Foundational Ontology

[bruces...@cox.net]

Interesting to see the call for articles and proposals for “foundational ontologies” posted here on April 15.

 

5th International Workshop on Foundational Ontology (FOUST)

Website: https://foust.inf.unibz.it/foust5/

 

I am experimenting with an idea for a “universal upper ontology”. I’d say this approach is something fundamentally different from existing domain-specific ontologies – but it might be intended as a foundation for them all.

 

One helpful critique of any program for a universal upper ontology is these remarks on Wikipedia:

 

https://en.wikipedia.org/wiki/Upper_ontology#Arguments_for_the_infeasibility_of_an_upper_ontology

 

The project that seems to be unfolding for me – is an exploration of this design:

 

• Develop an approach to a "universal upper ontology" that is defined in terms of general principles instead of specific definitions.
• Rather than attempting to define objects in a domain-specific list, create a synthesis of fundamental principles from mathematics and semantics -- on the basis of which any domain-specific ontology could be defined.
• What are the dimensions of this object?  What are the dimensions of this event?  What logical/arithmetic conclusions should we derive?

 

This is a very ambitious idea, which I see as possibly involving a “stack” of fundamental principles which are combined in a way so as to universalize the creation and definition of any semantic object.

 

What are the common foundational principles which govern the formation of a domain-specific ontology (make a list)?  Can they all be identified and sequenced, so that special-case languages can be defined from a common root, based on critical but common dimensions of variation?

 

This approach does involve confronting a venerable history of controversy and fragmentation and confusion, and a stack of problems that have not been adequately solved.

 

But just suppose – maybe something along the line of Wolfram “Mathematica” – it were possible to define a universal schematic for defining algebraic and semantic objects – like “words” or “word meanings” – and that this schematic was anchored in primary algebraic definitions – starting with the number line and the continuum, and then the linear sequencing of integers which creates a common ordering principle extending throughout the framework.

 

Years ago, I developed a dimensional model of semantics, based on the concept of “synthetic dimension” – which is a dimension with decomposable values as its “units”.  This approach creates an algebraic model of “qualitative” values – and is not unheard of in contemporary mathematics.  https://en.wikipedia.org/wiki/Level_of_measurement

 

Let’s grant this might be possible.  Synthetic dimensionality is at the least an open door to exploration.

 

So, the “stack” defined in a top-down way might include:

 

• The continuum
• The number system, including integers
• Arithmetic
• Boolean algebra
• A natural language semantics combining these elements and based on stipulation and intention – “words mean what we want them to mean”

 

Can we build any semantic or ontological object from these principles, in a systematic and common way?

 

The idea would be to spawn a comprehensive special-purpose vocabulary and domain ontology from a universal and internationally recognized body of relatively simple common principles.

 

It looks to me as though the difficulty or impossibility of creating a universal upper ontology is that professional ontologists have been trying build frameworks in a bottom-up way, since they are compelled to deal with the actual facts of some industry or domain as they are – as they are found today, in practice and observable.  Nobody knows how to define all these objects in a universal way, so the tendency is to give up and go the safe route.  Stick to  bottom-up empiricism.  Maybe not magic, but it works – at the price of some incommensurateness.

 

Could it be possible, is it possible, to entrain the fundamentals of mathematical definition in a universal way, to support all processes that any ontology must support, all spawned from a common body of relatively simple universal principles?

 

Maybe somebody can show me this is impossible.  But I am feeling a pretty strong push in this direction.

 

Bruce Schuman

Santa Barbara CA USA

bruces...@cox.net / 805-705-9174

www.origin.org  / www.integralontology.net

 

 

 

 

[Michael DeBellis]

Bruce, here is an idea I've been kicking around lately: In the last year or so when I build an ontology I often start with some entities from various common vocabularies like Dublin Core and FOAF. What about a practical upper model? I.e., an upper model that incorporates some of the most common reusable entities (e.g., foaf:Agent class) from the most widely used vocabularies? It seems to me if reuse is really are goal this would go a long way to making ontologies more consistent and reusable. 

