From: Hart, Lewis (lhart@grci.com)
Date: 08/30/01
John - Here are some general comments: 1. I would change the symbols indicating Classes (green) and Properties (blue) slightly. First, I would remove the line between multiple types. (e.g. shirtsize) Then every thing above the line is a type, and the ID is below. I would also use different shapes instead of color to differentiate Classes and Properties. Not every one has a color printer and some folks have color blindness. This is not too important for 'first order' entities that have "Class" or "Property" in in there type names, but second order entities could be confusing. For example, you can not tell directly that a Restriction is a Class. 2. Literals and Instances need to be differentiated. "short", "medium", and "tall" are instances of the Class Height, but "1" and "2" are untyped Literals. They should use different symbols. Also, instances need to indicate what class/property they are an instances of and perhaps use different symbols for instances of classes vs instances of properties. (Maybe dashed versions of the Class and Property symbols?) 3. I did not see the instance of a typed literal (e.g. <xsd:integer rdf:value="13"/> ) in your diagram. You need a symbol for these. Maybe the symbol for Literal with a type and separator line like Class and Property. 4. I also did not see a symbol for the ontology itself anywhere, but you will need to differentiate ontologies if there are multiple ones being used. 5. Multiple ontologies and namespaces need to be handled better. How does the viewer know what ontology/namespace "over12" comes from? How do I know that "over12" is not part of XSD? For that matter, how do I know what XSD: is? I would recommend that: - full URIs be used to identify "external" elements, or something analogous to xmlns:xsd ="http://www.w3.org/2000/10/XMLSchema#" be available. - the term "URI" in the type slot be replaced with the actual type (and use the correct shaped symbol) or left blank for unknown and - dashed lines be omitted, just use the regular symbols. -----Original Message----- From: John Flynn [ <mailto:jflynn@bbn.com> mailto:jflynn@bbn.com] Sent: Wednesday, August 29, 2001 9:48 AM To: <mailto:daml-graphical@daml.org> daml-graphical@daml.org Subject: VisioDAML >I have created a Visio template that includes a stencil with graphical representations of all DAML+OIL language >constructs. You can drag and drop the various DAML+OIL constructs onto the drawing page to create the representation >of an ontology. Its initial purpose is to provide a relatively easy means to graphically represent DAML ontologies. >I will post it on daml.org for download soon, but first I wanted to get some comments on some of the graphical >conventions. A JPEG file of the VisioDAML drawing of the DAML+OIL revised example ontology (less instances) is >attached. > >The example ontology is at: <http://www.daml.org/2001/03/daml+oil-ex.daml> http://www.daml.org/2001/03/daml+oil-ex.daml > >Some specific questions: Should the inverseOf and disjointWith connectors have arrow-points at both ends as shown in >the VisioDAML drawing? Should any of the other connectors have arrow-points at both ends? No, to both. I would have the arrows represent the triple that would be in RDF, with the arrow head on the object end. I believe you intend the diagrams as a representation of the DAML syntax, so semantics should not be shown. > >Also, the dashed-line boxes labeled URI represent references in the example ontology to data types that are remotely >declared. On the VisioDAML drawing when you mouse over the dashed-line box a comment appears showing the actual URI. >Any better ideas on how to represent these remote declarations? See above (#5). > >Any other comments on the drawing conventions are most welcome. > >Note that the graphical representation of the DAML+OIL revised example ontology makes readily apparent some of the >errors in the example. The Class FullTimeOccupation and the ObjectProperties, hasOccupation and hasSpouse, are >colored red in the VisioDAML drawing to illustrate they are not declared anywhere in the example ontology. Also, the >TransitiveProperties, has Ancestor and descendant, as well as the DataTypeProperty shirtsize, can be seen at the >right side of the drawing and are not connected to any other part of the example ontology. > >John > >John Flynn > >(703) 284-4612 > >DAML Integration and Transition PM > >BBN Technologies > > - Lewis ___________________________________________ Lewis L Hart GRC International lhart@grci.com 1900 Gallows Rd. Voice (703)506-5938 Vienna, Va 22182 Fax (703)556-4261
This archive was generated by hypermail 2.1.4 : 03/26/02 EST