RE: [tm-pubsubj] Requirements oddity

["Steve Pepper" <pepper@ontopia.net>]

* Patrick Durusau:

| > | <rdf:RDF
| > |      xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#";
| > |      xmlns:dc="http://purl.org/dc/elements/1.1/"/>
| > |     <rdf:Description
| > |         rdf:about="http://psi.fruits.org/apple.html";
| > |         dc:publisher="Johnny Appleseed"
| > |         dc:description="Apple: Round firm fleshy fruit of a rosaceous
tree"
| > |         dc:type="Published Subject Indicator"
| > |         dc:date="2003-09-14" />
| > | </rdf:RDF>
| >
| > This is fine as it is ... except that dc:description is misplaced.
| > You have provided a description of the subject itself, not of the
| > subject indicator (or identifier).
|
| I don't think "description of the subject itself" is quite right.

I agree.

| Dublin Core defines description as: "An account of the content of the
| resource."

So your use of dc:description is inappropriate, right?

A more appropriate "dc:description" might be

  dc:description="Published subject indicator for 'apple'"

| On the other hand, dc:type, which is defined as: "The nature or genre of
| the content of the resource." does seem to fit rather well.

I agree. In fact if you added

  dc:subject="apple"

then the combination of dc:type and dc:subject would render
dc:description superfluous...


| > I still don't understand why software needs metadata about identifiers
| > rather than indicators.
| >
|
| Hmmm, I suppose under the theory that each identifier is bound to an
| indicator (requirement for being a PSI), then it may be sufficient for
| the metadata to be "about" the indicator. It is by implication also
| about the identifier in some sense but it may not be necessary to go there.

I really do think that the identifier is an attribute (locator) of
the indicator and therefore in some sense "inherits" some of its
metadata. I don't see any need to have additional metadata about
the locator, especially given that metadata about an information
resource (e.g., a subject indicator) is a well understood and widely
used concept, whereas metadata about a locator is definitely not...

| What I was envisioning was a circumstance where I have software that
| trusts some identifiers and not others. The bare identifier, without
| more, does not give me any basis on which to distinguish those I trust
| (apart from domain name) from those I don't. Having the metadata in the
| Published Subject Indicator be "about" the identifier, I can distinguish
| (assuming required) on the basis of date, publisher, other required
| metadata? I might trust identifiers from Idealliance but only those with
| dc:creator of "Jane Harnad."

But where would you go to find that metadata? You would have to
resolve the identifier to the indicator. Once you have that, you
might as well use the indicator's metadata, which will tell you
who created the indicator and - by extension - who assigned the
identifier to it.

| As I noted, it may be a distinction without any real difference. I could
| treat the metadata in the Published Subject Indicator as determining
| whether I trust the Published Subject Identifier.

As should be clear by now, that is exactly what I think you should
do!

Steve

--
Steve Pepper <pepper@ontopia.net>
Chief Executive Officer, Ontopia
Convenor, ISO/IEC JTC 1/SC 34/WG 3
Editor, XTM (XML Topic Maps 1.0)