[Top] [All Lists]

Re: [ontolog-forum] standard ontology

To: "Ontolog-Forum-Bounces" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Sean Barker" <sean.barker@xxxxxxxxxxxxx>
Date: Thu, 12 Feb 2009 16:21:48 -0000
Message-id: <OOEEJGAPCAJOKOFFPHLHOEBCCAAA.sean.barker@xxxxxxxxxxxxx>
In the 90's (?) the phrase "Islands of Integration" went rapidly from
neologism to cliché, and people tried building integrated infrastructures
and developing corporate data models. This turned out to be far too
difficult even with a single organization, and we have been left with
archipelagos of integration. Even sorting out this level of problem is
likely to keep me gainfully employed until I retire.    (01)

>From a practical viewpoint, I would doubt that there is a business case for
building a single standard ontology. Even using traditional data
exchange/integration techniques:    (02)

1) It was once calculated that if we implemented all the IT projects
proposed, the benefits would be so great that we could give aeroplanes away
for free and still make a profit - benefit claims are taken with a pinch of
salt.    (03)

2) Any business large enough to invest in a "grand ontology" would be
managed at a set of independent projects/departments/etc. The marginal cost
of implementing such a system (rather than a point to point solution) is
generally greater than the benefit to any single department, and therefore
it is difficult to get approval at the departmental level. Also, such
managers are looking for a return on investment of 3 years.    (04)

3) Any business large enough to have lots of departments which could benefit
from corporate investment is so large that these investments only get two
minutes consideration at the levels which make such decisions. This sort of
problem is often too abstract to get the top level buy-in needed to get it
properly financed.    (05)

Therefore, I would agree with John that interoperation is generally specific
to a particular domain, but observe that, although the member of domains is
open ended, its often more effective to talk to people than try and
interoperate across multiple domains. I interact with the company purchasing
system a couple of times a year. The problem I have is not the semantics of
interoperation, but the process - which screen do I start with, and what
values to I type in. The semantic interoperation plan starts with "ask
someone who has used the system recently", followed by, "ring up finance and
ask which button to press".    (06)

Noting some of the comments earlier in the thread, in all these
interactions, one thing I never need to consider is whether clay or the
number seven is an individual. Is there any way of positively asserting that
I don't care? (RDF by preference)    (07)

Therefore: focused domain ontologies credible. Interoperation between
different ontologies for the same domain interesting. Grand unified ontology
across domains - most unlikely to get industrial support in the next ten
years.    (08)

Sean Barker
Bristol    (09)

On 11 Feb John Sowa wrote:
.................................................    (010)

Pat H, Pat C, and Ed,    (011)

PH> Hardly anyone needs to be able to interoperate with everybody.    (012)

PC> ... but a lot of people want to interoperate with a lot of
 > other people,    (013)

Actually, we all interoperate with an enormous range of different
kinds of people -- but usually on very specific domains.    (014)

Whenever we drive down a highway, we interoperate with all the
other drivers on the road according to a well-defined set of
conventions.  Whenever we buy anything at a store, we interoperate
with people with very different backgrounds on a narrow domain.    (015)

EB> Well, yes and no.  A lot of people (and more importantly
 > their software) need to interoperate with others in the same
 > domain (= company/industry/profession) and with others in
 > joint activities that cross specific domains.    (016)

Every company with more than 2 employees has people who
specialize in different areas.  The financial people working
for a company have more in common with financial people in
other companies than they do with the engineers in their
own company.  When you consider sales, legal, maintenance,
manufacturing, shipping, research, plumbing, air conditioning,
cafeteria, janitorial, etc., the number of different kinds
of people we interoperate with every day is enormous.  )    (017)

Fundamental principles:    (018)

  1. Interoperations are nearly always specific to a particular
     domain.    (019)

  2. But the number of different domains is open ended, even
     for a single company.    (010)    (020)

John    (021)

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    (022)

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