Names for the basic elements of a profile

[kcoyle]

One of the big questions is what terms we will use for the basic elements of vocabulary. The DSP uses:

Description Set Description Statement

While BIBFRAME uses:

Profile Resource Property

We also will need a name for the whole, and because it will not exactly follow the DSP it probably should be given a new name. All suggestions welcome!

[tombaker]

Actually, Description Set, Description, and Statement were defined not in DSP itself, but in DCAM. DSP, then, used these for defining Description Set Template (a cookie-cutter model for a Description Set), Description Template, and Statement Template. DSP as a whole was referred to as a description set profile constraint language. In other words, DCAM provided the abstract syntax for a constraint language.

Without looking at the BIBFRAME definitions, it looks to me like the BIBFRAME terminology is not quite aligned with the DCAM/DSP terminology:

• in DCAM/DSP, a Description Set Profile is a set of templates holding constraints - not just constraints on descriptions and statements, but also on the description set as a whole (e.g., the description set holds the description of a book plus one or more optional descriptions of authors). The DCAM/DSP notion of Description Set Profile is perhaps equivalent to BIBFRAME's Profile.
• In DCAM/DSP, a Description is a set of statements about exactly one Resource (in the RDF sense). Resource is already pretty well-defined in RDF as any thing that can have identity. Repurposing the term Resource in BIBFRAME to denote not just the thing itself, but also the description of that thing, looks like conceptual shorthand.
• In DCAM/DSP, a statement is not just a property, but a property-value pair (the subject of which is the subject of the encompassing Description). I'm guessing that what BIBFRAME calls Property actually refers both to Property and to Value - or rather, to the set of constraints on the property and on the value. As with Resource, Property already has a well-established meaning in RDF, so repurposing the term to refer, in effect, to "constraints on the usage of a property and its corresponding values" does not seem like a good idea.

I am assuming that the underlying models for DCAM/DSP and BIBFRAME are pretty much the same. If so, then I could live with (BIBFRAME) Profile as the equivalent of Description Set Profile, but I suspect that using Resource and Property to denote sets of constraints could actually make it more difficult for people to get their heads around RDF.

[nishad]

Description Set > Description > Statement > Constraints is a straightforward approach and self explainable, BIBFRAME vocabulary is rather confusing.

[kcoyle]

The problem that we run into with "Description" is that we also want a "description" property for each entity and each property. Most folks use dct:description for this. So we have two uses of description that get easily confused (Description and dct:description). I like Description Set a lot; I just wish we could easily disambiguate between the entity "Description" and the description of things within the metadata. Any ideas on that would be really helpful. Calling "Description" a "resource" (or some other term meaning "entity" or "thing being described") would really help here.

The other thing about description comes about because the library world uses the phrase "descriptive cataloging" for what they do when they create records in the catalog. To them, the description is the whole: author, title, publication information, physical characteristics, subjects. In DSP terms that would entail multiple descriptions and would be an instance based on a Description Set. We could possibly exploit this not-quite-shared use of description (but I don't have a good idea about that at the moment).

[tombaker]

I do not like the idea of mixing signifier (e.g., Description Set, Statement) and signified (Resource) in the same framework. If Description has too much distracting baggage, could it be qualified? Resource Description comes to mind, but that has its own baggage, starting with RDF and RDA! Single-Resource Description would be accurate, but it's too long. Using Description, but with a qualifier, would solve the problem - but I cannot think of any good qualifiers. Description Set => Entity Description => Statement?

[nichtich]

what terms we will use for the basic elements of vocabulary

What's wrong with starting with "basic elements" and differentiate later?

Moreover to me it looks like "constraint" (in a broad sense) is the first kind of element.

[tombaker]

Moreover to me it looks like "constraint" (in a broad sense) is the first kind of element.

@nichtich makes an interesting point. As I explained in more detail in another post, the 2008 draft Description Profiles constraint language has just templates and constraints. Analogously, ShEx has shapes and constraints. In both cases, constraints are indeed the "first kind of element".

The next question is: templates or shapes of what? And that is where we'd like to have the vocabulary of description set => description (or whatever) => statement. For DSP, those categories were defined in a DCMI Abstract Model.

[tombaker]

@kcoyle We had some good discussion here, but can we close this issue (at least for now)?

[kcoyle]

Done dctap