On Wed, 28 May 1997, Brian Kelly wrote:
> Dan
> Thanks for this message.
> I must confess to a certain confusion over how W3C sees the metadata
> architecture developing after the W3C sessions at Rutherford Labs in
> January and WWW6 in April.
> There has already been some breakdown in communications over the SCHEME
> proposal and the PICS WG. I am worried that once the other proposals
> related to metadata get to the stage of being discussed by W3C Working
> Groups we'll see similar confusion - potentially resulting in a lot of
> wasted time and effort by those involved in producing various proposals.
We definitely should strengthen communication links if they are
not yet well established.
> I'm also worried that even if the proposals are accepted, as there are
> potentially several ways of achieving the same effect, software vendors
> will have the final say in deciding the winner.
Hopefully, it will be software vendors and the user community.
> For example if I have a resource available in HTML 2.0, HTML 3.2, HTML
> with frames, PDF and English and French versions, I might want to use
> web collections (so that a user agent can give a site view of all English
> languge resources), the SCHEME proposal for embedding DC, TCN so that
> browsers can transparentally choose an appropriate format, DSig (so that
> I can assert that the document is a conforming HTML 3.2 document which
> will run on a browser without any plugins on a 8 MB 486 PC) and PICS
> (so that I can say it's contains no nudity).
This example very nicely shows that for many different reasons,
and at many different levels, there are various equally valid
approaches to metadata, and while we can expect that a few of
them will cease to exist, there is clearly not only one that
will persist. For example, we need various places where metadata
exist, we need various schemes for metadata to cope with the
variety of existing practices and to avoid information loss,
and so on.
> >From a metadata management perspective this seems a nightmare.
>
> Comments?
To a certain extent, this can't be helped. With good communication,
we can make sure that we converge instead of diverge in those
places where convergence is possible. But there is already
divergence in the bulk of existing metadata (just have a look at
all the variants of USMARC,...), and the web adds new requirements
(never had to indicate screen resolution and color capabilities
when lending a book at the library :-).
Regards, Martin.
|