Michael

[dr.matt...@gmail.com]

Dear Michael,

The problem is that if you keep doing that for long enough you will end up with and inconsistent mish-mash.

Regards

Matthew West

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/486a7263-dfb6-491f-a0d8-09dd8e38b89fn%40googlegroups.com.

[William Frank]

+1.  I'd put it this way. 

 A hierarchical organization of objects requires that the children of a parent partition the set of parent objects, so that each parent object is in exactly one child object.  That is to say, the organization is orthogonal.  In order to achieve orthogonality, we need to use a consistent discriminator to define each child.    For example, large beach balls and blue beach balls does not partition beach balls.  The color or the size needs to be the discriminator.  

It is odd to me how often this principle is ignored, leading to lots of practical confusions.   For example, a common U.S. money management division for securities is into fixed income, equities, and 'international' meaning non-us.    The discriminator used for fixed income is cash flow from the investment, (i.e., the other category should be variable income) for equities it is the way in which the asset owned is realized (debt, ownership of a corporation, commonidities, ..., options).  So, some equities, such as preferred stock, provide fixed income, and some debt provide variable or no income, and 'internationals' can be any of the above.)  Just so some marketing departments classify customers into high value customers and international customers and multi-product customers.  The settlement windows at a major bank were divided into those on the 23rd floor, the derivatives windows, and the hard asset windows. OTOH, this may be culturally, unconsciously deliberate, as obscurity is a way of protecting the exclusivity of a profession, as explained in Information Politics 

All this necessitates that classification schemes be designed to have certain mathematical properties, and I believe is part of Dr. West's consistency requirement in Developing High Quality Data Models.   It's also emphasized in the works of David Hay, and most everything about UML.  Knowledge of the old wines of data modelling and object modelling is the most important foundation for work using new ontology bottles.  Yet, like all other subjects in computer science and engineering, we don't really do engineering, because we don't build on past experience and knowledge, but start over again and again.  Ironically, ontology seems no exception. 

Wm

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/01c501d73dc0%24009bbaa0%2401d32fe0%24%40gmail.com.

--

Wm

We understand what other people say through empathy—imagining ourselves to be in the situation they were in, including imaging wanting to say what they wanted to say.  

– Zellig Harris    

[bruces...@cox.net]

Something like this issue came up during the Ontology Summit on Wednesday of this week, in the discussion of “terminology.”

 

It was (briefly) emphasized that the line of “semantic descent” from a broader abstraction to its particular subsets or “children” has to be 100% rigid, or a “mishmash” will result.

 

Of course that is true --  but with some caveats.  That rigid cascade of definitions can go out of date quickly and will require updating.  And of course, this is not how meaning works in human psychology, where adaptive flexibility and “context specific” definitions are essential.  “Word senses” constantly change in context-specific ways.  What is the meaning of the word “bank”?

 

In the summit meeting, somebody brought up the issue of metaphor, and mentioned the use of the term “my old flame” (my old romantic interest).  It was specifically mentioned that you could not really substitute the phrase “my old wildfire”.  No one would understand you – not at least without some explanation and maybe a wink of the eye….  

 

I’d like to see this problem of adaptive flexibility addressed in a universal context.  This is asking a lot, I know – but the existing industry standards seem to be very brittle, and a concession to the problem of silos and logical fragmentation (no capacity to adapt to specific context or changes).

 

A solution to this problem would have to be built into the basics of a universal (“one size fits all”) ontology.  Wondering how that might be possible is a fascinating exercise in creative algebraic semantics.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALuUwtBZHDifiX0-3i2yt32CdqQvsaddWH7Eu-0B5%2Bbs45hwhQ%40mail.gmail.com.

[Michael DeBellis]

The problem is that if you keep doing that for long enough you will end up with and inconsistent mish-mash.

