extensibility), RDF also gives meaning to all triples.
ACTION (Frank): update Peter's list based on our discussion
The following people have indicated their plans to attend:
Dan Connolly | no |
Mike Dean | yes |
Frank van Harmelen | no |
Pat Hayes | yes, but arriving late |
Ian Horrocks | yes |
Deb McGuinness | yes |
Peter Patel-Schneider | no (may call in) |
Lynn Andrea Stein | yes (2-5pm) |
rdf:parseType="daml:collection"
list syntax to specify values.
This is related to Dan Connolly's setOf and listOf properties.
Note that
daml:first
(and
daml:rest
)
used by
daml:List
is defined as an
rdf:Property
,
not
daml:ObjectProperty
or
daml:DatatypeProperty
,
limiting the semantics associated with
daml:List
.
Ian discussed a technique for preventing cycles in lists
(modulo an issue with pure, disconnected cycles),
and subsequently provided a more detailed
example
(which also includes a
ListC
class that addresses Jim's problem).
Deb and Richard Fikes had received email about most of these issues as they were discovered. They are working on fixes to the axiomatic semantics as part of an upcoming DAML+OIL paper for IEEE Intelligent Systems (which is an update to the DAML-ONT paper published in the MIT Web Semantics book), but probably won't get to finish this for at least 2 months. This will also include several fixes from other sources.
RESOLVED: We will add these updated semantics to the DAML+OIL (March 2001) page, alongside the initial version for this language release.
Stefan and Pat will plan to introduce their rules proposal.
$Id: 2001-06-26.html,v 1.7 2001/07/09 15:42:24 mdean Exp $