IT-ontology vs knowledge base

31 views
Skip to first unread message

Alex Shkotin

unread,
Feb 26, 2026, 4:15:40 AM (6 days ago) Feb 26
to ontolog-forum

Yesterday, at our wonderful OS26-1 session, the eternal question arose again: what is an ontology? Of course, in the sense in which the term is used in IT.

It seems to me that the closest term to this is "knowledge base":

In IT, every ontology is a knowledge base.

Then, outside of IT, we rise to the much more widely used term "knowledge": what is knowledge?

So, an ontology in IT is a computer-based artifact, a kind of knowledge base.

And then we need an example of a knowledge base that computer ontology experts don't recognize as an ontology. 

`Please please me`

If every knowledge base is an ontology, then these two terms are merely synonyms.


In passing, I'll note that a knowledge base has the property of being composed of units or elements of knowledge, such as sentences, words, etc.

But LLM and, in general, ANNs are knowledge processors: it's difficult to find a unit of knowledge by looking into them.


A few thoughts after a wonderful session.


Alex


John Antill

unread,
Feb 26, 2026, 9:24:59 AM (5 days ago) Feb 26
to ontolo...@googlegroups.com
Alex — great question, and I think the confusion is productive rather than semantic.
You’re right that in IT usage an ontology is a computer-based artifact. But it is not simply synonymous with “knowledge base.” The relationship is narrower and more technical.

1. Necessary distinction: knowledge base vs ontology
A knowledge base (KB) is any structured or semi-structured repository of information designed for retrieval or inference.
An ontology, in the IT sense (especially in semantic technologies), is a formal, explicit specification of a conceptualization (Gruber, 1993). That definition carries three important constraints:
Formal – machine-interpretable semantics (logic-based representation, e.g., OWL, description logic).
Explicit – clearly defined classes, properties, constraints.
Conceptualization – a model of the domain’s entities and relationships.
So:
Every ontology can function as (or support) a knowledge base.
Not every knowledge base qualifies as an ontology.
An ontology is not just content — it is a semantic model.

2. Example: a knowledge base that is not an ontology
Here are examples most ontology engineers would not classify as ontologies:
A traditional FAQ database (e.g., “How do I reset my password?” → answer text).
A SharePoint document repository.
A relational database of customers and transactions without formal semantics.
A vector database used for RAG retrieval.

A corporate KMS built on keyword search without formal class/property modeling.
These are knowledge bases because they store retrievable knowledge artifacts.
But they are not ontologies unless:
They define explicit classes and relationships,
Use formal semantics,
Support logical inference grounded in a conceptual model.

For instance:
A SQL table with columns CustomerID, Address, Balance is data.
An ontology would formally define Customer as a class, hasAddress as an object property, cardinality constraints, inheritance relationships, etc.
That semantic commitment is the differentiator.

3. Where your intuition is correct
You are absolutely right that ontologies are a subtype of knowledge base.
Formally:
Ontology ⊂ Knowledge Base
But the reverse inclusion does not hold.
If every KB were an ontology, then OWL reasoners would work on arbitrary document repositories — which they clearly do not.

4. On “units of knowledge”
You note that knowledge bases consist of identifiable units (sentences, words, entries). That’s correct in most symbolic systems.
Ontologies go further: their units are not sentences but axioms — logical statements about classes and relationships.
For example:
Dog ⊑ Mammal
has Mother ⊑ has Parent
Person ⊓ Male ⊑ Man
These are not just stored facts; they are formal constraints enabling inference.

5. LLMs and ANNs complicate the picture
Your observation about LLMs is important.
LLMs are not ontologies.
They are not knowledge bases in the classical symbolic sense.
They are statistical function approximators over high-dimensional parameter spaces.

Knowledge in an LLM is:
Distributed,
Sub-symbolic,
Not discretely addressable,
Not logically explicit.
You cannot “retrieve the axiom” that Dog ⊑ Mammal from inside the weights. Instead, the model approximates patterns learned from text.
This makes LLMs knowledge processors — but not structured knowledge bases.
They blur the boundary between:
Representation,
Storage,
Processing.

