ontology-summit
[Top] [All Lists]

Re: [ontology-summit] Reusability and Interoperability

To: Ontology Summit 2014 discussion <ontology-summit@xxxxxxxxxxxxxxxx>
From: "Barkmeyer, Edward J" <edward.barkmeyer@xxxxxxxx>
Date: Thu, 3 Apr 2014 15:13:15 +0000
Message-id: <bf91568d3a3e4a32b050b57f067a00da@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>

Matthew,

 

I agree that reuse allows one to identify and correct shortcomings in the reused ontology.  But I also agree with Jack and John that any change to an ontology that is reused *by reference* requires a regression test on every existing use. 

 

In most engineering disciplines, a new version of a reused component has a new identifier, and the decision to move the reference from the older version to the newer one is made by the re-user.  Ontology should not be an exception.  In a disciplined organization, creation of the new version causes the transmission of an Engineering Change Notification to all parties known to have an interest.

 

(In another forum, which I believe Matthew shares, we have just seen two emails on exactly this topic for the management of a large industry ontology:  ISO 15926-4.)

 

-Ed

 

 

From: ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:ontology-summit-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Jack Ring
Sent: Thursday, April 03, 2014 3:06 AM
To: Ontology Summit 2011 discussion
Subject: Re: [ontology-summit] Reusability and Interoperability

 

Matthew,

The other side of this coin is that the more an ontology is reused the more it is ‘adapted’ (usually complexified) then earlier users encounter bugs they never had before. 

 

As ontologies become laced with predicates the need to verify efficacy of process and invariants becomes very important. 

 

Andrea,

I will encourage Antonio Pizzarello to explain the progress and safety principles better than I. 

Meanwhile, it may be useful to look at the ontology ‘quality’ situation from another perspective. Consider a computer program to be an ontology decorated with predicates — a specially ordered set of operands and operators (wherein the operators are also ontic claims). Then how do we discover let alone guarantee the envelope throughout which that the program will produce correct results?

Jack Ring 

 

On Apr 2, 2014, at 11:19 PM, Matthew West <dr.matthew.west@xxxxxxxxx> wrote:



Dear Andrea,

Just to add to what you have below. Reuse also improves quality. The more an ontology is used, the more bugs are identified and eliminated, and  this is even more the case when the different uses are diverse. This in turn makes the ontology more reusable, so you can find yourself in a virtuous circle. This assumes that you have an effective method for fixing bugs, and that is not always the case.

 

Regards

 

Matthew West                           

Information  Junction

Mobile: +44 750 3385279

Skype: dr.matthew.west

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

Registered office: 8 Ennismore Close, Letchworth Garden City, Hertfordshire, SG6 2SU.

 

 

 

From: ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:ontology-summit-bounces@xxxxxxxxxxxxxxxx] On Behalf OfAndrea Westerinen
Sent: 03 April 2014 04:19
To: John Yanosy Jr.; Ontology Summit 2014 discussion
Subject: Re: [ontology-summit] Reusability and Interoperability

 

JohnY, JohnS and Jack,

Since the Track A co-champions are trying to write a synthesis, I want to come back to this email thread (which I let languish in my in-box waiting for time to write a reply) ...

I very much like the definitions from Adams that Jack noted.  So, I went and downloaded the paper for the 1993 panel where he participated.  Here is what he and others said that could be reused for ontologies:

* Why reuse?  To reduce the cost of development by developing less; To obtain a higher return on investment on what is developed

* (Brad Balfour) The implementation language will impact almost all the technical aspects of reuse.  The capabilities of the language impact what can be expressed and how you express it.  If languages are too diverse in their capabilities, then you can't switch among them.

* (Sam Adams) "Reuseful" means that the component provides something ("behavior", in the case of software) "that is commonly needed".  To be reusable, the component must be found, understood and trusted.  (Other points are made about iterating with the domain experts, having searchable metadata/information, ... that we have already noted in the synthesis.)

