Re: aboutEachPrefix: possible alternative

On Fri, 29 Sep 2000, Mark Nottingham wrote:

> 
> Hello,
> 
> I'm a member of the P3P working group. As you may know, we have a
> requirement to apply metadata to a broad selection of resources, often from
> a central location (The "well-known location").
> 
> This isn't an uncommon requirement; it's evidenced in several
> currently-deployed systems, and is of particular interest in web caching
> and content delivery networks (of which Akamai is an example).
> 
> aboutEachPrefix, as discussed here previously, isn't a good solution for
> these tasks. However, it _seems_ that this is still a good space for RDF to
> be used in.
> 
> At Akamai, I've spent a fair amount of time designing formats which allow
> our customers to apply metadata based on the URI namespace. We've gone
> through several iterations, and it has become apparent that this problem is
> not unique to us, and the various applications that face this problem could
> benefit from a common solution.
> 
> To this end, I've prepared a document that describes the problem and offers
> a straw-man solution. I'd very much like feedback as to whether this would
> be a useful problem to solve in such a general way, and comments on this
> particular proposal.
> 
> I'm especially interested to see what RDF people think about this; I'm
> curious as to how well it fits with the philosophy of RDF as a whole.
> 
> Regards,
> 
> 

Thanks Mark, interesting proposal! I think there's consensus about the
need to move beyond 'aboutEachPrefix' but a lack of clear proposals as to
how to do this. I hope your strawman stirs up some discussion.

Briefest of comments for now. BTW your attachment was unpacked by the list
archiver at
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Sep/0264.html
http://lists.w3.org/Archives/Public/www-rdf-interest/2000Sep/att-0264/01-urispace.html
although it doesn't seem to be being served as text/html


- I agree that often we'll want to pick out classes of Web resource using 
URI matching.

- often we'll want to use other, more abstract, mechanisms for picking out
classes

- I like the general drift of your paper in terms of providing some
mechanisms to support the former. I'm not sure at this stage about the
actual mechanisms proposed. Eg. one might imagine using theregex facility
from the XML Schema datatypes in a similar fashion, 
 

I need to re-read your proposal more carefully; my initial reaction
is: useful goal, not sure about the machinery for getting there. 

To flesh this out a little:

In the 'aboutEachPrefix' / PICS tradition, I might want to say that
every resource whose URI matches http://playboy.com/.*jpg is of rdf:type
foo:PotentiallyRudePicuture.

In other contexts, I might want to create an RDF class
bar:FriendRelativeOrCoWorker and populate it using an RDF query, so that
various people (or their mailbox URIs or whatever) are represented as
being of that rdf:type.

In both contexts, I might want to make generic descriptive claims about
the members of these classes. So mechanisms such as the one you sketch
might fit into a bigger picture involving other (less URI/regex
centric) ways of picking out collections of Web resources.

Thanks for opening the debate with a concrete proposal,

cheers,

Dan

Received on Saturday, 30 September 2000 10:11:14 UTC