W3C home > Mailing lists > Public > semantic-web@w3.org > March 2005

Re: tagging, scope, sweeping statements

From: Richard Newman <r.newman@reading.ac.uk>
Date: Sun, 27 Mar 2005 21:41:09 +0100
Message-Id: <08a3c38ea9c8adc3394fef933c064fe8@reading.ac.uk>
Cc: semantic-web@w3.org
To: Stan James <sjames@uni-osnabrueck.de>

Stan,

> Yes, I was worried about that too. My best rationalization is that my 
> tags are on par with a librarian filing a Dostoevsky novel under 
> "Fiction", even though it has Forward and Author Biography sections 
> that are very much fact.  Since the scope of the tag is "book", you 
> don't mind if sub-bits don't fit the tag.

That's a very good point! I'd be inclined to agree with you, at least 
on the Web; I don't know what tagging a URI  which is used for a SemWeb 
resource would mean, or a mailto: URI. I don't really want to go there 
-- too hard! ;).

> Your comment about not looking in URI's was a good point. I'm looking 
> into Phil Archer's "RDF Content Labels" as suggested by Benjamin.

I did mumble "d'oh" when he suggested that, given that I'd read that 
email :D
2 + 2 didn't get put together.

> ...and if I can ask another newbie question (of the whole group)...
>
> My del.icio.us-ish RDF sometimes includes bookmarks of other RSS 
> files.  (Recursion!)  But how should I indicate that the target file 
> is [a similarly formatted] RDF file and not plain URL?  My quick 'n 
> dirty approach has been to use the dublin core "Type" element, eg: 
> <dc:type>myformat</dc:type>.  Is this appropriate?

I'm not sure I know precisely what you mean, but I'll weigh in anyway.

1. why would you want to indicate what the object of the bookmark was? 
Are you indicating that some are JPEGs, some are RDF, some are PDFs? 
This may inform your decision -- e.g. to use rdf:seeAlso in the case of 
RDF.
2. you could make a subproperty of the bookmark property that you are 
using -- such as "pdfBookmark", "rssBookmark".
3. you could tag on some type property to the bookmarking, which is the 
approach you seem to be taking.

I seem to recall solving a problem like this when drafting a 
bibliographic ontology, but grepping hasn't found my solution. I expect 
it lived for a while on my office whiteboard :D

The rest of the list will have much more to contribute.

HTH,
-R
Received on Sunday, 27 March 2005 20:41:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:41:45 UTC