[rdfweb-dev] Which Person wrote this FOAF?
wkearney99 at hotmail.com
Tue Jul 29 14:57:49 UTC 2003
Look at my foaf for one example.
It has a description element at the end. Here it's stating the URL "" (as in,
"this" document) is described as being created by me, on a given date, modified
on another and assurable using the ascii armored signature at a particular URL.
Now, the leap here is the use of dc:creator in this manner. There's some
ambiguity as to whether this is or isn't correct.
But yes, when triples are 'dumped together' into a larger store there are
definitely going to be some issues with proving who's authoritative over what.
I've long requested a 'best practice' on making authoritative statements about
one's self and to consistently validate that programmatically. It seems tools
are only now emerging that make such things possible. The need has, I'd argue,
always been there. It's almost like it's been left to the "we can't do that yet
so we don't talk of it" state of limbo...
----- Original Message -----
From: "Julian Bond" <julian_bond at voidstar.com>
To: <rdfweb-dev at vapours.rdfweb.org>
Sent: Tuesday, July 29, 2003 4:24 AM
Subject: [rdfweb-dev] Which Person wrote this FOAF?
> I was going round and round last night trying to understand this.
> Most FOAF files in the wild follow a pattern. They have one foaf:Person
> with lots of information about the person that wrote the file and then a
> long list of foaf:Persons with minimal info linked via foaf:knows.
> There's an implied suggestion that the Primary foaf:Person information
> is authoritative because it's about the person that made the file.
> Several of the Foaf Explorers (even those written by RDF-heads) display
> this implied information because the main person is displayed first at
> the top of the html page.
> When you look at the file as triples, this implied authority is lost.
> It's just a list of foaf:Persons one or more of which have a list of
> foaf:knows. There's no explicit triple to say that this foaf:Person
> wrote the file authoritatively and it's mostly about them.
> As well, in triples, there's no reason why a file shouldn't have more
> than one "primary" person.
> Now I'm not trying to go down the route of pgp signing and proof of
> ownership. I'm just trying to recover the information that this file is
> mostly about this person. In the XML view it's mostly obvious because
> the file only has one top level node and that's a Person. But how am I
> supposed to extract this information from the triples?
> We could use foaf:maker and foaf:made to store the data but these are
> new and not widely used yet.
> For a spider, it probably doesn't matter because this file just adds to
> the data store of triples and they're all related via an mbox_sha1sum
> key. But for a FOAF explorer that html-izes a single file, I think it
> Is this just an artifact of RDF-API for PHP (RAP) and other RDF parsers
> have a getTopNodes() function or something to extract the lower level
> XML structure info if required?
More information about the foaf-dev