I think that is a risk for any standard. A long time ago I was active in the HL7 Healthcare standard for messaging across healthcare systems. Each organization (HP, Accenture, various electronic medical record vendors, middleware vendors, etc.) had their own ideas on what was most important to add or change in the standard. If each group had developed their own version of HL7 it would have been an "inconsistent mish-mash" but that is why as much as they are a pain we need standards groups.  HL7 hashed out their ideas, made compromises, and came away with something that no one was completely satisfied with but actually worked pretty well in enabling integration of what were highly siloed Healthcare systems for labs, admissions, EMRs, billing, etc. If on the other hand they had started from some armchair philosophy model rather than looking at what practitioners actually do I think the result would have been something that few people would actually use. 

So I agree if I take what I need from FOAF and DC and you take what you need, etc. it's a mess. What I'm proposing is that whatever group is working on an upper model standard for OWL focus on actual practice rather than armchair philosophy. I've been reading some of the philosophical papers that provide the justification for various upper models. They are IMO filled with highly debatable interpretations of the history and the current state of philosophy (and they essentially ignore psychology which if we are going to look for theoretical foundations is at least as important as philosophy). So I challenge the whole dogma that arm chair philosophy should be the basis for upper models in the first place. Set theory is well established as the formal foundation for mathematics. The various attempts at modern metaphysics are highly debatable as having any value at all. They claim to be a foundation for science but no science I've ever read cares about continuants or occurents. And I'm only mentioning BFO because it is the most popular, I think the same thing applies to any armchair philosophy based upper model.

I realize that term may seem a bit harsh but I think that it is appropriate in this case. There is philosophy that I don't consider "armchair". For example, the work by Frege, Russell, Zermello, Frankel, Godel, etc. to define a formal model that is the foundation for mathematics (which ultimately resulted in success in the form of ZFC set theory) or the work of Tarski to develop model theory or Chomsky's work to extend the theory of computation to apply to the various kinds of possible languages and the type of sets they can recognize and generate. That work has formal proofs and consequences for design of theorem provers, compilers,  NLP, and theoretical analysis of software. The kind of philosophy where people just speculate about various concepts that they think are somehow more fundamental than others is what I think of as "not even wrong" because there is no way to actually evaluate it against alternatives, at least I haven't seen any. 

So I think it is worth considering a whole new approach to upper models based on practitioners not on philosophy. It may be that such a process would create new models (e.g., for common patterns like lists, part-whole relations) or update some existing vocabularies (last time I looked FOAF was in need of an update to take full advantage of OWL). But ultimately, I think that kind of a process would yield an upper model that would be much more useful and would facilitate reuse more than one based on armchair philosophy. 

Michael

You received this message because you are subscribed to a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/VVjS-nKbp6I/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/01c501d73dc0%24009bbaa0%2401d32fe0%24%40gmail.com.

[William Frank]

I agree with you completely about the futility of 'armchairing' one's way to the "right" upper ontology.   But more so, I am quite sure, considering the varieties of human languages, that there *IS* no 'right' one.   There may be useful ones, and the way to construct the same is indeed to determine what has proved useful for domain ontologies.  

My point was only that the orthogonality of classifications is something, which, if you avoid when you choose to classify, you start with inconsistent mismatches, as it seems pretty obvious.   "OK, Michael, I' ll take the red balls and you can have the large balls?"    "Sally and I divide up our responsibilities, so we don't step on each other's toes: I am responsible for the communications equipment, and Sally here is responsible for the telephones????" I think people need to walk on the foundations created by others, before they can run.  And, here, even intuitionists who reject the law of the excluded middle (they say that not everything is either an X or is not an X) do not replace it by rejecting the rule against inconsistency (They don't say, some things are both x's and not x's at the same time).  In rigourous domains, where we expect a clear definition of x).  

The more fundamental thing is that hierarchical classifications, while often critical for special purposes, just *don't reflect the human thought scape*.  But the vagueness of words (edge cases), Bruce, the fact that they change their meaning over time, the fact that there are special uses, that meanings are often a matter of chains of family resemblances, is an entirely different issue than that of orthogonality.  People have mentioned George Lakov often enough in these mails.   

