[seek-kr-sms] ontology/folder UI design in Kepler

[Ferdinando Villa]

Hi folks,

reading through the chat I was wondering that maybe something like
Serguei's idea of 'graphic extension' may be helpful here. Graphic
extension (I'm sure there's a better name) associate a fragment of
knowledge from an ontology with a particular display mode, graphic
symbol or whatever - e.g. different-looking arrows, folder icons, or
stock/flow symbols a' la Stella. Used within an interface that's capable
of handling them, they can be used to interface a conceptual model
tailored to a specific user to the actual ontology it uses, hiding its
complexities. Serguei has put some examples online, see e.g. 

http://ecoinformatics.uvm.edu/dmaps/tut42.html
http://ecoinformatics.uvm.edu/dmaps/tut43.html

It's something we're planning to integrate into GrOWL, to build
generalized modeling environments where after you load the mapping (as
an OWL file) you can use domain-specific icons and create instances
whose actual OWL definition is more complex than it looks. Users see an
intuitive/familiar graphical depiction of their domains, but the machine
thinks OWL internally. E.g. this can be used to build Stella on top of
GrOWL, and have it output full OWL specifications of the models. I'm
leaving for a few days and writing this in a hurry - anyway, any chance
that this concept (as is or used with GrOWL) can help integrating KR
into Kepler? It could be used, for example, to help users define new
concepts that extend others in ways more complex than simple
subsumption, without having to properly link them to an ontology that
they have no chance of understanding well enough to do it right.

ciao, 
ferdinando


On Fri, 2005-02-25 at 19:43, Matt Jones wrote:
> Messed up the list address ... sending again... sorry for the duplicate 
> laura.
> 
> Matt
> 
> -------- Original message ------------------
> Hi Laura,
> 
> Chad, Shawn, and I have been discussing the tree control on the left
> side of the Kepler window.  Chad and Shawn are currently modifying it to
> allow creation of new nodes in the ontology.  In doing so, we've
> recognized a bunch of problems with using the informal 'folder' concept
> to represent an ontology class. The Growl editor is a more acurate
> representation of the ontology but has limitations given the small space
> we are trying to use in Kepler to discuss this stuff.  We need to
> resolve this to make the UI clear and not have the users mess up the
> ontology as they add new concepts and actors.  The attached log of our
> IRC conversation describes some of the issues.  Could you review it, and
> then we can discuss how to reach our goal.  I think its important to
> have you involved at this design stage.  Chad and Shawn are implementing
> now so we need to provide feedback asap.  Maybe we can have a conference
> call next week?  Thanks.
> 
> Matt
-- 
Ferdinando Villa, Ph.D., Associate Research Professor, Ecoinformatics
Gund Institute for Ecological Economics and Dept of Botany, Univ. of Vermont
http://ecoinformatics.uvm.edu