Peter Ansell wrote:
2009/3/24 Kingsley Idehen <kide...@openlinksw.com>:
If you are going to honor the Identity principle on the Web, in an
unobtrusive manner (i.e., leverage ubiquity of HTTP) there is no way around
the above.
I never thought of it as honour. The large number of ideals
surrounding the redirect law still don't impress me.
In the context of my comments about "honor" means: "adhere to" :-)
The ideal of separating the Identity of a datum from its representation
isn't a feature unique to the Linked Data realm.
I assume you are familiar with the concept of "Object Identity" [1] ?
The linked rdf
web is coming along despite these ideals increasing the general apathy
to the whole concept so far IMO.
The mechanics are surfacing in the wrong spaces, and in any such
scenario apathy is a risk. Stuggles with 'C' pointers comprehension
produced similar
reactions back in the day. In short, many object oriented languages
tried to take pointers out of scope in order to address similar issues.
This matters are quite old, but very
important all the same.
The fact that this debate is still
going is worrying, but hopefully it won't stop people discovering the
usefulness of shared resolvable identifiers which form the real basis
to the usefulness of linked rdf.
Do you want to reference and de-reference at the datum level or not? Do
you see any utility in this at all?
Just curious though, how is it *actually* obtrusive not to implement a
redirect if ones ideals tell one that there is no functional
difference in the nature of the resource whether its representation
was derived from the result of a redirect or not?
No, redirection is not the point of "obtrusion". Using HTTP based URIs
is how you unobtrusively apply "Identity" to the current ubiquitous Web
of HTTP based
user agents. Otherwise, we would be using URNs and the Handle system.
If people want to
track provenance as Michel says, they can either recognise that some
triples have predicates which relate to provenance and others are
independent of the provenance, or they can store the provenance
information in a different graph and query across the two if needed.
That part is up to the interested end-user and really shouldn't raise
the entry bar to new producers.
We are not interested in the documents (units of persisted
contextualization of data), we are interested in the "raw data" in the
documents. We want to reference them and de-reference negotiated
representations of their descriptions.
The real issue here is not the "Identity" principle but how poorly this
has been explained e.g. by sometimes describing these concepts as being
unique to the Web.
We are applying an age-old pattern: Data Access by Reference, within
the context of the Web. This pattern is as old as computer science itself.
Kingsley
Cheers,
Peter Ansell
Links:
1.
http://www.cs.cmu.edu/afs/cs.cmu.edu/user/clamen/OODBMS/Manifesto/htManifesto/node4.html
-- Object Identity
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com