So, *If* you want to accommodate more accurate models of people's labeling of stuff, then a taxonomy tool such as SKOS, which eschews hierarchies in favor of overlapping word relations like 'broader' and 'narrower' will work better than the hierarchy which ontologies expect.  Ontologies, and object oriented software are modelled after Aristotle's classifications, done initially for animals.  Works well in biology and other sciences.   Has not proven nearly so useful in representing the kinds of human artifacts that most software is about, where inventions involve permutations and combinations of features.  

Standards become a mess for lots of reasons, but I have never seen a standard that used a non-orthogonal classification system.    I doubt that people contemplating such a standard would not recognize its inconsistency.  

Wm

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFikcq_2V2ypk-Uwq5XS1Hz%2B77Kt0ofUYVPqi0_rW--%2BMAOA%40mail.gmail.com.

[John F. Sowa]

William and Michael,

I mostly agree with both of you.

.
WF> I am quite sure, considering the varieties of human languages, that there *IS* no 'right' [ontology].  There may be useful ones...

If there were a universally valid top-level ontology, the 200,000 years of natural language evolution by Homo saps would have converged to it.  Instead, all NLs are sufficiently flexible that they can be adapted to anything.  Today, all of us are using the core syntax and vocabulary of our stone-age ancestors, but we can adapt it to anything and everything that any scientist, engineer, artist, or politician can throw at us.

MDB> HL7 hashed out their ideas, made compromises, and came away with something that no one was completely satisfied with, but actually worked pretty well in enabling integration of what were highly siloed Healthcare systems for labs, admissions, EMRs, billing, etc.  If on the other hand they had started from some armchair philosophy model rather than looking at what practitioners actually do, I think the result would have been something that few people would actually use.

The reason why HL7 works is that it avoided anything that philosophers -- either armchair or university chair -- concocted for them.  There are only three philosophers I would trust:  Peirce, Whitehead, and Wittgenstein.  The reason why I trust them is that they would never trust anybody who concoctd a Top-Level Ontology.

I wrote an article about them:  "Signs, Processes, and Language Games, Foundations for Ontology", http://jfsowa.com/pubs/signproc.pdf

The ontology I would recommend has a top level with just one node, called Entity.  Beneath it are infinitely many branches, but you only select as many as you need.

John

[Ítalo Oliveira]

Dear colleagues,

Let me give you my two cents on the feasibility and utility of an upper ontology (aka foundational ontology).

The lack of a foundational ontology supporting the construction of a domain ontology is not a problem per se. However, studies have shown that foundational ontologies significantly contribute to prevent and to detect bad ontology design [1], improving the quality and interoperability of domain and core ontologies [2].

Modeling domain and core ontologies without making explicit the underlying ontological commitments of the conceptualization gives rise to semantic interoperability problems. In fact, there is a strong connection between the ability of articulating domain-specific notions in terms of formal ontological categories in conceptual models, and the interoperability of these artifacts [3].

Semantic interoperability is also hindered by the sole use of languages such as OWL, which merely address logical issues neglecting truly ontological ones [3,4]. For example, ontologically neutral logic languages usually make no difference between being red (color) and being an apple; though people understand these things as fundamentally different, they are both formalized in FOL as unary predicates  like in (Red(x) ∧ Apple(x)). On the other hand, once meaning negotiation and semantic interoperability issues have been established by the usage of an ontologically well-founded modeling language, knowledge representation languages such as OWL can be employed for ontology implementation if necessary [4].

Moreover, there are things that can only be made by using foundational ontologies, provided you have the right tools. Ontology-driven conceptual modeling languages, such as OntoUML which is based on Unified Foundational Ontology (UFO), present several modeling patterns derived from its underlying ontology; instantiating these patterns allows the user to make more consistent models and more fastly. Another thing is model summarization for dealing with the complexity of the model: once you can take advantage of underlying ontological distinctions, you have very clever and quick algorithms for partitioning or summarizing the model according to the important things in it; traditionally graph algorithms have been using for this, ignoring those ontological differences.

