[rdfweb-dev] Proposal: UsingDublinCoreCreator
Dan Brickley
danbri at w3.org
Thu Jul 24 12:58:46 UTC 2003
* Karl Dubost <karl at la-grange.net> [2003-07-24 08:40-0400]
>
> Le mercredi, 23 juil 2003, à 21:39 America/Montreal, Masahide Kanzaki a
> écrit :
> >>My conclusion is somewhat conservative, somewhat radical: to only use
> >>dc:creator to relate a document to a name of a creator of the
> >>document.
> >>And to use new RDF relations (eg. foaf:maker) when we want to relate a
> >>document to an agent that created it.
> >
> >I read through wiki page and agreed with the discussion, but don't
> >think
> >it's good to 'restrict' dc:creator to name only.
>
> You mean physical person ?
>
> >As Dan noted on the wiki, there are many many implementations which use
> >dc:creator as a container of Agent/People. The change of (or adding new
> >restriction on) dc:creator syntax will make so many people unhappy.
>
> As defined in the spec:
>
> """
> http://dublincore.org/documents/2003/03/04/dcmi-terms/#H2
> Term Name: creator
> URI: http://purl.org/dc/elements/1.1/creator
> Label: Creator
> Definition: An entity primarily responsible for making the content of
> the resource.
----------------^^^^^^
>
> Comment: Examples of a Creator include a person, an organisation, or a
> service. Typically, the name of a Creator should be used to indicate
---------------------------^^^^
(this entity versus it's name thing is the problem)
> the entity.
>
> Type of Term: element
> Status: recommended
> Date Issued: 1999-07-02
> """
>
> A key of the interoperability is not to redefine what others have made.
I don't think anyone was proposing we refine dc:creator. The problem is
that Dublin Core hasn't yet come up with a way of allowing names and
other information about creators to be written down in an unambiguous
way.
I've been involved with Dublin Core since 1997, am a co-editor on two of
the DC-in-RDF specs, was briefly co-chair of the DCMI Architecture WG
and generally have spent a lot of time worrying about doing the 'right
thing' by DC. I'm sad that dc:creator is still confusing after all this
time, and that I've played a part in that mess.
> If you are not satisfied with the original property, create a new one
> which is more complex and suit your needs.
What I've done is the opposite: created a fresh new property that is
_less_ complex than the original. Basically I think the thing to do here
is stand back from the legacy of dc:creator and get some experience with
a simpler property that does nothing at all except relate things to the
things that made them. Most of the problem with dc:creator is mixed
expectations. People want it to do too many things at once...
>
> Is the problem that you can't distinguish between a human and a
> machine? If it's that, you can add a property to qualify the nature of
> the creator, something ala type="person".
No, the problem is that there is no clear answer to questions such as
'what is the dc:creator of this document?', since the allowed answers
include (i) the name of the creator, and (ii) the creator him/herself.
In RDF, this is really confusing, having a single property point to
radically different things. A person is not his/her name...
> So an agent can have created the document, now I think, the debate is
> that you have a human who has created a document with a tool or a
> series of tools and in this sense the property made is interesting or
> madeby.
No, that's a side-discussion (on whether non-Person Agents can create
things). The issue / problem here is purely to do with the vague
definition of dc:creator, which allows both people and their names to
'count' as being the creator of a thing.
>
> I have one question though. What about if your tool is not ONE tool but
> a series of tools. You have used this and that to create, and you
> finally finished it by hands. Would you like to keep the history of the
> tools used to edit it?
Keeping track of workflow in that style is definitely useful, I think
modelling events in the workflow is the way to go.
http://www.ilrt.bris.ac.uk/discovery/harmony/docs/abc/abc_draft.html for
some old ideas on that...
cheers,
Dan
More information about the foaf-dev
mailing list