Let’s take an example: titles. How does one distinguish a uniform title, from an abbrviated title, from a trnasliterated title in a relatively “simple” format like MODS? You have one complex XML structure, with four or five elements, and variety of attributes. That complex XML structure is designed to basically say “all of these are tiles.”

OK, but when you deal with complex data, you end up with quite complex XML; data which is simply difficult to work with.

What if instead you just gave each separate URIs …

http:loc.gov/standard/mng/title http:loc.gov/standard/mng/abbreviatedTitle http:loc.gov/standard/mng/transliteratedTitle

[let's not worry about details, because the above is likely not quite right; but I'm sure you get the idea]

In RDF/XML, you then end up with three different elements You define the relations between them (that the last two are sub-properties of the first) in a separate RDF/OWL schema.

In essence, you separate out the concerns (encoding and semantics).

Another example is the constant use of the uncontrolled authority attribute in MODS and MADS. Again, more complicated (and losely-controlled) XML to do work that can be done more easily and precisely. We live in the 21st century, with a huge HTTP and URI infrastructure; so use it.

<mng:subject rdf:resource=”http://loc.gov/topics/x”/>

Did I answer your question?

Bruce

I’m not following how you do this.