RE: [xtm-wg] A challenge on "the graph"

[Sam Hunting <sam_hunting@yahoo.com>]

[daniel]
> Perhaps it is also a further requirement statement for the processing
> model - a more formal definition of the interoperability requirement.
> I like the feel of this.

Agreed... Notations are hard to do, though.

[nikita]
> > May be we should introduce a notion of "Universal ASR" ?

[daniel]
> I think this is what Groves aim to provide. It would be of real value
> to see a Grove representation of a topic map, and a formal 
> description of how to map from XTM syntax to such a representation.

That would certainly be one solution that a lot of people could live
with. It would seem to answer the question of how to represent a graph
as a set of nodes with properties (which seems to be the request of one
set of PM advocates).

[daniel]
> Looking at Lars's approach based on XML Infoset, I am extremely
> concerned that it is far too close to the XTM syntax. I don't want
> information objects corresponding to the elements in the XTM 
> document. I want the information objects to reflect the meaning of 
> those elements, which are much closer to the conceptual model.

Agreed -- astonishing, eh? -- but, Daniel, what requirement do you see
this want that you have fulfilling? Some may find this desire
unpersuasive in the absence of requirements.

[daniel] 
> Lars finds the conceptual model "too conceptual". I find his PM
> approach "too syntactic".

Which is why we need a middle layer, eh? :-)

[daniel] 
> I did submit work in progress towards a UML model of the mapping from
> the Conceptual Model to the syntax. No-one has responded to this, as
> far as I am aware. 

Yes, we do have some "talking past each other" in this area. That's why
I ewncourage people to raise objections so they can be answered...

However, I did answer and raise the Bic Picture question of why one
would wish to map directly from CM to syntax without going through the
middle layer of the PM. Above, you seem to be arguing for exactly this
approach, unless I have misrepresented you. What am I not
understanding?

[daniel]
> In doing this mapping work (which is far from complete - but I
> wanted feedback on the approach before continuint) I found it 
> necessary to create some "abstract classes". These are classes that
> are never instantiated in the syntax. Could these be the middle layer
> objects that Lars really wants?

Hmm... Can you help me out with some words? 

My thought is that this is similar to the idea that in "the graph",
when the angle brackets are stripped away, there is a an a-node that
instantiates the topic-basename template, with one arc labelled "topic"
connected to the t-node demanded by the topic elements, and a second
arc labelled "basename" that connects to the t-node demanded by the
basename string element. (This is nice, because now we have a simpler
though therefore more verbose mechanism of associations. We probably
also have a model of "meaning" -- when roles are taken into account as
subjects, "the graph" starts to look a lot like a conceptual graph, a
la Sowa.) 

Now, the template does not appear in the angle brackets, nor do the
arcs -- though there are rules that show how the one demands the other.


Leaving out the details -- and, as TS Eliot says, "I gotta use words
when I talk to ya", so we are ALWAYS using syntax ;-) -- is this is
sort of thing you had in mind, Daniel?

S.


=====
<!-- "To imagine a language is to imagine a form of life."
     - Ludwig Wittgenstein, Philosophical Investigations -->

__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/

------------------------ Yahoo! Groups Sponsor ---------------------~-~>
Do you have 128-bit SSL encryption server security?
Get VeriSign's FREE Guide, "Securing Your
Web Site for Business." Get it now!
http://us.click.yahoo.com/EVNB7A/c.WCAA/bT0EAA/2n6YlB/TM
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/