Re: PROFILE as HEAD attribute

Andrew Daviel (andrew@andrew.triumf.ca)
Tue, 15 Apr 1997 18:22:10 -0700 (PDT)


Date: Tue, 15 Apr 1997 18:22:10 -0700 (PDT)
From: Andrew Daviel <andrew@andrew.triumf.ca>
To: Paul Miller <apm9@york.ac.uk>,
cc: www-html@w3.org
Subject: Re: PROFILE as HEAD attribute
In-Reply-To: <335406E7.D81F443@w3.org>
Message-ID: <Pine.LNX.3.91.970415174956.13696L-100000@andrew.triumf.ca>

On Tue, 15 Apr 1997, Dan Connolly wrote:

> Paul Miller wrote:
> 
> > Presumably, the DC. prefix model suggested by many for the Dublin Core
> > could be called into service here, giving you
> > <META NAME="ADS.creator"
> > There are issues here of registry, as with SCHEME names, in order to avoid
> > 'ADS' being 'Archaeology Data Service' AND 'Alberta Dustbin Systems' or

> is needed. Or to look at it another way: the registry is the web.
> excerpt from
> http://www.w3.org/pub/WWW/Search/9605-Indexing-Workshop/ReportOutcomes/S6Group2.html
> <LINK REL = SCHEMA.dc HREF = "http://purl.org/metadata/dublin_core">

In the current absence of a defined machine-readable way of declaring a 
schema, this means in effect that a resource discovery agent must tag 
each metadata item with the URL of the schema. Since the same prefix may be 
used for a different schema in other documents, it is not sufficient to
store the schemas independantly. So effectively I have a data item with name
"DC.creator;http://purl.org/metadata/dublin_core_elements" and value 
"J. Fish". Then I have to fold URL aliases, such as "..re_elements/" and
"..re_elements", "http://purl.ORG..", etc. and some people may leave the 
schema off altogether ... Oh well; disks are cheap ...
 

Minimalist Musings:

If DC is supposed to be a "core" set of common metadata (which I 
understood it was), I'd like to fold the discovered common
<meta name="author"> into <meta name="DC.creator">
and <meta name="description"> into <meta name="DC.description">.
In fact, I'd really like to drop the "DC." for DC metadata altogether in the 
interests of simplicity, since we have "scheme" and type suffixes where
a restricted vocabulary is required.

Of course, if DC is not considered an "official" metadata set, but just 
one of an infinite variety of possible sets, that's different.

Andrew Daviel

(( Have I misunderstood here ? Although the DC was dropped
from http://purl.org/metadata/dublin_core_elements I presumed that
that page merely listed the elements in a format-independant way and
that the DC was required in HTML. At least, on even-numbered days I did 
:-} ))

((( Is XML or PICS-NG going to supersede HTML before DC is finalised ? 
Are we wasting our time ? Is DC going to be incorporated in XML ... ? )))