As far as I can see, foundational ontologies and well-founded domain ontologies are under-explored in a sense that there are many open issues and research directions to be explored. Looking at the UFO/OntoUML's development and its tools (OWL mapping, anti-patterns detection, patterns instantiation, syntax checking etc.), we can see there is a lot of work to do.

[1] Schulz, S.: The role of foundational ontologies for preventing bad ontology design.In: 4th Joint Ontology Workshops (JOWO). vol. 2205. CEUR-WS (2018)

[2] Keet, C.M.: The use of foundational ontologies in ontology development: an em-pirical assessment. In: ESWC. pp. 321–335. Springer (2011)

[3] Guizzardi, G.: Ontology, ontologies and the ”I” of FAIR. Data Intelligence2(1-2),181–191 (2020)

[4] Guizzardi, G.: The role of foundational ontologies for conceptual modeling and do-main ontology representation. In: 2006 7th International Baltic Conf. on databases and information systems. pp. 17–25. IEEE (2006)

Kind regards,

--

All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/096d7baf282b3834f652099540560eeb.squirrel%40webmail2.bestweb.net.

--

Ítalo J. S. Oliveira

+55 81 9 9645 4629 (Whatsapp)

+39 3703672409 (Telegram)

Academic Website

Blog: Not so Short Notes

[dr.matt...@gmail.com]

Dear Colleagues,

I agree with Italo.

When people say they don’t have an Top Level Ontology, what they are really saying is that they do not know what it is.

It is true that there is more than one possible Top Level Ontology, but that does not mean it is unhelpful to know what yours is. In constructing your ontology there are a number of ontological choices that you make, knowingly or unknowingly. It is better to make them knowingly, since then you have some insight. The worst thing is not to make a clear choice, because then you might flip-flop in different parts of your ontology leading to internal inconsistency, or allowing inconsistency in its use.

In the UK National Digital Twin programme, as part of our due diligence on TLOs we undertook an analysis of a wide range of ontologies and the commitments they made. Here is a link to the overview of the way we categorized different ontologies.

https://digitaltwinhub.co.uk/a-survey-of-top-level-ontologies-ontological-commitment-overview/4.1

Regards

Matthew West

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAOEC0q9P%2BUjS8ab_j3gPjYVGTBhKJ-e2yLjhOsEDf%3DHFHvidug%40mail.gmail.com.

[Tom Tinsley]

Michael DeBellis said, “So I think it is worth considering a whole new approach to upper models based on practitioners not on philosophy. It may be that such a process would create new models (e.g., for common patterns like lists, part-whole relations) or update some existing vocabularies (last time I looked FOAF was in need of an update to take full advantage of OWL). “

 

I fully agree. As a beginning student of Charles S. Pierce, I believe he gave us the direction we need with the Piercean categories. The focus of ontologies appears to be only on Possibility and Actuality. Necessity seems to be ignored. In Pierce’s time, the interpretant in Necessity was a person. When an ontology is applied today in digital systems, the interpretant is the digital system.

 

Where are the ontologies for the digital systems?  They do exist in systems technology. They can be found in existing products from IBM, Oracle, Microsoft, RedHat and others. The predominant model is the implementation of Services.

 

Service models are universal across all knowledge domains for digital systems. I wrote my first paper on this years ago, https://www.scitepress.org/Papers/2011/37193/37193.pdf. Efforts are underway to recognize the importance of identifying the services associated with an ontology. For example, the intended service usage of HL7:  http://www.hl7.org/special/committees/soa/overview.cfm.

Ontologies must be prepared so they can be understood by the interpretant. For a digital system interpretant, the services and messages should be defined in Actuality within a Possibility ontology, as in the following example:

The Actuality of this model is not a guideline for implementation, but rather the architecture implemented by a framework that verifies consistency and manages all service execution. Each specific service could be implemented using a selected method: program languages, rules, neural networks, quantum computing, etc. If a service implementation can be whiteboxed, even further consistency could be verified.

 

I agree with Michael DeBellis, specifically if the interpretant is a digital system.