6. A cleaner taxonomy
To reduce confusion, I would suggest this hierarchy:
Data – raw symbols.
Information – structured data.
Knowledge base – organized, retrievable information.
Ontology – formal semantic model enabling logical inference.
LLM – probabilistic model approximating language patterns (knowledge-like behavior without explicit representation).

7. The core issue
The problem is not synonymy — it is abstraction level.
“Knowledge base” is functional.
“Ontology” is structural and formal.

One answers:
Where is the knowledge stored?
The other answers:
How is the domain conceptualized and formally constrained?
They overlap, but they are not identical.

John Antill
MS KM, MCKM, CKS IA & KT, KCS
MS AI Student at Purdue


--
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 visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROSCZB5-eif06erOp4bqAcd%3DB70ci9tNGFz58RUsvwo8zQ%40mail.gmail.com.

Jake Waas

unread,
Feb 26, 2026, 11:31:13 PM (5 days ago) Feb 26
to ontolo...@googlegroups.com
Anecdotally, about the upper ontology I use - my fork of Erik Sandewall's Leonardo System - a knowledgebase contains entityfiles of entities. Zero or more entityfiles defining entities of type Type for a knowledgebase are construed as ontologies. There is a basal Ontology knowledgebase which contains ontology entityfiles that are always present. (Viz https://codeberg.org/screwlisp/pawn-75/src/branch/main/Pawn-75/remus/Ontology https://codeberg.org/screwlisp/pawn-75/src/branch/main/Pawn-75/utilus/Dtc-lib/dtc-onto.leo ). I imagine someone in the ontolog forum knows more about the system I use than me.

Alex Shkotin

unread,
Feb 27, 2026, 5:50:15 AM (5 days ago) Feb 27
to ontolo...@googlegroups.com
John,

It's good that you cite Gruber and use "semantic" terminology, which is typical for ontologists.

But you don't provide any references to the definition of a knowledge base. Where did you get your cute definition?

ISO has long had a short and concise definition of a knowledge base [1]:

"a database that contains inference rules and information about human experience and expertise in a domain."

Of course, that was back in the days of Prolog. And of course, OWL2's remarkable achievement is the development of knowledge processing algorithms.


So you claim that an ontology is a knowledge base, but not every knowledge base is an ontology.

And here are your criteria with my comments:

"But they are not ontologies unless:

They define explicit classes and relationships, --which are rules. 

Use formal semantics, --which is a property of the processor that processes the knowledge base.

Support logical inference grounded in a conceptual model. --which is a property of the processor that processes the knowledge base.

"

So we need to have classes and relationships that are "grounded in a conceptual model." This is a rather fragmented understanding of what experts know about the subject area.


For now, it's probably safe to say that some structured knowledge storage may not be suitable for use as input to a universal Hermitt-class knowledge processing algorithm. But if we look at the hets.eu project, we can conclude that there may be a universal converter of knowledge bases into structures accessible to certain formal knowledge processing algorithms.


It seems like Prolog knowledge bases aren't an ontology for you?

So, you're saying that if knowledge bases aren't directly accessible to Hermitt, they aren't ontology?

If we have certain knowledge processing algorithms on one hand, and knowledge bases on the other, then the question naturally arises as to what format our knowledge bases should be used as input to the algorithms.


I propose, based on the ISO definition of knowledge bases, to reiterate the criteria for being an ontology.


It's great to come up with examples, of course, but it would be better to refer to a real knowledge base that you don't consider an ontology.


Once again, if my knowledge is presented in a format not directly accessible to Hermitt, that's no reason to claim it's not an ontology.

The ontological nature of knowledge shouldn't depend on the format of presentation.


