Learning objects

["Butler, Mark" <Ma...@hplb.hpl.hp.com>]

Hi team,

RE: Learning objects

Just noticed this thread on learning objects, and it turns out I have been
looking at this area for a while now in my day job. So here is some relevant
background, although obviously you need to decide how relevant this is to
Cocoon. 

The main point about "learning objects" is they are packages, e.g. a set of
HTML pages and images rather than a single page. The relationship between
the pages is important. As they are packages they are transportable, unlike
web pages where it is not clear what resources are required. Also people are
interested in reusing learning objects, so there are some metadata schemas
that facilitates learning object re-use by providing metadata about the
learning object. For an introduction see

"A Primer on Learning Objects." Warren Longmire. 2000. 
http://www.learningcircuits.org/mar2000/primer.html

"All about learning objects." 
http://www.eduworks.com/LOTT/tutorial/learningobjects.html

More Background on IMS (a metadata format for LO's) can be found here
http://www.imsglobal.org

Two example projects using LO's are
http://belle.netera.ca/about.htm
http://careo.ucalgary.ca/

RE: Complexity of Protege

Personally I think Protege is probably the best open-source tool for doing
ontology creation. Ontologies (in the RDF, OWL sense) are a bit of a
buzzword, but the key point is tools like Protege require you to make an
explicit distinction between classes and properties, unlike data formats
like XML where the distinction is implicit. Protege is now a tool adopted by
people working on the Semantic Web, but it predates it and is therefore a
lot more user focused than lots of tools that have originated since the
introduction of data formats based on RDF e.g. IsaViz. For a good
introduction to Protege, see

"Ontology Development 101: A guide to creating your first ontology." Natalya
F. Noy and Deborah L. McGuinness.
http://protege.stanford.edu/publications/ontology_development/ontology101-no
y-mcguinness.html

RE: Ontologies versus thesauri

Thesauri and ontologies actually have a lot in common. Ontologies, apart
from describing classes and properties, also describe their relationships a
way that allows certain kinds of inferencing (machine processing). Thesauri
on the other hand deal with relationships between terms, and there is a
correspondance between these terms and classes used to represent property
values in ontologies. However, in thesauri, the relationships between the
terms are less formally defined, and hence less amenable to inferencing, but
conversly seem easier to understand e.g. narrower term versus subproperty. 

For an excellent discussion of the relationship between ontologies and
thesauri, see
http://jodi.ecs.soton.ac.uk/Articles/v01/i08/Doerr

kind regards,

Mark Butler