 

 

From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of Michael DeBellis

Sent: Friday, April 30, 2021 1:56 PM
To: ontolo...@googlegroups.com
Subject: Re: [ontolog-forum] Re: Foundational Ontology

 

The problem is that if you keep doing that for long enough you will end up with and inconsistent mish-mash.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFikcq_2V2ypk-Uwq5XS1Hz%2B77Kt0ofUYVPqi0_rW--%2BMAOA%40mail.gmail.com.

[Jon Awbrey]

Dear Tom, Michael, Matthew, John, and All ...

Welcome to Pragmatism, where we detail praxes directed toward pragmata
and inquiries as activities directed toward knowledge as an objective.

Having encountered these epiphanies time and time again I know what
is more than likely to happen next ... people give a nod to process
over commodity and just as quickly nod off again ... slipping into
the slogmatic dumbers of espying Peirce through Frege–Russell eyes.

Looking back over the past two decades of watching this happen,
it strikes me part of the reason for the regress may be a habit
of sticking foundations on top of the building rather than where
they belong at the bottom. Analyzing activities, practices, and
procedures, as we do in developing a program, may appear to work
from the top down in a stepwise refinement fashion but the nodes
at the top are nothing but nominal tokens till we manage to cash
cash them out and fund them in the fundamental procedures at the
recursive base where the real action starts.

Taking this view leads to radically different sorts of orderings
from the Great Chains of Ontological Being we are wont to see in
the ontology biz today.

Just for instance —

Fig 1. Normative Science Rests Largely On Phenomenology And On Mathematics
https://inquiryintoinquiry.files.wordpress.com/2014/08/peirce-syllabus.jpg

Regards,

Jon

On 5/1/2021 12:27 PM, Tom Tinsley wrote:> Michael DeBellis said, “So I think it is worth considering a whole new

approach to upper models based on practitioners not on philosophy. It may be that such a process would create new models
(e.g., for common patterns like lists, part-whole relations) or update some existing vocabularies (last time I looked

FOAF was in need of an update to take full advantage of OWL). “

>
> I fully agree. As a beginning student of Charles S. Pierce, I believe he gave us the direction we need with the
Piercean categories. The focus of ontologies appears to be only on Possibility and Actuality. Necessity seems to be
ignored. In Pierce’s time, the interpretant in Necessity was a person. When an ontology is applied today in digital
systems, the interpretant is the digital system.
>
> Where are the ontologies for the digital systems? They do exist in systems technology. They can be found in existing
products from IBM, Oracle, Microsoft, RedHat and others. The predominant model is the implementation of Services.
>
> Service models are universal across all knowledge domains for digital systems. I wrote my first paper on this years
ago, https://www.scitepress.org/Papers/2011/37193/37193.pdf. Efforts are underway to recognize the importance of
identifying the services associated with an ontology. For example, the intended service usage of HL7:
http://www.hl7.org/special/committees/soa/overview.cfm.
>
> Ontologies must be prepared so they can be understood by the interpretant. For a digital system interpretant, the
services and messages should be defined in Actuality within a Possibility ontology, as in the following example:
>

> The Actuality of this model is not a guideline for implementation, but rather the architecture implemented by a
framework that verifies consistency and manages all service execution. Each specific service could be implemented using
a selected method: program languages, rules, neural networks, quantum computing, etc. If a service implementation can be
whiteboxed, even further consistency could be verified.
>
> I agree with Michael DeBellis, specifically if the interpretant is a digital system.
>
>

> From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of Michael DeBellis

> Sent: Friday, April 30, 2021 1:56 PM
> To: ontolo...@googlegroups.com
> Subject: Re: [ontolog-forum] Re: Foundational Ontology
>

> The problem is that if you keep doing that for long enough you will end up with and inconsistent mish-mash.
>
> I think that is a risk for any standard. A long time ago I was active in the HL7 Healthcare standard for messaging
across healthcare systems. Each organization (HP, Accenture, various electronic medical record vendors, middleware
vendors, etc.) had their own ideas on what was most important to add or change in the standard. If each group had
developed their own version of HL7 it would have been an "inconsistent mish-mash" but that is why as much as they are a

pain we need standards groups. HL7 hashed out their ideas, made compromises, and came away with something that no one
was completely satisfied with but actually worked pretty well in enabling integration of what were highly siloed

