Drummond,
> Im
interested in the direction you are going because XDI also has the requirement
to be self-describing. I like what youve done so far. Keep
going.
We already have a set of well-defined IModel interfaces which
describe the context perfectly. Is there any reason/benefits to
replace the existent model with a new proposed?
Jim,
> It may be better to not define special interfaces for this purpose,
but instead simply re-use the existing
> interfaces that are used to access normal information within a Context
(nodes and their attributes).
Why it is better to use INode interface to describe a
model of IContext, IAttribute, IAttributeValue and others? Really, we
see only disadvantages in describing a context model in such a
way. Using of new model does not look convenient for a
user. For an example, he will need to investigate a
complicated sructure of INode instances to know the list of possible values
of simple attribute (defined by owl:oneOf data
range). Moreover, in case of such a model we
need to create a new special schema to describe HOWL
schema. So, we think that re-using of INode interface is not a
reason to reject existent IModel interfaces. I only agree with "top node". It
will be helpful to add getTopNodeModel() method to IContextModel interface,
and getSuperNodeModel() and getSubNodeModels() methods to INodeModel
interface.
Thanks, Sergey Lyakhov
----- Original Message -----
Sent: Friday, February 29, 2008 6:45
AM
Subject: RE: [higgins-dev] preliminary
work on "IdAS data model as nodes"
Jim,
Im interested in the
direction you are going because XDI also has the requirement to be
self-describing. I like what youve done so far. Keep
going.
=Drummond
On a number of occasions, I've
mentioned my desire to explore the notion of using existing IdAS APIs to
access a context's data model (schema).
This page (http://wiki.eclipse.org/Higgins/ModelAPIs) is
a start, but it needs polish. It has a number of known (to mee) problems
that need to be dealt with. I just thought I'd mention it in case anyone
is interested in at least seeing the direction I'm going.
Jim
_______________________________________________ higgins-dev mailing
list higgins-dev@xxxxxxxxxxx https://dev.eclipse.org/mailman/listinfo/higgins-dev
|