[ontolog-forum] Role of SME in Ontology Deployment [was Using controlled

[Ali Hashemi <ali@xxxxxxxxx>]

Sorry for cross posting, I thought this would be relevant to the making the case summit as well. Please feel free to branch into whichever forum is more appropriate.

In the previous thread [1], it seemed to be taken for granted that it is desirable for a subject matter or domain expert to be able to follow the formalism of an ontology.

I would like to query this forum for people's views on the role of SMEs in the creation and deployment of ontologies.

In a later email to that thread, 

John Sowa wrote:

Fundamental principle:  the primary purpose of the CNL is not
for the SME to verbalize the ontology, but for the SME to read
and verify the verbalization created by the KE.

While I find this to be in principle, a laudable goal, I wonder if it is actually necessary and might not possibly be a distraction of attention and resources. 

Note: I am not suggesting that CNL's are not useful or anything like that, the focus is on how much we should expect an SME to follow the machine readable representation of the ontology.

In the same vein, Dave and Rich lamented:

David Eddy wrote:

HOW can ontologies be delivered to people who have zero interest?  If
it is "necessary" to have a solid grounding in formal logic, the
effective market for ontologies will be zero.
...
If there is to be sustained use of ontologies, the ontologists MUST
step back & get out of the loop.

Rich Cooper wrote:

Darn those SMEs, they just won’t take these amazing tools even if you give them to the SMEs and make them use them right!  I wonder if there is ANY domain oriented reason why they don’t follow the approved methods.  Could it be they know something?  Naaah.  

Why is it beneficial to have an SME be able to read the innards of a deployed ontology? In his email, Dave had mentioned how a car driver does not need to know the inner workings of an engine to drive a car [2], and in the same thread I mentioned how nuclear engineers don't need to know how the control room was designed to use the control room [3]. What benefit do we gain from exposing SME's to a particular formalism or machine readable segment of an ontology?

At this point, I should mention that I have found this work to be very useful in helping clarify for me, the role of ontologies in an organization (http://stl.mie.utoronto.ca/publications/ker.pdf) [4], specifically pages 28-32 (in the print numbering, starting at section 6.2) regarding (in)formal competency questions.

In my mind, the verification of whether the ontology is true to the fragmented mental model (I use the term loosely) of a particular subject matter expert, comes from the ability of the ontology to answer the competency questions accurately, more so than following how their knowledge was represented in a particular machine readable formalism.

I submit the following reasons for this claim:

• First, drawing on the analogy to cars and control rooms above, it seems curious to expect the same familiarty from the SME. Or is the analogy incorrect, and in fact the SME is the car manufacturer or a user-interface specialist? In this case then, the SME is the ontologist, yet I don't think this is accurate. 
• Second from the fact that in many ontology deployments, any subject matter expert will only be able to follow only a fragment of the whole. Meaning, that a particular representation choice that takes into account other fragments might not superficially strike a particular SME as intuitive or "right". 
• Thirdly, because a particular choice of an axiomatization might be unwieldy and meaningless to an SME. Again, let us consider the role of orthocomplementation or pseudocomplementation in mereotopologies [5]. I can imagine that large number of people would consider themselves experts in parts, wholes and connections. A cartographer or geologist might have intimate knowledge of what it means for some rock formation to be part of another, and connected to another. Would they really understand why the notion of orthocomplementation (however it is rendered, whether in OWL, CLIF or some controlled english) is written as is? Would their understanding not be illuminated rather by seeing what models (Tarskian, but rendered in an accessible way) are allowed (or ruled out) by this representation choice. And does this not actually come into focus, only by considering the competency questions and the actual output of the ontology?
• Bringing us to the last point (which also hopefully addresses Dave's query) - the real utility of the ontology is how it fares with respect to the specified competency questions. Indeed, aside from a key, valuable resource helping the ontology designer develop axioms which match the intended models of the SME [6], the creation of the competency questions seems to be one of the most important roles for a subject matter expert.

Lastly, in many deployments, the ontology is part of a larger system. It might have been captured at a highly expressive level, and deployed in pieces and fragments throughout a complicated system, perhaps each fragment only capturing a less expressive part of that larger bit. Should an SME follow both the highly expressive and more restricted choices? 

What do others think?

Is it the case that most ontology deployments will have formalization choices that are readily transparent and grokable to an SME? Is it necessary? Is it the best use of an SME's expertise or time? Why might evaluation of an ontology with respect to competency questions not be adequate? In what ways does the car and/or nuclear control room analogy fail? Which of the assumptions that I've made are contentious or strike you as having missed the point? Why?

PS 

I have training in something called Cognitive Work Analysis[7] and Ecological Interface Design, where one does need to consult subject matter experts, among many other resources to build a view (model / conceptualization, i use the term loosely) of a complex socio-technical system. This information is then used to design and deploy varyingly from the level of control rooms in oil refineries, to the actual displays on a particular screen to present the most useful information to a particular user for a given function. It is not expected that the domain expert understand and follow the laborious process of the creation of the design specification for the control room layout or how each item is to appear on their screen or other interface media. What is expected, is that the system aid the user and/or SME in their "cognitive work".

Thank you kindly,

Ali Hashemi

[1] http://ontolog.cim3.net/forum/ontolog-forum/2011-03/msg00025.html

[2] http://ontolog.cim3.net/forum/ontolog-forum/2011-03/msg00084.html#nid07

[3] http://ontolog.cim3.net/forum/ontolog-forum/2011-03/msg00063.html

[4] Uschold, M. and Grüninger, M.(1996), Ontologies: Principles, Methods, and Applications, Knowledge Engineering Review, 1:96-137. ( http://stl.mie.utoronto.ca/publications/ker.pdf )

[5] Hahmann, T., Gruninger, M. and Winter, M. (2009) Stone Ortholattices: A new approach to the mereotopology RT0, to appear in Journal of Artificial Intelligence. ( http://stl.mie.utoronto.ca/publications/RT-journal.pdf )

[6] Guarino, N. (1998), Formal Ontology in Information Systems In N. Guarino (ed.) Formal Ontology in Information Systems. Proceedings of FOIS'98, Trento, Italy, June 6-8, 1998. IOS Press, Amsterdam, pp. 3-15 ( http://www.loa-cnr.it/Papers/FOIS98.pdf )

[7] http://cel.mie.utoronto.ca/research/frameworks/cwa.htm

--

(•`'·.¸(`'·.¸(•)¸.·'´)¸.·'´•) .,.,

_________________________________________________________________
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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)