Healthcare systems for labs, admissions, EMRs, billing, etc. If on the other hand they had started from some armchair

philosophy model rather than looking at what practitioners actually do I think the result would have been something that

few people would actually use.
>

[John F. Sowa]

Dear Matthew,

MW> When people say they don’t have a Top Level Ontology, what they are really saying is that they do not know what it is.

I would go one step farther: I don't believe that 99% of the people who use the term "Top Level Ontology" have any idea of what is needed for a good top level ontology.

Short test:  If they mention the word 'OWL', they have no idea what a good ontology is.

I have been in the ontology business for about 40 years.  I was one of the early adopters of the word 'ontology' in the computer field.  I promoted that word in my 1984 book, but I now wish that I had done more to promote the word 'terminology'.  There is more than a century of solid work in terminology.  By contrast most of the work in ontology is amateurish.

For some discussion of the issues, see the slides for my keynote talk at the 2020 European/Extended Semantic Web Conference:  http://jfsowa.com/talks/eswc.pdf .

There is much more to say, but those slides (which are an extended version of the ones I presented) discuss the directions for the future, which does not include hand-coded ontologies.

John

[Azamat Abdoullaev]

John F. Sowa wrote:

"I would go one step farther: I don't believe that 99% of the people who use the term "Top Level Ontology" have any idea of what is needed for a good top level ontology.

Short test:  If they mention the word 'OWL', they have no idea what a good ontology is.

I have been in the ontology business for about 40 years.  I was one of the early adopters of the word 'ontology' in the computer field.  I promoted that word in my 1984 book, but I now wish that I had done more to promote the word 'terminology'.  There is more than a century of solid work in terminology.  By contrast most of the work in ontology is amateurish".

A sad truth...

--

All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/98a992de2b6b9b7ead0ca3c50643fe7a.squirrel%40webmail2.bestweb.net.

[Tom Tinsley]

John Sowa,

 

With great respect for your knowledge and contribution, let me counter your test with this:

 

The thousands of Software Engineers that use the language “OWL” follow the design principle of “Good Enough”.

 

Although, I do agree there must be a better understanding of what is a “foundational ontology”.

 

What I see is a failure of Academia to provide foundational ontologies. The CIP (Classification of Instructional Programs) used by universities in the U.S., modularizes instructional programs into a hierarchy of dependency. Shouldn’t there be foundational ontologies introduced in each module that builds upon lower-level modules and modules from other hierarchies?

 

What I also see is Business developing foundational ontologies. In the U.S., the NAICS (North America Industry Classification System) is used to define a business function. Most of the ontologies developed could be classified within one of these codes. Unfortunately, most efforts are not modularized, obscure potential foundational ontologies and do not draw from foundational academic ontologies. They actually appear to be more like the conceptual model for a subject-matter database.

 

One change agent is the Textbook Open Knowledge Network. As you know, they are working to uncover foundational ontologies as a means to deliver a new form of a digital textbook. They are up against publication companies and professors that benefit greatly from selling books. The loser in this is the student and society as a whole.

 

Foundational ontologies are important. They should represent the principles on how we view the world. They used to change slowly. The encyclopedia once was a set of books with annual updates, now we have Wikipedia. Once I did my research at the library, now I use Google.

 

I set up a demo to show how ontology can be used to manage a pizza store chain and their suppliers. The core process is one of buying and selling. I looked at FIBO, but it is not modularized and was an overkill for my demo. I prepared a simple ontology of the REA foundational model as described in academia. I had to prepare the ontology because I could not find a formal REA ontology.

 

So how can Academia become the leader? Clearly, Academia needs to build foundational ontologies as the basis for education and that can be applied by Business. Following one of the largest universities in the U.S., I see an enormous pushback from professors over computer-based capabilities to enhance learning. My first reaction is they do not want to change, but I do not think that is a fair assessment. There must be other drivers.

