On Jan 17, 2006, at 11:00 AM, Florian Reuter wrote:
>> The complexity of the metadata issue is really just in the details.
>> They are not hard to work out though.
>
> You would be surprised how complex the issue of metadata can be in
> case of OpenDocument. E.g. I recently talked to philologist
> researchers, who e.g. would like to annotate greek texts with meta
> data. We had a short discussion about meta data support in
> OpenDocument, which is different to your requirements. What I learned
> from that is, that we definitely need to do some fundamental research.
> As a researcher I'm sure you agree, that a broader approach is called
> for, given the depth of the issue. We should not rush this and focus
> on special cases (e.g. bibliography). I also believe that users of
> OpenDocument will appreciate this.
Sigh ...
With all due respect Florian, I have always been of the opinion that we
ought to do this carefully, and do this right. I have not said that
annotation is not worth considering, but it is yet another issue
separate from the more basic issues of how to allow extensible metadata
description, and how to use that to describe the content of document
parts.
I find it VERY hard to believe that any researcher would consider
detailed annotation support in OpenDocument more important than
flexible document level and bibliographic (and other sub-document
metadata) support. One of the reasons LaTeX was a big success in many
research fields was precisely because of BibTeX.
In any case, good annotation support presupposes good general metadata
support. So why can't we proceed on that basis?
If you insist on solving ALL of this at one time -- including the
annotation stuff -- you will spend the next few years doing it. It has
been a subject that the DocBook people have been looking at for years,
and they only just settled on a solution a few months ago [1].
The alternative is that we at the OpenOffice bibliographic project will
just go off and write our own metadata support that will be
incompatible with the current document metadata support, and will not
be standardized. If someone decides to implement a bibliographic module
in KOffice, they will do the same thing. Does anyone really believe
that's a good idea?
We need this support, and it needs to be standardized.
I know in general companies don't care much about higher ed, but I'm
giving you all a real use case, a way to separate out the different
issues (description, storage, annotation), and I'm even giving you a
reasonable way to address it in a way not limited to this use case.
What is so hard about agreeing we ought to allow extensible metadata
descriptions in the file wrapper?
What I object to is not that people recognize it's a complicated issue
that bears further research and discussion, but that Michael has
proposed to stop the conversation, except in the context of a
face-to-face that I cannot reasonably attend.
Bruce
[1] http://norman.walsh.name/2005/10/17/annotations