Alex


[1] https://www.iso.org/obp/ui/en/#iso:std:iso-iec:2382:ed-1:v2:en

2123771 knowledge

<artificial intelligence> collection of facts, events, beliefs, and rules, organized for systematic use

Note 1 to entry: knowledge : term and definition standardized by ISO/IEC [ISO/IEC 2382-28:1995].

Note 2 to entry: 28.01.03 (2382)

[SOURCE:ISO-IEC-2382-28 * 1995 * * * ]

2121399 knowledge base K-base

database that contains inference rules and information about human experience and expertise in a domain

Note 1 to entry: In self-improving systems, the knowledge base additionally contains information resulting from the solution of previously encountered problems.

Note 2 to entry: knowledge base; K-base: term, abbreviation and definition standardized by ISO/IEC [ISO/IEC 2382-1:1993].

Note 3 to entry: 01.06.18 (2382)

[SOURCE:ISO-IEC-2382-1 * 1993 * * * ]

2123867 knowledge base K-base

database that contains inference rules and information about human experience and expertise in a domain

Note 1 to entry: In self-improving systems, the knowledge base additionally contains information resulting from the solution of previously encountered problems.

Note 2 to entry: A new abbreviation is added to that mentioned in ISO/IEC 2382-1:1993.

Note 3 to entry: knowledge base; K-base; KB: terms, abbreviation and definition standardized by ISO/IEC [ISO/IEC 2382-28:1995].

Note 4 to entry: 28.04.06 (2382)

[SOURCE:ISO-IEC-2382-28 * 1995 * * * ]

2123801 knowledge tree

hierarchical semantic network represented by a tree-like directed graph

Note 1 to entry: knowledge tree: term and definition standardized by ISO/IEC [ISO/IEC 2382-28:1995].

Note 2 to entry: 28.02.11 (2382)

[SOURCE:ISO-IEC-2382-28 * 1995 * * * ]

2123773 knowledge-based system

information processing system that provides for solving problems in a particular domain or application area by drawing inferences from a knowledge base

Note 1 to entry: The term "knowledge-based system" is sometimes used synonymously with "expert system", which is usually restricted to expert knowledge.

Note 2 to entry: Some knowledge-based systems have learning capabilities.

Note 3 to entry: knowledge-based system; KBS: term, abbreviation and definition standardized by ISO/IEC [ISO/IEC 2382-28:1995].

Note 4 to entry: 28.01.05 (2382)

[SOURCE:ISO-IEC-2382-28 * 1995 * * * ]




чт, 26 февр. 2026 г. в 17:24, John Antill <djant...@gmail.com>:

Alex Shkotin

unread,
Feb 27, 2026, 6:02:27 AM (5 days ago) Feb 27
to ontolo...@googlegroups.com

Jake,


For me "Zero or more entityfiles defining entities of type Type for a knowledgebase are construed as ontologies." means "Any".

And I hope not every Lisp program is a knowledge base?


Alex



пт, 27 февр. 2026 г. в 07:31, Jake Waas <ja...@waas.nz>:

John Antill

unread,
Feb 27, 2026, 8:49:57 AM (4 days ago) Feb 27
to ontolo...@googlegroups.com

Alex 

Thank you. Let me answer this section directly and narrowly, since you’re asking for references and for clarification of the criteria.

1. On the definition of “knowledge base”

You are absolutely correct to cite ISO/IEC 2382. That is a legitimate and authoritative definition:

“A database that contains inference rules and information about human experience and expertise in a domain.”

My usage is not meant to replace ISO, but to operate at a different abstraction layer.

In addition to ISO 2382 (1993/1995), the broader AI literature defines knowledge bases functionally, e.g.:

  • Russell & Norvig, Artificial Intelligence: A Modern Approach – A knowledge base is a set of sentences in a formal language.

  • Brachman & Levesque, Knowledge Representation and Reasoning – A KB consists of a representation language plus a set of assertions in that language.

  • Classic expert systems literature (Newell, Feigenbaum) – A KB is the knowledge component separated from the inference engine.

All of these definitions are intentionally broad.

