Jim, I think there *is* interest in “determining an authoritative
data model source”, and in fact in the whole topic. I for one am vitally
interested, and would like to see constant improvement in the documentation of
the Higgins data model, since it is essential to the success of Higgins as
whole. I know Paul and Markus and I have a deep interest in the topic as I’ve
been working with the two of them and others on the XDI TC to map Higgins data
model to the XDI RDF data model (initially documented at http://www.oasis-open.org/committees/download.php/25531/xdi-rdf-model-v7.pdf
-- better documentation coming soon). Doing a completely clean mapping is
the easiest way to enable Higgins to speak XDI and vice versa. We plan to give
a presentation at the F2F on this.
I talked to Paul earlier today and he
apologized for being snowed under and said he would be getting to this thread
as soon as he can.
In my own review of the http://wiki.eclipse.org/Higgins_Data_Model
section of the Higgins wiki, I’ve found the typical situation with a
fast-moving technical wiki: some pages are great, most are average, and some
are just confusing deadwood. It’s often easier for a newcomer like me to
see that – once you get “into the flow” of the wiki you tend
to concentrate on the new stuff and ignore the older cruft.
So leading up to (and after) the F2F, I’d
be happy to volunteer to help garden the http://wiki.eclipse.org/Higgins_Data_Model
section of the wiki to see if we can’t turn it into accurate, up-to-date documentation
of the current state of the data model and clear assessment of all open issues.
higgins-dev-bounces@xxxxxxxxxxx [mailto:higgins-dev-bounces@xxxxxxxxxxx] On Behalf Of Jim Sermersheim
Sent: Monday, January 14, 2008
Subject: Re: [higgins-dev] What
defines the Higgins data model?
well there doesn't seem to be a swelling interest in determining an
authoritative data model source. I guess these emails will be come that
spawn new threads for the different topics we're discussing in hopes that we
>>> "Tom Doman" <tdoman@xxxxxxxxxx> 01/14/08 10:29 AM
It's possible they'll each be constrained or shaped by the other. If
for no other reason, necessity.
At any rate, yes, let's get some consensus on this and make some progress
refining things! Let's at least make a list of the contentious
>>> "Jim Sermersheim" <jimse@xxxxxxxxxx> 01/11/08
11:34 AM >>>
Before addressing all three model threads recently spawned, I'm interested in
understanding what defines the Higgins Data Model we use in IdAS and HOWL.
Do we define HOWL, and expect IdAS to follow? I know we don't intend
to start with IdAS and make HOWL conform (though it has
happened). Or do we have a conceptual data model which is not
formalized in any modeling language from which we try to make both HOWL and
So far, I've believed the latter. And that where we do this is
starting with the Higgins Data Model ( http://wiki.eclipse.org/Higgins_Data_Model )
page. From there, we define what is a Context ( http://wiki.eclipse.org/Context ),
Digital Subject ( http://wiki.eclipse.org/Digital_Subject ),
and Attribute ( http://wiki.eclipse.org/Identity_Attribute ). On
the Attribute wiki, we talk about the fact that:
- an Attribute has 1..N values.
- value types are XML- Schema literals or complex (I'm unclear on
whether we consider xsd:anyType as a literal)
- values are unique. (Note that determining uniqueness
across disparate types would add complexity)
Anyway, before we dive deeper, can anyone (I'm thinking Paul) provide clarity
on what's authoritative?
higgins-dev mailing list