* (David Wade) Reuse requires education of the business and technical advantages, metrics providing evidence of increased productivity, and full participation in/relevance to the development effort.

Jack went further to say ...

Seems to me that any chunk of an ontology may be or may not be Reuseable depending on whether the specification of its attributes (delimiters) is sufficient. Further, that John Sowa's multifacted diamond structure specifies those attributes.

Seems to me that any chunk of an ontology may be reuseful when it is consistent with the system principles of Progress Properties (an end state can be reached from a current state in a finite number of transitions) and Safety Properties (certain aspects of the system are invariant during the Progress.

 

I want to look at these two statements independently, since I want to make sure that I am interpreting them correctly ...

1. Image removed by sender.Specification of a ontology's (or ontology chunk's) attributes, as given by Sowa's diamond structure - I am interpreting this to mean that a reusable semantic should define what objects and relationships are described (including whether the objects/relationships/states reflect changes over time), provide one or more schemas/definitions for them, include a history, describe why the ontology (or chunk) was created and how it is used for one or more purposes, etc.  

2. Consistent with the system principles of progress and safety properties - What might these properties be for ontologies (or chunks of ontologies)?  Does this translate to very specific properties like completeness and consistency?  Or, does it go further to describe more complex characteristics of ontologies such as whether reasoning, probabilities, functional expressions, n-ary predicates, etc. can be supported?  Or, did I misunderstand the paragraph entirely?

Lastly, I want to take JohnY's first sentence and translate it for ontologies ... Reuse is the ability to include one ontology (or piece thereof) in another, which requires that the concepts (including any axioms and rules), assumptions and _expression_(s) of the included ontology meet a need and match the concepts, assumptions and _expression_ of the including ontology.


 

On Thu, Mar 13, 2014 at 1:15 PM, John Yanosy Jr. <jyanosyjr@xxxxxxxxx> wrote:

Confirming this is the ability to swap out sw components within an application and framework as long as the framework interfaces and use of common services offered by the framework are satisfied by the component and as long as the interfaces and services implemented and exposed by the new component satisfy the application component model typically expressed as a set of methods contained within the interface signature. For ontology reuse what is the equivalent of a sw framework and the defined exposed app component interface and services offered by methods. Not sure the analogy is equivalent but it does demonstrate that practical reuse has and is occurring with sw components and web services. In both cases there were Meta model specs that had to be satisfied in order to reuse a component as well as some dynamic service behavior specs. In a new app model with reuse of a component the components behavior was well defined but the larger app which reused a component would have its own behaviors. Encapsulation and polymorphism concepts for sw components  seem to enable reuse in a sw component model. What are the equivalents for ontology. What is an ontology like component. What is encapsulation in this ontology component model. What is polymorphism.

Best regards,
John A Yanosy Jr
Mobile: 214-336-9875

On Mar 13, 2014 11:00 AM, "Jack Ring" <jring7@xxxxxxxxx> wrote:

In the early 1990's Sam Adams, guru of object technology, in the (then) Object Technology Practice, IBM Consulting Group, suggested the distinctions between reuseable, reuseful and reused (successfully). My recall may not be perfect but it went something like this.
Reuseable when disambiguated sufficiently to be referenced/selected without confusion,
Reuseful when content and structure are consistent with the context in which it is envisioned to participate, and
Reused (successfully) when it (as is or as modified) expanded the repertoire of the 'system' in which it participated without fostering unacceptable unintended consequences.

Seems to me that any chunk of an ontology may be or may not be Reuseable depending on whether the specification of its attributes (delimiters) is sufficient. Further, that John Sowa's multifacted diamond structure specifies those attributes.

Seems to me that any chunk of an ontology may be reuseful when it is consistent with the system principles of Progress Properties (an end state can be reached from a current state in a finite number of transitions) and Safety Properties (certain aspects of the system are invariant during the Progress.

