W3C home > Mailing lists > Public > www-rdf-interest@w3.org > June 2003

Re: Proposal: "Canonical" RDF/XML

From: Charles McCathieNevile <charles@w3.org>
Date: Sun, 29 Jun 2003 12:07:15 -0400 (EDT)
To: Benja Fallenstein <b.fallenstein@gmx.de>
cc: rdf-i <www-rdf-interest@w3.org>
Message-ID: <Pine.LNX.4.30.0306291151370.23294-100000@tux.w3.org>

Ergh. I get the feeling I am not explaining this very clearly, or I am
missing something fundamental.

On Sun, 29 Jun 2003, Benja Fallenstein wrote:

>Hi Chaals,
>Charles McCathieNevile wrote:
>> On Sun, 29 Jun 2003, Benja Fallenstein wrote:
>>>The first application for this is that my working group wants to
>>>maintain a TODO list/calendar/schedule/note board/etc. as an RDF graph
>>>in our CVS repository. We want to be able to mix many different
>>>vocabularies there.
>>>I think that your approach doesn't apply there... :)
>> I think it does. You keep RDF data, with whatever mixed vocabs and so on you
>> have, in a document. But when you want to work with it you have an interface
>> that gives a canonical form - an for different tasks you maight have
>> different forms.
>Hm, I don't think that I want the file we keep in CVS to be a
>task-specific XML form of the RDF data. Again, I see how task-specific
>XML can be useful when "working with it," but I think that my use case
>is more about "keeping it." I would definitely not want to change my
>canonicalization tool and adapt it to a new XML schema each time I want
>to use a new vocabulary in my RDF.

No, I don't want to keep the XML - the tools write it out and 'turn the XML
into RDF' (most trivially by sticking it into rdf:RDF tags...)

The idea I have is that what you provide the tool is a dataset to query, and
the XML template you want it to match (this time). Typically that template
will fall rapidly out of the application you have at the time, and you may
not even bother to keep it.

The kind of query interface that RDF Author suggested, where you have some of
a graph and you want it matched according to some conditions is in my mind.
Maybe I should find some time to write it up in detail.

>(BTW, I also have a second, related, and to me actually more important
>use case: Identifying versions of RDF graphs by cryptographic hashes, so
>that I can reconstruct a version using diffs and then check it against
>the hash. -- Hm, I think I mentioned this once before on this list,
>months ago *scratches head* (anyway) :-) )

for that you just need a serialisation that you can check, and your proposal
makes sense. (I presume you specifically want to use diff rather than a
triple-wise diff that understands blank node IDs...)

>> My use case is as follows:
>> I collect data mixing foaf information about who knows who, what they are
>> interested in and working on with information about what languages they
>> speak, stuff about where they are when, (both in generic location and in
>> terms of attending events), what they look like, and perhaps a few other
>> things.
>Sounds interesting -- out of curiosity: is there a practical application
>you have in mind for this or is it mostly just for fun? :-)

I started doing it for fun (because my friends were). Then I discovered I
wanted to find out what a person looked like (because I was trying to find
them at a conference), and that I was wondering if there is a danish-speaking
person interested in the semantic web, around copenhagen. (not surprisingly,
there is. Even better, the semantic web told me so...)

I did want to find out if there were hungarian speakers working on the
semantic web, but it turns out they don't have a lot of this data available
yet so I googled - which is a terribly inefficient way of finding out stuff
like this.

I hope to make available a bunch of information that says there are people
speaking spanish and interseted in a few fields. I think this kind of data
has very wide application in the non-english speaking majority of the world.
And I find it helpful stuff, plus a good space to work with demos.

the xform stuff is at
http://www.w3.org/2001/sw/Europe/200305/foaflang/xforms/ if you're


Received on Sunday, 29 June 2003 12:07:21 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:43 UTC