So I am not proposing a “cute” alternative definition. I am operating within the mainstream KR definition: a knowledge base is any structured set of knowledge artifacts that supports inference.

ISO’s definition is fully compatible with that.

  • Russell & Norvig, Artificial Intelligence: A Modern Approach (4th ed., 2021) define a knowledge base as a set of sentences in a formal language to which inference procedures are applied.

  • Brachman & Levesque, Knowledge Representation and Reasoning (2004) describe a KB as a representation language plus a set of assertions expressed in that language.

  • Newell (1982), “The Knowledge Level,” Artificial Intelligence, distinguishes the knowledge level from the symbol level and inference mechanism.

  • Feigenbaum (1977), “The Art of Artificial Intelligence,” IJCAI, separates the knowledge base from the inference engine in expert systems.

2. “Ontology is a knowledge base, but not every knowledge base is an ontology”

This is not a rhetorical move. It is a distinction established in the ontology engineering community.

The differentiator is not:

  • Whether inference occurs

  • Whether rules exist

  • Whether a reasoner like Hermit can process it

The differentiator is whether the artifact is a formal specification of a conceptualization, not merely a rule-based problem-solving system.

This distinction appears in:

  • Gruber (1993)

  • Guarino (1998)

  • Studer, Benjamins & Fensel (1998)

  • The W3C OWL specifications

The key issue is ontological commitment, not computational accessibility.

  • Gruber (1993), “A Translation Approach to Portable Ontology Specifications,” Knowledge Acquisition.

  • Studer, Benjamins & Fensel (1998), “Knowledge Engineering: Principles and Methods,” Data & Knowledge Engineering.

  • Guarino (1998), “Formal Ontology in Information Systems,” FOIS 1998.

3. “Explicit classes and relationships — which are rules”

Classes and relationships are not “rules.”

They are vocabulary primitives.

A rule in Prolog:

grandparent(X,Z) :- parent(X,Y), parent(Y,Z).

A class axiom in OWL:

Grandparent ≡ Person ⊓ ∃hasChild.(∃hasChild.Person)

The first is procedural Horn-clause inference.
The second is a logical definition under description logic semantics.

The difference is not superficial. One defines a predicate extensionally via resolution; the other defines a concept intensionally via model-theoretic constraints.

That is why ontology engineers distinguish:

  • Vocabulary (classes, properties)

  • Axioms (subsumption, equivalence, restrictions)

  • Rules (SWRL, production rules, etc.)

They are not the same category.

  • Baader et al., The Description Logic Handbook (2003), distinguish TBox (terminological axioms: classes, subsumption, properties) from ABox (assertions about individuals), and separate these from rule systems.

  • The W3C OWL 2 Structural Specification (2012) clearly distinguishes class axioms and property axioms from rule extensions such as SWRL.

4. “Formal semantics is a property of the processor”

This is technically incorrect.

Formal semantics belongs to the language specification, not the processor.

OWL 2 has a model-theoretic semantics defined independently of:

  • Hermit

  • Pellet

  • Fact++

  • Any specific implementation

A reasoner is an algorithmic realization of consequence relations defined by the semantics.

If every OWL reasoner disappeared tomorrow, OWL would still have a formal semantics.

So ontology status does not depend on Hermit accessibility.

It depends on whether the representation language has well-defined declarative semantics.

  • The OWL 2 Direct Semantics (W3C Recommendation, 2012) provides a model-theoretic semantics independent of any specific reasoner.

  • Description logic semantics are formally defined in Baader et al. (2003).

  • The model-theoretic tradition originates with Tarski (1956), Logic, Semantics, Metamathematics.

5. Prolog knowledge bases

Some Prolog KBs can function as ontologies.

But many do not.

Why? Because they often:

  • Mix domain knowledge with control strategy

  • Use procedural constructs (cuts, ordering effects)

  • Lack explicit ontological commitments (identity criteria, typing discipline, subsumption structure)

A clean, purely declarative first-order theory encoded in Prolog syntax could absolutely count as an ontology.

