- From: Frank van Harmelen <Frank.van.Harmelen@cs.vu.nl>
- Date: Thu, 09 Jan 2003 13:46:06 +0100
- To: Dan Connolly <connolly@w3.org>, Jim Hendler <hendler@cs.umd.edu>, Deborah McGuinness <dlm@ksl.stanford.edu>
- cc: www-archive@w3.org
Report of breakout session on Feature Synopsis Documemt
Overall conclusion:
- make it more introductory,
- less technical,
- stronger on motivation/purpose
- rename it "overview",
Main chunks of work:
- CHUNK "GUIDE/ROADMAP":
- CHUNK "MOTIVATION FOR THE LANGUAGE"
- CHUNK SUBLANGUAGES:
- CHUNK FEATURE LIST:
- CHUNK EDITORIAL
- Some suggestions for Guide document
CHUNK "GUIDE/ROADMAP":
- needs a reading guide/roadmap between the documents
- make Synopsis non-normative (ie informative)
- explicit statement of purpose, intended audience, follow-up reading
- remove ref. to WG expect in header
- title: "Web Ontology Language (OWL): Overview"
- Intended readership:
- anyone who wants to decide if they want to use OWL
- knowledge of RDF(S) is useful but not essential
- mention UML as a metaphor
CHUNK "MOTIVATION FOR THE LANGUAGE"
- need motivation for the language: lift par. from Requirements Doc
+ pointer to ReqDoc for further reading.
Relation with RDF(S) and XML
(both in terms of motivation and what's possible)
Shortest statement = expressive language for defining shared
vocabulary, with formal semantics that makes it suitable for machine
usage.
[see list of Dan]
XML: surface syntax, no/unconstrained semantics
XML Schema: restricting form of XML documents
RDF: datamodel for things + relations between them,
some simple semantics
RDF Schema: vocabulary for documenting properties and classes,
with semantics for generalisation-hierarchies of
properties and classes
OWL : more vocabulary for modelling concepts such as cardinality
("exactly one"), equality, richer typing of properties
("all carnivores eat meat")
This will help in:
- what does upgrading from RDF involve
CHUNK SUBLANGUAGES:
- Needs better stuff on sublanguages and semantics-layering:
- motivation "why each sublanguage",
- explanation "what it is"
- guide on "when to use what"
- consequences of this to choice
point to conformance issues
- strategy for feature-list should be additive
CHUNK FEATURE LIST:
- make feature list more introductory & shorter,
interlinked
for each feature: "this language element allows you to..."
(only scrap feature list alltogether if this doesn't work)
- incorrectness on sameClassAs/sameIndividualAsa
- max length 10 pages
if nec. economise on details and not on motivation etc.
- PostPone until issue is clearer:
- importing documents vs language levels
- we deliberately stay don't list RDF syntax
CHUNK EDITORIAL
interdocument consistency of terminology,
(dialect/language/vocabulary, instance/individual/member)
interlinking etc
- check against http://www.w3.org/Guide/pubrules#head
- Brian McBride and/or Dave Becket comments
- replace "RDF namespace" by "features already present in RDF(S)",
replace "OWL namespace" by "features introduced by this language"
----------------------------------------------------------------------
- SUGGESTION FOR GUIDE: Description of the language levels in Guide
document is misleading/backwards (not consistently addititive or
diminuitive)
- SUGGESTION FOR GUIDE:
needs list of interdefinability between language items
- SUGGESTION FOR GUIDE:
needs difference-list with DAML+OIL
- SUGGESTION FOR GUIDE: explain what it takes to turn an RDF Schema
document into an OWL Lite document
- SUGGESTION FOR GUIDE: explain how to say that two classes are
different
Received on Thursday, 9 January 2003 08:46:10 UTC