ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Architectural considerations in Ontology Development

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Barkmeyer, Edward J" <edward.barkmeyer@xxxxxxxx>
Date: Fri, 22 Feb 2013 12:10:56 -0500
Message-id: <63955B982BF1854C96302E6A5908234417D4F59FE3@xxxxxxxxxxxxxxxxxxxxxxxxxx>

+1

 

“Reverse engineering”, i.e., engineering knowledge from one or more formal (and informal) sources into “formal ontologies”, is exactly what three active NIST projects, in cooperation with the related industries, are doing.  The process industries for example, have a plethora of documents in formal forms.  The problem is to get a common model of the intent of the 6 terms on the 6 different “product data sheets” and of the column titles on the two alternative spreadsheet forms, each of which corresponds to a specification with a natural language definition somewhere. 

 

Similarly, the idea of “lifting” an XML schema or an SQL schema or spreadsheet to a fake “ontology” expressed in OWL or RDF is a cottage industry in academia.  The fake ontology is just using OWL (or RDF) as a common language for expressing the schema itself (table to class, column to property), but then one can talk about “ontology integration” or “ontology mapping” as what one is doing in engineering that thing into a real knowledge model.  It is all reverse engineering.  This approach just labels the result of the first step with a more grandiose title, so that the author can call his knowledge extraction techniques by one of the vogue buzzwords.  Those buzzwords were originally intended to refer to efforts to combine independently developed ontologies, which is “quite another thing entirely”.  So don’t be misled by the terms. 

 

-Ed

 

--

Edward J. Barkmeyer                     Email: edbark@xxxxxxxx

National Institute of Standards & Technology

Systems Integration Division

100 Bureau Drive, Stop 8263             Work:   +1 301-975-3528

Gaithersburg, MD 20899-8263             Mobile: +1 240-672-5800

 

"The opinions expressed above do not reflect consensus of NIST,

 and have not been reviewed by any Government authority."

 

 

 

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Mike Bennett
Sent: Friday, February 22, 2013 11:43 AM
To: [ontolog-forum]
Subject: Re: [ontolog-forum] Architectural considerations in Ontology Development

 

One comment:

 

But I'm pretty sure such practical reverse engineering is of minimal interest to the Ontology world.

 

MW: If that were true, why would we have run a session on Methodologies for developing integrating ontologies? In fact it is easily the largest practical use of ontologies in industry.

 


This certainly has been the thrust of our work. The financial industry has lots of standards for messaging, and some at the level of common logical data models, but the requirements for integration within financial firms has led to an increasing realisation that what's really needed is common semantics: individual terms, with agreed definitions and formally framed meanings, independent of all these physical and logical data models. Those technical models have been very good for their purposes, but each time a firm buys in a new system, adds a new data feed, connects to a new broker or other market participant via another feed, or merges with or acquires another financial institution with a whole lot of similar applications (most of them having different data models), then there is a case to be made for common semantics.

Now of course one could address some of the need for common semantics with a good old low tech vocabulary, but it makes a lot more sense to use the available expressive power of formal ontology languages (or least a sub-set of those formalisms that can be shown back to the business for validation). So that is what we've done with the Financial Industry Business Ontology (though, for most business participants, we still end up describing this in terms of a common business vocabulary - at least until business folks start to understand what formal ontologies can do).

So the industry needed a formal, standardized ontology for financial industry concepts, and this has been and is being built up along Semantic Web principles. Would we have started with a blank sheet of paper to do this? No, there are all those logical and physical data standards which represent considerable work and considerable business knowledge (and in the more widely used ones, considerable feedback from the real world so we know they actually worked).

The message standards and logical data model standards are therefore a good source of knowledge, the only problem with them being that, with a few honourable exceptions, the business knowledge that business people brought to the table was immediately hard coded into XML schemas or UML class models without first being documented in a technology-neutral and design-independent way. That is, while business knowledge was brought to the table there was a double obscuration of this knowledge from the business itself: through the chosen design of the designed artefact (XML schema, RDB or UML Class model etc.), and through the technical language in which those designs are typically expressed (XMLSpy, ERM, UML etc.).

This is symptomatic of a deeper problem in many message or model standards: someone has designed a design without ever documenting the business requirements against which that design was designed, and in defiance of most accepted engineering practice (other than that which is applicable to games or to one-off prototypes). That is, we have message models without some corresponding conceptual models.

This leads to not one but two motivations for reverse engineering from designs to ontologies:

1. It provides the "missing" conceptual model which sets out the business concepts against which those technical standards models were apparently designed; and

2. It provides the industry with a conceptual model for integration, common message hub creation etc.

This is why I tend to talk in terms of conceptual models or conceptual ontologies rather than "Integrating" ontologies per se: integration is one of the use cases for a conceptual model (including a conceptual model for data, i.e. an ontology). Other use cases are in the ongoing maintenance and business validation of just about anything for which there is intended to be some future versioning and maintenance. Like standards. Or anything which can lose you money if it messes up.

So yes, reverse engineering into ontologies is very much of interest to us business problem solvers. Whether it's of interest to the "ontology world" or not is not something I'm in a position to comment on, but it's very much of interest to the business world :)

Mike



 

Regards

 

Matthew West                           

Information  Junction

Tel: +44 1489 880185

Mobile: +44 750 3385279

Skype: dr.matthew.west

matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx

http://www.informationjunction.co.uk/

http://www.matthew-west.org.uk/

 

This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.

Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.

 

This week I've seen what appeared to be a promising corporate glossary.  On first use with 5 terms which were not found... oops!  That's one prospect who's never going to return.  And this was just a simple glossary.

 

- David

 




 
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
 

 


_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)

<Prev in Thread] Current Thread [Next in Thread>