On Mar 12, 2014, at 11:11 PM, John F Sowa wrote:

> Andrea, Jack, and Gary,
>
> I agree that a very underspecified definition can cover the many
> senses of a word.  But it adds very little information.
>
> AW
>> How about the following for our basic definitions:
>> * Reusable - "Capable of being used again" (from WordNet)
>> This begs the question of what makes semantic content
>> "capable" of being used again.
>
> As you said, it doesn't answer the question.  It relates the word
> 'reuse' to the word 'use', which has very little meaning in itself.
> If I say "I used X", that means I performed some action in which X was
> involved as a tool, a resource, a part, etc.  That doesn't say much.
>
> If I say "I reused X", it says even less.  I can reuse a DVD disk
> as a coaster for my coffee cup.  That's similar to Jack's example
> of reusing an article of clothing.
>
> AW
>> * Interoperability - "Ability of systems or organizations to work
>> together" (from Wikipedia)
>
> That's another underspecified definition.  Computer systems have
> been interoperating from the earliest days.  Whenever some human
> carried a deck of punched cards from computer A and loaded it in
> the card reader of computer B, the two computers interoperated.
> A high-speed line can make the interoperation faster, but there's
> no difference in principle.
>
> AW
>> There was a thread on the public-lod mail list that specifically
>> asked for this, "Evaluation of ontology reuse choices in real-world
>> scenarios" (Feb 20).  Unfortunately, there were no specific answers.
>
> That's the critical point.  Scenarios and case-histories get into
> the details.  It's too bad that there were no specific answers in
> that case.  I suspect that the many different partial answers were
> so diverse that it was hard to find useful generalizations.
>
> GBC
>>> Haven't we made progress on understanding several areas of semantic
>>> relations that can be reused?  Distinctions among different types
>>> of Part relations come to mind and are post-Aristotle.
>
> I didn't summarize all the discoveries of the Sumerians, Egyptians,
> Chinese, Indians, Greeks, Romans, Arabs, Indians, medieval Scholastics,
> Renaissance, etc.  Aristotle most certainly did analyze the many kinds
> of part relations.  As I said, Wilkins' ontology (17th c) was as good
> or better than many ontologies on the WWW.
>
> Roget's Thesaurus (19th c) is still widely used -- and it has a better
> coverage of adjectives and adverbs than WordNet.  I suggest that you
> compare the two.  Take any sample of words (nouns, verbs, adjectives,
> and adverbs) and look at the "concept neighborhoods" linked to them
> in both Roget's and WordNet.  It's very easy to compare them:
>
>  1. For Roget's, take any word and type it into the following demo:
>     http://www.ketlab.org.uk/roget.html
>
>  2. For WordNet, type the same word to
>     http://www.ketlab.org.uk/wordnet.html
>
> You might start with the verb 'explore'.  Then compare 'happy' and
> 'happiness' on both systems.  For software and documentation about FCA
> (Formal Concept Analysis), see http://www.upriss.org.uk/fca/fca.html
>
> AW
>> I agree that we do have common semantics already that need to be
>> highlighted - part, dependency, generalization/specialization, events...
>
> Yes, but...  There is a huge difference between the kinds of detailed,
> very precise ontologies needed for designing an airplane and the looser
> ontologies needed for answering a Jeopardy question.  LOD is somewhere
> closer to Jeopardy than to airplane design.
>
> AW
>> All of these semantics need more than just a definition and encoding
>> in RDF or OWL.  They need the backing tooling, metadata and repository
>> that we have highlighted.
>
> I certainly agree with the first sentence.  On the second, I believe
> that current research is just scratching the surface of what needs
> to be done.
>
> John
>
> _________________________________________________________________
> Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
> Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
> Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
> Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
> Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
> Community Portal: http://ontolog.cim3.net/wiki/


_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
Community Portal: http://ontolog.cim3.net/wiki/



_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
Community Portal: http://ontolog.cim3.net/wiki/

 


_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014  
Community Portal: http://ontolog.cim3.net/wiki/

 


_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014  
Community Portal: http://ontolog.cim3.net/wiki/     (01)
<Prev in Thread] Current Thread [Next in Thread>