[rdfweb-dev] Re: PHP RDF Parser

[Julian Bond]

Libby Miller <libby.miller at b...> wrote:
>On Fri, 10 Jan 2003, Julian Bond wrote:
>> 1) I got an app working about 30 minutes ago and have been browsing
>> round FOAF files on the web. I'm seeing a number that look like this.
>> <rdf:Description>
>> <foaf:name>XX YYYYYY</foaf:name>
>>
>> with no <foaf:Person>
>
>there isn't really the concept of valid foaf.

Ooh err!

>Event if domain of name
>was Person a resource could in fact be a Person even if it doesn't say
>so rihght here - you just don't know - there's an open world assumption
>in RDF (you could always find some more information). In fact also the
>namepsace document http://xmlns.com/foaf/0.1/ doesn't specify a domain
>for name.

OK. Bad example. So how about

<rdf:Description>
...
<foaf:workplaceHomepage rdf:resource="xxx.yyy" />
...

http://xmlns.com/foaf/0.1/ appears to say that workplaceHomepage should
be of domain Person.

Give the source, the file this came from is clearly meant to be FOAF. So
the author appears to be saying that the object might be a Person or
might not, they just don't know. So they've used <rdf:Description>
instead of <foaf:Person> as the top level domain. Curious, when it's
themselves they're describing. ;-)

Later in the same file, they've got <foaf:knows> entries within the same
<rdf:Description> that contain more <rdf:Description> instead of
<foaf:Person>. (note that the foaf spec also says that foaf:knows should
have both a domain and range of foaf:Person)

It's as though the author has deliberately not used <foaf:Person> which
strikes me as obtuse. Perhaps there's a reason I don't see yet, perhaps
it's a mistake.

-- 
Julian Bond Email&MSM: julian.bond at v...
Webmaster: http://www.ecademy.com/
Personal WebLog: http://www.voidstar.com/
CV/Resume: http://www.voidstar.com/cv/
M: +44 (0)77 5907 2173 T: +44 (0)192 0412 433