Sean and Adrian,
My comments are below:
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of sean barker
Sent: Monday, August 16, 2010 1:02
PM
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum]
Looking for a razor
Adrian,
My
problem space is aerospace and defence, where getting things wrong costs lives.
One way of restating the problem is that I am looking for systematic ways
of evaluating the risk that a term will be misapplied, or applied inaccurately,
and also of minimising that risk, where risk is the product of the
probability of an occurrence and the impact of that occurrence.
From your
slides, Executable English is not concerned with that problem. In fact,
rhetorically, calling something "executable English" is a claim that
people already understand what terms mean. I would contrast your statement on
slide 19 "A term is defined by
the set of its superclasses in the taxonomy, and by its properties" with
my final comment "ontology classes
used by a business process are exactly those classes which label the
alternative routes onward from a decision process, and therefore define the
grounding of terms." The way many people would read your definition is
that terms are acontextual, whereas I would insist that terms are context
sensitive, particularly where conventional terms are used with a particular
technical sense by a business process. For example, in one business, a
"drawing" was a concept reified as a database entry, whereas a thing
with lines on showing the shape of a part was called a "drawing
sheet".
The problem of accurate
communication is therefore one of alerting people to the actual meaning of a
term in the context in which it is being used, rather than allowing them to
assume they already know what the term means. In terms of Executable English, I
can see that it might have applications where the impact of misinterpretation
of terms is limited, but I see it as part of the problem exactly because it
draws attention away from the need to manage such risks. This is more a
statement of requirements than a criticism of Eexecutable English
Adrian, I can echo Sean’s concern with precision in mission
critical systems. Words are thoughtlessly thrown around casual conversations,
but in ANY large database system, and ESPECIALLY in mission critical ones, the
interpretation of words is not evenly or well distributed; people have uniquely
different interpretants for the same word. This is the subjectivity problem again.
Ideally, we could insist that all people
be informed of, and use the one true unique meaning of the word among us, and
that would result in slightly improved synchrony for a tiny little while, but
the issue is that different people in the organization do different things with
the concepts and thereby have different issues to be resolved using the same
words. So the issues of overall effectiveness, even of identification and tracking
of all the Things involved, requires that every agent in the organization actively uses
the decreed meaning. Even if this polyannaish notion is entertained, by
definition, only terminal nodes in the problem reduction graph can source or sink new
information. Everything produced by groupings of those nodes (vocabulary, named classes, enumerations,
records, tables, columns, domains, ..) is NONterminal. (I hope my definition is familiar and not confusing to
anyone reading this).
But Periceans would say (or correct me) that
the vocabulary decreed and shared among the agents in the organization above is
all that need be considered.
I would respond (if they were to actually
ask me) that only one more property need be added to the system – the agent’s
unique designator, which should have thereby been included in the decreed ontology
as of the human kind, i.e. the kind of machine with valuable capabilities and very
little planning or control that can be both predicted and effectively enforced.
Further use of that designator domain is to track certain events of already
known kind, or of any exception message text, to create a problem trail as a
system runs. That only leaves the scatological evidence of a system problem.
So its NOT Executable English that is the
problem; this simply means that subjectivity HAS to be architected into
information systems down to the individual person. That’s why there are
so many views of the “same” database, when supposedly only one for
each job role in the organization is needed. EE, like a database developer,
has to recognize and treat that issue to be successful in solving practical
problems. So does the custom programmer and that is why there is such a gap
between the theorists and the actualists.
-Rich
From:
ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx]
On Behalf Of Adrian Walker
Sent: 09 August 2010 21:08
To: [ontolog-forum]
Subject: Re: [ontolog-forum]
Looking for a razor
*** WARNING *** This message has originated outside
your organisation, either from an external partner or the Global Internet. Keep
this in mind if you answer this message. Sean --
You wrote:
I am looking for a razor that can cut
between the "ontologies as a formal system" and "ontology term
grounding" parts of the discussion, and so ensure that both parts are
solved.
A candidate for your desired semantic razor is Executable English.
The diagrams on slides 14-17 and slides 51-52 of [1] illustrate this.
There's a runnable example on slides 35-43.
Cheers, -- Adrian
[1] www.reengineeringllc.com/Internet_Business_Logic_e-Government_Presentation.pdf
Internet Business Logic
A Wiki and SOA Endpoint for Executable Open Vocabulary English Q/A over SQL and
RDF
Online at www.reengineeringllc.com
Shared use is free, and there are no advertisements
Adrian Walker
Reengineering