 

I am a software engineer, not an academic, but many on this forum are academics. So, why is Academia not embracing the development of foundational ontologies? If the problem is the OWL language, then it needs to define another. In any case, Academia should provide the foundational ontologies needed for both education and business.

 

With a sincere hope for the future of reusing shared knowledge,

 

Tom Tinsley

--

All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/98a992de2b6b9b7ead0ca3c50643fe7a.squirrel%40webmail2.bestweb.net.

[John F. Sowa]

Tom, I sympathize with that point:

TT> The thousands of Software Engineers that use the language “OWL” follow the design principle of “Good Enough”.

I worked at IBM for30 years on both the research side and the engineering side.  I never worked in sales, but I visited customer sites to talk to customers while being accompanied by the sales guys.  I understand the trade-offs.

Engineers have a problem to solve within the limits of time, budgets, and available resources.  If the current system uses OWL, and you need to parch it up, you make do with what you have.

I also appreciate and admire some great engineering feats, and I cringe at disasters caused by engineers I won't name.  The reason why I am so disgusted with OWL is that it was obsolete in 2005.  It is a cringe-worthy obstacle that killed far better systems, and it is a dead end for future developments.

John

[Tom Tinsley]

John,

No sympathy needed here. The design principle of “Good Enough” is a way to manage and improve quality. What would have happened if John F. Kennedy had said, “We will send a man to the moon and return as soon as the technologies are perfect”?

 

This principle is the foundation of Six Sigma that applies the universal problem-solving methodology, DMAIC. In software engineering, this concept has led to the development of Agile project methodologies. Unfortunately, having been an expert witness in multiple court cases involving Agile projects, I believe we have a long way yet to go.

 

So, keeping in mind the subject of this thread, are you saying that there can be no foundation ontologies with OWL? If so, what do you recommend? Specifically, what language and tools do you recommend?

My interest is to use an ontology language to describe the Piercean category of Possibility. Then, to verify consistency in Actuality and Necessity by applying Possibility reasoning. My current prototype uses OWL and the HermiT reasoner, but a user only sees the Piercean view, so the behind the scenes could be changed.

 

Having served in IT from developer to CIO, I also, as you said, “cringe at disasters caused by engineers”. I’ve been brought in to clean up some of these disasters, I know the IT industry can do better. I believe the “better” will come from the proper application of ontology. I expect this belief is shared by all system engineers on this forum.

 

I am awed by the experts like yourself on this forum, so when you say, “it (OWL) is a dead end for future developments”, I sincerely want to know the other options.

 

Tom Tinsley

 

 

From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of John F. Sowa
Sent: Monday, May 3, 2021 12:46 AM
To: ontolo...@googlegroups.com
Subject: RE: [ontolog-forum] Foundational Ontology

 

Tom, I sympathize with that point:

--

All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/901475a5ea01b1ddde2bf72439c6b64a.squirrel%40webmail2.bestweb.net.

[Adrian Walker]

Tom,

You may like to look at the reasoner and system that's live online at www.executable-english.com .

Apologies if you have seen this before, and thanks for comments (either on  or offline).

Thanks,  -- Adrian

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/020401d74048%24dec16c00%249c444400%24%40tampabay.rr.com.

[Tom Tinsley]

Adrian thanks for the response.

 

I have seen this before. It brought back memories of COBOL (All good). I’ll take the tutorials and get back with comments.

 

Thanks again, Tom

 

From: ontolo...@googlegroups.com <ontolo...@googlegroups.com> On Behalf Of Adrian Walker
Sent: Monday, May 3, 2021 4:16 PM
To: ontolog-forum <ontolo...@googlegroups.com>
Subject: Re: [ontolog-forum] Foundational Ontology

 

Tom,

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CABbsESdGu-4a%2BJTfAGXb%2BP3HX1AM7QLw5t0eTu%3DTfvQDZ4365g%40mail.gmail.com.