Standard IVORNs, data model identifiers

[Norman Gray]

Mark and all, hello.

On 2014 Feb 18, at 17:47, Mark Taylor <m.b.taylor at bristol.ac.uk> wrote:

> I have nothing elegant to offer I'm afraid, but the idea of a new ID
> for every minor standard version sounds unpalatable.  In practice
> I find that for most of the versioned/versionable things I access,
> it's more important to be able to identify the thing with
> (the version-free Platonic ideal of) the standard than to know
> exactly which version I'm talking to.  So unless somebody argues
> persuasively otherwise, I'd vote for (a).

If record proliferation, and identification, are taken to be distinct issues, then a resolution might be (to resolve) to indicate the version-free standard with an unversioned URI, and indicate a version in a fragment.  Since any reasonable URI-parsing library will have a way of easily extracting the non-fragment parts of the URI, you retain the option of comparing URIs in the way you want, without discarding the option of identifying individual versions where necessary.

This does mean having more registry records; is that really a challenge?

Non-issues (which I mention in my self-appointed role as standards lawyer):

If one wanted to express this in terms of URI resources/subresources, and all that, then I think it's perfectly reasonable to conceive of the versionless standard as being a 'resource', and individual versions being 'subresources', but I don't see this as being an important thing.

The URI spec has a section on URI equivalence (RFC 3986, Sect. 6.2), which also doesn't add anything significant to this question.  It ends up being mostly about URI normalisation, but notes in passing that two URIs with different fragments will always be deemed inequivalent by that spec.

All the best,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK