Hello,
I would like to issue a call to those who disagree (Aaron I know you disagree so
you're already on the list) with the pattern for a DCMI namespace described in:
http://dublincore.org/documents/2001/03/09/dcmi-namespace/
to offer feedback with respect to alternate patterns.
I would like to determine:
(NOTE: assume that http://purl.org/elements/1.1/ will remain unaltered).
1) How many applications/implementors are impacted by the proposed pattern (for
the 2 approved sets (dcq and dctypes) and any furture sets (e.g., dc-ed))?
(NOTE: please try to include implementations that may not have a representative
monitoring this list)
2) Who feels that purls should be used instead of dublincore.org URLs (or both)?
3) Who feels that a different arbitrary number (by different I mean not the
creation date of the vocabulary package) should be used (e.g., qualifiers 1.0,
dcq 1, ...) as a versioning string?
Please if these are issues we need to hear about them. Please see Aarron's note
below he has provided a summary of some of the stonger canidates we have
consisdered.
Cheers Tod Matola
DCMI/OCLC Office of Research
[log in to unmask]
Aaron Swartz wrote:
> Matola,Tod <[log in to unmask]> wrote:
>
> > This revised draft is available for 1 week of review and comment. After
> > which the revised draft will be submitted to a expert committee (selected by
> > the DCMI Directorate) for final review.
>
> Will the comments on the choice of namespace not be heard? Will the Working
> Group be allowed to participate in the selection of the expert committee?
> Can the committee be given several choices? I'd like to suggest the
> following alternate proposals be presented:
>
> 1) http://purl.org/dc/package/version/
> Ex: http://purl.org/dc/qualifiers/1.0/
>
> This also has the added benefit of being previously endorsed by the Dublin
> Core website (see
> http://www.google.com/search?q=cache:www.purl.org/dc/qualifiers/1.0/ for
> evidence), as well as used in a number of Dublin Core files (and possibly
> applications). I think that it is wrong of Dublin Core to renege on it's
> agreement to metadata creators. A quick Google search shows over 125 pages
> use this namespace, including several Dublin Core Working Drafts, and the
> CellML specification.
>
> There is little in the way of good reason to change these URIs, and a number
> of good reasons to keep them. If this is a demonstration of the persistence
> of Dublin Core identifiers, it does not look very good.
>
> 2) http://purl.org/dc/package/id#
> Ex: http://purl.org/dc/qualifiers/1#
>
> 3) http://dublincore.org/package/id#
> Ex: http://dublincore.org/qualifiers/1#
>
> 4) http://dublincore.org/ns/package/id#
> Ex: http://dublincore.org/ns/qualifiers/1#
>
> While I believe number 1 is the best choice, I'd be willing to live with any
> of the others (or minor modifications of them) as a replacement to the
> namespace currently proposed.
>
> Other benefits of these alternate choices:
>
> - They are compatible with the "examplotron-style" of namespaces
> proposed on XML-DEV. (see http://xmlns.com/ for similar usage)
>
> - They are much easier to parse/understand.
>
> - They can be decomposed to find the latest version, and other
> information.
>
> Ex: http://purl.org/dc/qualifiers/1.0/ -> http:/purl.org/dc/qualifiers/
> which has pointers to later/older versions of qualifiers, etc.
>
> I hope that the Dublin Core will reconsider this decision before making an
> awful mistake and bringing confusion and complication to the world of Dublin
> Core users.
>
> --
> [ Aaron Swartz | [log in to unmask] | http://www.aaronsw.com ]
|