A heuristic expert system implemented in Prolog typically would not.

So the issue is not “Prolog vs OWL.”
It is declarative domain theory vs procedural problem-solving system.

  • Mix domain knowledge with control strategy (Lloyd 1987).

  • Depend on operational features such as ordering and cut (Kowalski 1979).

  • Lack explicit ontological commitments in the sense articulated by Guarino (1998).

6. On Hets and universal converters

Hets translates between logical systems that already have formal semantics.

It does not convert arbitrary knowledge storage into ontology.

You cannot feed a SharePoint document repository into Hets and obtain a formal ontology automatically.

Translation requires that the source already be expressed in a logic.

So again, the distinction hinges on logical formalization, not file format.

  Hets (Mossakowski, Maeder & Lüttich, 2007, TACAS) is grounded in institution theory (Goguen & Burstall, 1992  

7. A real knowledge base that is not an ontology

You requested a concrete case.

Consider a commercial fraud detection knowledge base in a financial institution:

  • Thousands of production rules

  • Pattern thresholds

  • Heuristic scoring rules

  • Operational exceptions

  • Stored procedures tied to execution workflows

It satisfies ISO:

  • Contains inference rules

  • Encodes domain expertise

  • Supports automated decision-making

But it lacks:

  • Explicit conceptual modeling of domain entities

  • Identity criteria

  • Subsumption hierarchies grounded in ontology design

  • Declarative model-theoretic semantics

It is a knowledge base.
It is not an ontology.

Similarly:

  • A vector search index for RAG

  • A CRM relational schema

  • A rules engine for pricing

All qualify under ISO’s KB definition.
None are necessarily ontologies.

8. “The ontological nature of knowledge shouldn’t depend on format”

Philosophically, yes.

But in IT, we are not talking about knowledge-in-itself.

We are talking about engineered artifacts.

An ontology in IT is a formal artifact with explicit semantics.

A conceptual model written in English may describe an ontology.
It is not an ontology artifact until formalized.

This is the same distinction between:

  • Mathematical truth

  • A formal proof in a specified logical system

Ontology engineering operates in the latter space.

  • Noy & McGuinness (2001), Ontology Development 101, emphasize formal class and property specification.

  • Smith (2003), “Ontology,” in Blackwell Guide to the Philosophy of Computing and Information, distinguishes philosophical ontology from computational ontology artifacts.

9. Reiterating criteria clearly

Based on ISO and ontology engineering literature, we can state:

A knowledge base becomes an ontology when:

  1. Its vocabulary explicitly defines domain classes and relations (Gruber 1993; Studer et al. 1998).

  2. Its assertions are intended as domain axioms expressing ontological commitment (Guarino 1998).

  3. The representation language has formal declarative semantics independent of execution strategy (Tarski 1956; OWL 2 Direct Semantics 2012).

  4. Inference derives from logical consequence under that semantics (Baader et al. 2003).

That does not require Hermit.
It does not exclude Prolog.
It does require logical formalization.

The disagreement here is not about ISO being wrong.
It is about ISO defining a superset category.

Ontology, in the semantic-technology sense, is a stricter subclass of knowledge base.

That subclass distinction is structural, not tool-dependent.

John Antill

MS KM, MCKM, CKS IA & KT, KCS
MS AI Student at Purdue

Paola Di Maio

unread,
Feb 28, 2026, 11:44:16 PM (3 days ago) Feb 28
to ontolo...@googlegroups.com
Alex and everyone

asking what is ontology and what is knowledge are  entry level *esp for learned communities
who have been answering the question for all possible angles for half a century

I wonder if you have ever thought of creating a primer  and an evaluation sheet  to guide for those who are new
so that we do not have to keep repeating the same things over and over

Without foundational knowledge, it is difficult to advance discourse

Other questions that may be of interest
- what is new in ontology?  that has not been said/done in the past?
What is knowledge in 2026, in the light of scientific and technological advances
of the last decade?
how does our understanding of the world, informed by ontology and knowledge actually impact our lives *as scientists and as citizens
in terms of belief, awareness of what is going on, ability to continue to learn, adapt and evolve etc

PDM


Alex Shkotin

unread,
Mar 2, 2026, 6:12:40 AM (yesterday) Mar 2
to ontolo...@googlegroups.com

John


Great reply. We're discussing a rather technical problem. We need two definitions: one for KB and one for IT-ontology.

