Rachel Heery wrote:
>
> Can you clarify what RDF expects there to be at the namespace URL, in
> terms of what data is there? does there have to be anything there in
> reality? Is this URL just a label, or is the implication that there is
> some sort of machine readable DC definition there? Of course I can see it
> would be useful if there was something there, particularly if human
> readable, but does there have to be?
As you suggest, there is no requirement that there be anything at all at
that URL. XML namespaces are simply a mechanism for creating globally
unique names. So if we write, say:
<?XML:namespace href="http://purl.oclc.org/dublincore" as="DC"?>
<DC:Title>Hello world</DC:Title>
we are confident that the combination of the two strings:
http://purl.oclc.org/dublincore
and:
Title
is globally unique. As Eric wrote, the XML namespaces spec does not
specify how the two strings are to be combined when compared with another
two strings. But that doesn't matter. We're used to situations where our
name and date of birth are used to identify us, without being concatenated
in some fashion.
The "DC" string plays no part in the string comparison; think of this
string as a macro, which is replaced by the URL when it comes to matching.
The following example results in precisely the same data as the one above:
<?XML:namespace href="http://purl.oclc.org/dublincore" as="abcd"?>
<abcd:Title>Hello world</abcd:Title>
Charles mentioned in his mail the possibility of a Distributed Computing
namespace. I might construct a document using both namespaces, eg like so:
<?XML:namespace href="http://purl.oclc.org/dublincore" as="Dublin"?>
<?XML:namespace href="http://www.distcomp.org" as="DC"?>
<Dublin:Title>Hello world</Dublin:Title>
<DC:Mips>12000000000</DC:Mips>
I have avoided a name clash in my file by changing the name of my Dublin
Core namespace from "DC" to "Dublin".
Returning to RDF, the idea is that eventually there may be a machine-
readable schema description at that URL, but it isn't compulsory.
> If we use this same convention for DC:scheme (i.e. state the namespace
> for the scheme by URL) will there be valid URL's for all the schemes DC
> implementors want to use? in other words can all the schemes we might want
> to use be valid namespaces.
You touch on a sore point. I strongly disagree with the DC:Scheme idea,
though I probably used it in examples not so long ago. In my view, there
is no such thing as DC:Scheme. If a DC:Subject uses LCSH, then it is using
a Library of Congress Scheme. If the DC:Subject uses another one of the
twenty or so subject schemes provided in the Scandinavian DC template, then
it is using that scheme and not some DC:Scheme. There is no such thing as
a DC:Scheme. Sorry for the outburst :-).
Returning to your mail, yes there should be a URL for each scheme used.
There are a number of senses in which such a namespace URL could be valid:
1. Will a browser understand it?
2. Does it correspond to a resource owned by the Authority looking
after that scheme?
The former is not essential; the latter is.
---------------------------------------------------------------------------
Misha Wolf Email: [log in to unmask] 85 Fleet Street
Standards Manager Voice: +44 171 542 6722 London EC4P 4AJ
Reuters Limited Fax : +44 171 542 8314 UK
---------------------------------------------------------------------------
12th International Unicode Conference, 8-9 Apr 1998, Tokyo, www.unicode.org
7th World Wide Web Conference, 14-18 Apr 1998, Brisbane, www7.conf.au
------------------------------------------------------------------------
Any views expressed in this message are those of the individual sender,
except where the sender specifically states them to be the views of
Reuters Ltd.
|