Thursday, March 31, 2005

Lecture Thursday 31 March 2005

Handbook of ontologies (Springer)

Ontology: defines the basic stuff of the world
not just terms but objects/concepts and their relations

relationships:
classification
part/whole
material/made of
tool/function
domain-specific relations

Thursday, March 24, 2005

Lecture Thursday 24 March 2005

Thursday, 3/24/05: Classification, cont.

* Parsons, J. & Wand, Y. (1997). Choosing classes in conceptual modeling. Communications of the ACM 40(6), 63-69.
* Priss, U. & Jacob, E. (1999). Utilizing faceted structures for information systems design. Proceedings of the 62nd Annual Meeting of the American Society for Information Science, 203-212.

Concern with coding hierarchy: notation, a remnant of pre-computing and early computing systems where memory was a concern

facet: fabric


notation
order in an array--in which order would you like a set of facets (how to order types of fabrics)?
schedule order--in what order are the facets themselves (e.g., chief complaint facets, such as body system, then type of pain, etc.)
citation order
cotton green shirt sounds funny
but green cotton shirt sounds normal




version of ICD-9

how do you order nominal data?

Tuesday, March 22, 2005

Lecture Tuesday 22 March 2005

Classification made simple


things to take out of your house in a fire
Do we have a word for this?
No.
But we have a word for "cat" representing things that are cat.

This is an ad hoc category: highly situational, improvised as necessary

What is an essential difference between an ad hoc category and a named one?

What I would remove:
family & pets
fire box
computer
old tax records
pieces of art
photo albums



we also have:
- scientific/expert classes
- folk/everyday classes
- class by definition (arbitrary)

coherence/convergence - ad hoc classes are difficult to converge

defining a class
----------------
extensional - listing all members
intensional - listing essential properties/conditions that members must meet (the rules)
subjective
intensional -



mutually exclusive classes are wonderful but really don't happen very much in reality or if they do they're quite unwieldly

exhaustive
classes are all-inclusive; the category "other" contitutes a failure

hierarchical classifications are too rigid; often classifications depend upon multiple layers of characteristics

Thursday, March 10, 2005

Lecture Thursday 10 March 2005

rules logic proof and trust

trust seems a rather insidious requirement

can you actualy make trust an IT implementation.

Washington site shows value of metadata


rules
logic framework
trusted systems

the gettier proposition is true by justification of something in the present text
but it's not the right justification
and so the truth is accidental and therefore there's no knowedge


but these systems could have the social function of insisting propositions qualify for knowledge


computer game world: everything is there
social function

completeness of a game, a premise of fiction
when the premise is lost is where the danger ensues
but the web makes no such assumption

re: banking industry and computer reality


what the representations are
can we manipulate them
yes, but we need to maintain the premise of fiction

services: going beyond the resource discovery....

agency/confidence in the service

we do operate without the certitude that truth has
but we do have notions of reliability based on experience

when it comes to business you either never find out or its too late

qualification: the semantic web needs qualification

greater data reuse for purposes not originally conceived

existential question: when computers are defining a foul reality
similar to a deity defining a foul reality for a protagonist

language is somehow anarchic
notions of centralization of these language
functions is centralizing the power
and power is dangerous

starts with writing
dictionaries
semantic web

Tuesday, March 08, 2005

Lecture Tuesday 08 March 2005

For Thursday's class

1. Reflection paper due
2. What components of the semantic web do we have now? Which ones are on the way? Which are out of reach?

Thursday, March 03, 2005

Lecture Thursday 03 March 2005

Dublin Core!
http://dublincore.org/documents/dces/
http://dublincore.org/documents/dcmi-type-vocabulary/

What do you need to add to a description to distinguish items?

Dublin core: here are tools you have to point out similarities & differences
dimensions of description
value fields are dimensions

delivery mechanism?
header & file



html: will it stick around for longer than ever imagined?

semantic web