W3C home > Mailing lists > Public > semantic-web@w3.org > December 2005

RE: How will the semantic web emerge

From: Hans Teijgeler <hans.teijgeler@quicknet.nl>
Date: Sat, 17 Dec 2005 18:33:41 +0100
To: <semantic-web@w3.org>
Cc: <collin@w3china.org>
Message-ID: <001201c60330$0c63abd0$6c7ba8c0@2e1a3193d4c544a>
Hi,
 
This thread was interesting, but I missed the following:
 
Some problems-to-be-solved we have in the process industries (oil & gas,
chemicals, pharma, etc) are:
*         how do we integrate data about one plant that originate from a
plethora of systems used in a large number of cooperating companies?
*         how do we do that over the lifetime of that plant? (typically
10-40 years)
*         how do we do that where we don’t know upfront who will be our
partners on a project?
*         how do we selectively hand over data to our client (the plant
owner), whilst maintaining referential integrity?
 
The semantical soundness of the data must be 100%, because we are
dealing with utterly expensive facilities that may be utterly dangerous
(remember the recent fires in Hemel Hempstead!). Misunderstandings can
be fatal, or at least very expensive.
 
Interfacing on a system-to-system basis leads to an exponentially
growing number of interfaces.
 
Our choice has been to map the data of those systems that need to be
shared and integrated to one standard common format (now defined in OWL,
perhaps later in a successor of OWL), and integrate them in a number of
standard triple stores. Each system has its own standard triple store as
a kind of Extranet, and requires only one mapping to that store.
 
Using SPARQL anyone involved (and entitled) can obtain any information
about the plant (or part thereof), during any point in time in the
lifetime of that plant, from what we call a predefined CPF
(Confederation of Participating Fašades) (Fašade = triple store).
 
The standard common format is defined in templates as per ISO 15926-7,
that are based on the data model of ISO 15926-2, in conjunction with
standard classes and properties defined in ISO 15926-4. This appeared to
fit like hand-in-glove with OWL (Full).
 
We are setting up a system that allows ISO-accredited domain experts all
over the world to propose new classes and object information models,
that ultimately will acquire ISO IS (=International Standard) status.
 
For an overview please see [1].
 
Regards,
Hans
 
[1] http://www.infowebml.ws/introduction/introduction.htm 
Received on Saturday, 17 December 2005 17:34:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 February 2013 14:24:52 GMT