Re: [office-metadata] Groups - Metadata_Model_Proposal_10May2007 (07-05-10-ODF-Metadaten_pld.odt) uploaded

[marbux <marbux@gmail.com>]

Thanks for the enlightenment, Bruce.

When the finished Metadata proposal goes to the TC, I hope that it
might be accompanied by a discussion of what changes to the
conformance section will be needed to adequately support the Metadata
SC's work product. We have major work to do on the conformance section
that has the potential to affect many portions of the specification.

Best regards,

Marbux

On 5/11/07, Bruce D'Arcus <bdarcus@gmail.com> wrote:
>
> On May 11, 2007, at 10:07 AM, marbux wrote:
>
> > The fundamental issue raised by the XML Unique IDs proposal is whether
> > conformant applications will be required to preserve RDF metadata
> > inserted by other applications, including XML Unique IDs.
>
> We have talked about some of this previously, but we should clarify
> the details with the TC.
>
> Preserving xml:ids per se -- both the attribute and its value -- is
> not per se important because of the binding abstraction we use. What
> is critical is that metadata files be preserved (including
> manifest.rdf) and that the binding between xml:ids and IRIs also be
> preserved and maintained.
>
> [for background, it is possible that an implementation will decide to
> rewrite xml:id values at times; that's fine so long as they update
> the binding too; deleting xml:id attributes would be bad though,
> because the bindings would break]
>
> But as Michael pointed out, there are some tricky details. What
> happens if a user deletes content with some metadata fields. Do we
> require the ODF application to preserve the original field metadata?
> I sure hope not.
>
> So the question is how to write this into the spec as precisely as
> possible, and we have tried to do that. But it's definitely something
> the engineers on this committee should pay close attention to, and
> make sure we cover it once it gets to the TC for discussion (very
> soon hopefully!).
>
> Bruce
>
>