[rdfweb-dev] FOAFnet and the future of FOAF in

[Julian Bond]

Jeremy Gray <jeremy at jeremygray.ca> wrote:
>On Behalf Of Joel De Gan

>> Anyway, the module approach (ala:
>> http://www.perceive.net/schemas/20021119/relationship/ ) is an
>> excellent way to add in things to foaf without forcing anyone to
>> implement your particular flavor of foaf. But, obviously if everyone
>> does that, it is going to be a nightmare.
>
>True. One of RDF's great freedoms is that we _can_ all go off, invent our
>own little modules, and then infer between similar redundant modules

If we're talking about adding in data from another namespace (like
vcard) I'm in complete agreement. But if we're talking about replacing
core FOAF elements with subclassed versions like the relationship
schema, I'm not sure I am. Maybe I'm not using a powerful enough parsing
tool, but ISTM that a FOAF reader has to have hard coded knowledge of
this namespace unless both the FOAF superclass and namespace subclass
are included in the FOAF file. eg

Person
 yyy
 knows
  xxx
 acquaintanceOf
  xxx
Is fine because if I don't understand acqauintanceOf I can just ignore
it.

If its written like this

Person
 yyy
 acquaintanceOf
  xxx

then you're saying that yyy has no "knows" until I process the
information that acquaintanceOf is a sub class of knows. For this one
namespace it's ok because it's becoming a standard and reasonably well
known. Until the tools automatically discover this information, it's
hiding information that could have been used.

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