I've posted the definitions you found, your own, and a few more in the IT-knowledge base DEFINITIONs. It's open for comments. And perhaps, if Ken and other trustees are interested, we could move it to our website.

We need to choose a base definition from which to derive the IT-ontology definition. For which we might need to create another document. I'll work on that in a few days.


Regarding your two definitions JA001, JA002. (see IT-knowledge base DEFINITIONs)

JA001!

A knowledge base is any structured or semi-structured repository of information designed for retrieval or inference.

URL PID:JA

JA002!

A knowledge base is any structured set of knowledge artifacts that supports inference.

URL PID:JA

In your first definition, "or" implies that DB is also suitable.

Moreover, for me, these are two different definitions, and one needs a sufficiently developed theory to prove that they are equivalent. It would be good to write it out: primary terms, axioms, definitions.


Regarding point 7, "A real knowledge base that is not an ontology." You misunderstood me. I need a real-life example of someone who created a knowledge base—that's what they call it. And you look at it and say, "Yes, it's a knowledge base, but it's not an IT ontology."


All details about IT-ontology definitions later.


Just one time my just initial point: if we have knowledge this is our problem to get tools for reasoning with it. Well it may be incomplete or even contradictory. Is the IT-ontology not contradictory and complete knowledge about the domain?

For me it's a theory or model for theory.


Alex



пт, 27 февр. 2026 г. в 16:49, John Antill <djant...@gmail.com>:

John Antill

unread,
Mar 2, 2026, 8:38:30 AM (yesterday) Mar 2
to ontolo...@googlegroups.com

Alex,

Thank you — this helps clarify exactly where the real disagreement lies.

You are correct: JA001 and JA002 are not equivalent definitions. They represent two different abstraction levels, and treating them as interchangeable would be a mistake.

Let me state my position cleanly.

1. On JA001 vs JA002

JA001 (“retrieval or inference”) is intentionally broad. It captures common ITSM and documentation practice, where “knowledge base” often means a structured repository for reuse, even if no formal inference occurs.

JA002 (“supports inference”) is narrower. It aligns with:

  • ISO/IEC 2382 (inference rules + expertise),

  • Russell & Norvig (KB = set of sentences in a formal language),

  • Classical KR tradition.

These are not equivalent. Formally:

JA002 ⊂ JA001

Every inference-supporting KB satisfies JA001.
Not every JA001 KB satisfies JA002.

For purposes of deriving an IT-ontology definition, I agree that we must choose a base definition. I propose that we adopt JA002 (inference-essential) as the base, precisely to avoid the “database = KB” collapse.

If we choose JA001, the derivation to ontology becomes structurally unstable.

2. On “a real KB that is not an IT-ontology”

Thank you for clarifying what you want.

Here is a concrete example:

ServiceNow Knowledge Base (or Atlassian Confluence Knowledge Base).

  • It is explicitly labeled “knowledge base.”

  • It is structured.

  • It supports search, categorization, workflow approval, versioning.

  • It may even contain decision trees or procedural logic.

It is unquestionably a knowledge base in industry usage.

However, it is not necessarily an IT-ontology because:

  • It does not require formal semantics.

  • It does not require explicit class/property axiomatization.

  • It does not define ontological commitment in the Gruber/Guarino sense.

  • Its structure is organizational, not model-theoretic.

So we can say:

Yes — it is a knowledge base.
No — it is not (by default) an IT-ontology.

That is the type of real-world discriminator you requested.

3. On Contradiction and Completeness

You raise an important philosophical question:

Is the IT-ontology complete and non-contradictory knowledge about the domain?

