Should PROF roles be in a their own namespace?

[aisaac]

Currently PROF roles are defined in their own namespace, separate from the "base" one for PROF. In w3c/dxwg#535 and w3c/dxwg#536 points have been made for and against having them in separate namespace. But that discussion was mixed with the discussion on whether they should be documented in the same document as PROF, which is slightly different. In the end I am not sure there has been a definitive argument for keeping roles in the same namespace. So I would like the WG to re-examine this. Maybe to there has been progress on the front of possible W3C registry (such registry was envisioned at one point as a possible place for a role vocabulary).

In the meantime I think there should be a note acknowledging the issue in the role section of the PROF spec, asking for commenters to give feedback.

NB: I believe this is not a blocking issue for moving to CR.

[nicholascar]

I support them being in their own namespace as they currently are so that it may be managed separately from the main PROF namespace.

PROF's namespace was specifically set up to be a slash URI (as opposed to the more common, for W3C ontologies, hash URI) to allow for derivative, separately managed, namespaces e.g. /prof/role/ which is currently used for the roles.

[nicholascar]

@aisaac Since they are currently in their own namespace, I've added the following sentence to the section you suggest:

Note that these Roles are defined within a namespace that is derived from, but separate to, the namespace for PROF. This namespace is: http://www.w3.org/ns/dx/prof/ with the suggested RDF prefix of role.

Can you indicate whether this was the sort of note you had in mind?

[kcoyle]

I've seen some discussion on the w3c github about a registry and there isn't a decision yet, so we can't lean on that. (Actually, their idea of a registry is very different from what I've always thought of and what little I've participated hasn't gotten any take-up from others.)

[aisaac]

@nicholascar this goes in the direction I was envisioning, certainly. But if we want to invite feedback, now or later, and acknowledge the uncertainty of the context of registry, the note should probably end with "This situation may however evolve, should this be needed (for example resulting from changes in the W3C approach to vocabulary management)."

[rob-metalinkage]

I dont think we need to concern ourselves with a "registry" yet. Having a set of roles in a namespace that a community group can extend it easily - as well as baking in a set of uncontroversial ones into the core - covers all the bases and shows a way forward. (We'd probably just end up in pain trying to define what a registry was anyway)

[aisaac]

@rob-metalinkage indeed my suggested addition does not try to define what a registry is or would be. It doesn't even mention the name.

As I've said earlier, there are pros for the current solution (like yours), there are cons (like yours), we've not completely sorted it out (i.e. I guess there was no vote from the WG on this) and we feel that the context could bring drastic changes to the discussion. None of this has changed in the past 6 months, so it feels safe to add a little a flag that we can wave in case some (including ourselves) make comments that we would have overlooked the issue.