[foaf-dev] dc:creator vs foaf:maker
Peter Williams
pwilliams at rapattoni.com
Sun Mar 9 19:41:40 GMT 2008
Axel:
I got frustrated with the topic you introduced, as I went to some trouble t=
o try and understand the issue you raised : starting with FOAF as-is. I nev=
er really got far enough into the topic to really understanding your sugges=
tion, given your hint about updates in Dublin Core. I cannot (practically)=
understand the FOAF document as it is currently stands (which is my lack o=
f basic expertise in core techniques, clearly).
1. I captured Henry Story's foaf file, as rdf/xml. I cleaned it up to remov=
e the non American characters from his friends names. This allowed to me vi=
ew his published resources, with the tool's owl reasoning engine turned on.
2. Reading your email's citations properly took me to the FOAF spec - and t=
o its textual description of the concepts of agents, persons, and makers.
maker - An agent that made this thing. =
Status:stable
Domain:http://www.w3.org/2002/07/owl#Thing
Range:foaf:Agent
The foaf:maker property relates something to a foaf:Agent that foaf:made it=
. As such it is an inverse of the foaf:made property. =
The foaf:name (or other rdfs:label) of the foaf:maker of something can be d=
escribed as the dc:creator of that thing.
For example, if the thing named by the URI http://rdfweb.org/people/danbri/=
has a foaf:maker that is a foaf:Person whose foaf:name is 'Dan Brickley', =
we can conclude that http://rdfweb.org/people/danbri/ has a dc:creator of '=
Dan Brickley'. =
3. Aha! we have a nice example of owl being (simplistically) useful, I said=
to myself, noting the domain of the property is owl:Thing and a defined ag=
ent notion is properly and intuitively responsible for making this particul=
ar type of "document". This will allow me to conclude (from last quoted par=
agraph) dc:creator from foaf:maker is some "creating" individual in practic=
e, distinguished from the declared agent that "makes it". =
4 So, I imported the foaf ontology into the dataset to start reasoning with=
foaf rules. Lots of additional inferences duly occurred. However, inferenc=
es about foaf:maker and dc:creator were not among them, in the individual f=
oaf:PersonalProfileDocument.
5. On the foaf:maker property decryption itself, only rdfs:sameAs <uri> was=
usefully inferred
OK. We are back to 'Peter is just too dumb' for the SemWeb. The fact that c=
onclusion are dc:creator are discussed in the foaf standard , does NOT mean=
that a tool using the rules in the foaf ontology's rdf file will ACTUALLY =
cause a machine to make the discussed inference/conclusion. Sigh. =
Perhaps after all I'm really not that dumb. A more valid conclusion could b=
e : its really hard to tell what is settled in FOAF - vs. what is ongoing b=
asic research.
Then I read http://lists.foaf-project.org/pipermail/foaf-dev/2003-August/00=
5657.html (2003 era)
At that point, I decided I'd really had enough of the topic of maker and cr=
eator. =
Peter.
From: Axel Polleres
Sent: Sat 3/8/2008 5:02 PM
To: Peter Williams
Cc: foaf-dev
Subject: Re: [foaf-dev] dc:creator vs foaf:maker
Peter Williams wrote:
> As a non expert, mere "user", it comes across as a "my name is more =
> viable than your name" playground statement, age 7.
not sure what you are rferring to, I am not referring to identifiers on =
the web, this is a different issue and of course I am aware of that ;-)
I was referring to basically whether dc:creator should be used as an =
objectproperty or as a mere datatypeproperty.
the current foaf spec gives a clear recommendation in that (see [1,2])
whereas the new DMCI spec seems to treat this issue differently in a way =
that does seem to make a separate foaf:maker property no longer necessary.
Axel
> Either the semweb is about everyone being peers in how they name things =
> (with cited equivalencies, possibly ) or its not.
> =
> If the community in the core schema world is undermining the notion that =
> all names are equally valid (and certain naming schemas are not in fact =
> more "valid" than others), we cannot expect folks to let semweb take =
> off. SemWeb has not broken out of the notion that there evn exists a =
> "dominant" naming doctrine - for any layer.
> =
> Speaking as a peon.
> ------------------------------------------------------------------------
> *From:* Axel Polleres
> *Sent:* Sat 3/8/2008 2:04 PM
> *To:* foaf-dev
> *Subject:* [foaf-dev] dc:creator vs foaf:maker
> =
> Hi all,
> =
> What I read about foaf:maker vs. dc:creator at [1,2] seems to be =
> outdated by the latest DCMI recommendations [3].
> =
> Also if you look at http://dblp.l3s.de/d2r/page/authors/Axel_Polleres, =
> it seems that the D2R guys - somewhat ahead of their time ;-) -
> already seem to use dc:creator like that, i.e. referring to URI =
> identified resources directly.
> =
> for me personally it seems that one should deprecate foaf:maker and make =
> it a subproperty of dc:creator.
> =
> Opinions?
> =
> Axel
> =
> 1. http://xmlns.com/foaf/spec/#term_maker
> 2. http://wiki.foaf-project.org/UsingDublinCoreCreator
> 3. http://dublincore.org/documents/dc-rdf/
> =
> -- =
> Dr. Axel Polleres
> email: axel at polleres.net url: http://www.polleres.net/
> =
> rdfs:Resource owl:differentFrom xsd:anyURI .
> _______________________________________________
> foaf-dev mailing list
> foaf-dev at lists.foaf-project.org
> http://lists.foaf-project.org/mailman/listinfo/foaf-dev
-- =
Dr. Axel Polleres
email: axel at polleres.net url: http://www.polleres.net/
rdfs:Resource owl:differentFrom xsd:anyURI .
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.usefulinc.com/pipermail/foaf-dev/attachments/20080309/515=
152f4/attachment.htm
More information about the foaf-dev
mailing list