My answer is: no.

An IT-ontology is not complete knowledge. It is not guaranteed consistent in practice. It is not a total theory of a domain.

It is a formalized conceptual model intended to:

  • Constrain interpretation,

  • Enable logical consequence,

  • Make commitments explicit.

It may be incomplete.
It may contain modeling errors.
It may require revision.

So I would not define IT-ontology as “complete knowledge.”

I would define it as:

A formally specified conceptualization expressed in a representation language with explicit semantics.

It is closer to a theory schema than to a complete theory.

Your formulation — “a theory or model for theory” — is actually very close to how ontology engineers think about it.

4. On Primary Terms and Axioms

I agree with you that we should not leave this at the level of informal prose.

If we proceed rigorously, we could structure it as:

Primary terms:

  • Knowledge artifact

  • Representation

  • Inference

  • Conceptualization

Definition (KB):
A structured set of knowledge artifacts in a representation supporting inferential use.

Definition (IT-ontology):
A knowledge base whose artifacts constitute explicit axioms about domain entities and relations, expressed in a language with formal declarative semantics.

Then the subclass relation becomes demonstrable rather than rhetorical.

I am happy to work through that derivation formally.

5. On Format vs Ontological Nature

You are correct that the ontological nature of knowledge should not depend on format in a philosophical sense.

But in IT we are not classifying knowledge-in-itself. We are classifying engineered artifacts.

An IT-ontology is not “knowledge as such.”
It is a formally encoded artifact with defined semantics.

The distinction is epistemic vs engineered

Summary

  • JA001 and JA002 are not equivalent.

  • For derivation purposes, I propose adopting JA002.

  • A ServiceNow or Confluence KB is a real KB that is not an IT-ontology.

  • IT-ontology is not complete knowledge; it is a formalized conceptual model.

  • I agree that the next step should be a structured derivation document.

If you draft the next document on IT-ontology, I will happily work through the formal derivation with you.

John Antill

MS KM, MCKM, CKS IA & KT, KCS
MS AI Student at Purdue

Alex Shkotin

unread,
7:50 AM (12 hours ago) 7:50 AM
to ontolo...@googlegroups.com

PDM,


For now, it's suggested that we define ontology in IT based on the assumption that it's a special type of knowledge base. And then we have to start with what knowledge base is. Join the IT-knowledge base DEFINITIONs.

We will gradually assemble a collection of definitions of knowledge base and IT ontologies.


To your first question, my answer is no. There are already plenty of them.

Well, your other three questions are certainly interesting.


And your last topic is a matter for philosophers, not IT.


Alex



вс, 1 мар. 2026 г. в 07:44, Paola Di Maio <paola....@gmail.com>:

Alex Shkotin

unread,
10:55 AM (9 hours ago) 10:55 AM
to ontolo...@googlegroups.com

John,


I propose focusing primarily on collection. Each collection entry should only include ontological attributes (so-called references to reality):

URL and/or bref: a reference to a real-world location where you can go (preferably with one click) to verify that the definition is cited correctly, i.e., it actually exists.

AID: the definition's author identifier. You can contact the author for clarification.


You've provided two definitions, and I think two more. I've formatted the first two correctly: there's a URL, so anyone can go and see them, and there's a link to the author, so anyone can contact you.


Let's develop all the other topics outside the collection in the Discussions section. There will also be sections for different authors.

Perhaps some consolidated point of view will emerge.

I propose working on the document structure and the procedure for working with it at the IT-knowledge base DEFINITIONs.

And then we'll create a similar one for IT-ontology. You've already provided several references.

I hope we can find an acceptable format and procedure.

Take a look at our work from about 2.5 years ago (t)"service". the word meanings, term definitions; theories involved(-:PUBLIC:-) 📒


Alex



пн, 2 мар. 2026 г. в 16:38, John Antill <djant...@gmail.com>:
Reply all
Reply to author
Forward
0 new messages