About the model: Question 9
9. There are times when you have an expression of an expression (as when a specific edition is the basis for a new edition, or when there are two Japanese translations, etc.) or a manifestation of a manifestion (as when a reproduction is made of a specific other manifestation). If I put in subexpressions and submanifestations as subclasses does that somehow require every resource to use all of these levels even when they are not necessary? Is there another way to deal with a hierarchy that needs to be expansible and contractible in RDF?
“I would just reuse expression and manifestation classes ; they are “sub” by inheritance. I usually point to the difference between ; <topic> <topicName> some name </topicName> </topic>and <topic> <name> some name </name> </topic>Here, classes and instances are described in XML, but the point is that we can reuse <name> at will, as we certainly will define what “name” mean somewhere. In this case, “name” is a sub-element of “topic”, so the topic’s name. This is straight class-instance theory, and doesn’t need a lot of explaining. The former example needs an explanation of what “topicName” mean, and especially if it means anything different from a normal “name”.So, the same with “expression” and subexpression” ; we can already see that the one is a sub of the other, so does “subexpression” express more than just “sub”? In my eyes, the less “things” you need to define and the fewer relationships you need to explain, the easier it will become to express and explain them. Maybe these are too simplistic examples?”–September 11, 2007 email from Alexander Johannesen.
DECISION: Rather than define subexpression and submanifestation class levels, I defined expression-to-expression relationship properties.