ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] master data vs. ontologies

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Jack Hodges <jhodgesatmb@xxxxxxxxx>
Date: Thu, 12 Feb 2015 08:15:54 -0800
Message-id: <CALqBbwsGTuQ5tmO74ZbWoZk=V96-ZzV30sPb83rreEA5nvriGQ@xxxxxxxxxxxxxx>
MDM is generally considered in the context of database systems, and intranets within enterprise networks, but is certainly about managing models just as ontologies are. At this level the differences are mostly about the differences between database system technologies and semantic system technologies. These differences are well known.

On another level ontologies, or should I say the greater ontology community, focuses more on the scope of use. MDM systems are still localized, and thus the models are still vetted within the context of an enterprise and its use cases. The focus/goals of the semantic community is to develop models that will be vetted on the broadest scale possible, starting perhaps in localized communities but branching outward when and where possible. So the goals are similar, and the kinds of people working the problems have a similar mindset and approach, but the tools and languages are different, and the focus (sometimes because of those tools and languages) is different.

Logic is one way to build models but it is only one way so I do not think that bringing it into the discussion is on point.

On Thu, Feb 12, 2015 at 6:14 AM, John F Sowa <sowa@xxxxxxxxxxx> wrote:
Erick, Alex, Ravi, and Kingsley,

EA
> I need some help to better define the line (sometimes apparently
> grey) between master data and ontologies.

It's grey because those terms are not mutually exclusive.  If you
clearly specify the definitions of your data and relate them to
the subject matter, you would have an ontology.

The critical distinction is not between master data and ontology,
but between logic and ontology.

EA
> I guess most of us are familiar with Gruber's one: a formal
> specification of a shared conceptualization.

That definition is widely quoted, but Barry Smith has said that
it's a terrible definition -- and I agree with him:

  1. It depends on four terms that are as difficult or even more
     difficult to define than 'ontology':  'formal', 'specification',
     'shared', and 'conceptualization'.

  2. The word 'shared' is irrelevant.  Anyone can define an ontology
     for a particular purpose without sharing it.  And it's still
     an ontology.

  3. The word 'conceptualization' depends on the word 'concept',
     which most people have heard.  But if you ask anyone who is
     not a philosopher or psychologist to define it, you won't get
     anything useful.  If you ask a philosopher or psychologist,
     you'll get a dozen definitions, most of which they don't like.

The critical distinction is not between master data and ontology, but
between logic and ontology.  I recommend the following definitions:

  1. Logic is the study of conditions that distinguish true
     statements from false statements.  People use informal logic
     every day -- whenever they agree or disagree with anyone else.

  2. Ontology is the study of existence.  As Quine said, the basic
     question of ontology is "What exists?"  Answer:  "Everything."
     The hard part is to analyze, catalog, and define everything.

  3. An ontology is a catalog of the kinds of things that exist
     in some domain with sufficiently precise definitions to
     distinguish and relate the various kinds.

  4. Logic is used to reason with those definitions in order to
     derive their implications.

AS
> we need programmer to work with data and mathematician to work
> with ontology.  A big difference.

There is no difference -- because anybody who has ever written a
program that runs correctly on a digital computer is a mathematician.

It's true that logic has become very mathematical in the past 150 years.
But Aristotle's subset of logic and ontology doesn't require anything
more than Venn diagrams:  http://www.jfsowa.com/talks/aristo.pdf

RS
> Ontology also deals with relationships but not as RDBMS only, these
> can be rich predicates and hence should enhance MDM if Ontologies
> are used these would in many cases would show more light on entities
> and also types of relations and also can relate to affinities.

I agree.  But any well-defined specifications for any database
(relational or network) can be mapped to an ontology.  Aristotle's
logic (Venn diagrams) is sufficient for many purposes.  In fact,
most OWL ontologies don't use much, if anything beyond Aristotle.

KI
> The problem is that MDM is a marketing moniker, just like "Big Data",
> SOA, and the like. All of these monikers are deliberately generic
> and borderline meaningless (by design). An ontology (in my eyes) is
> the antithesis of fluffy marketing buzzwords since they force you
> to look at the actual characteristics of an entity and define said
> characteristics meaningfully.

I agree.  One reason why I like to start with Aristotle's subset is
that it gets rid of the fluff.  It shows how to specify an ontology
with a simple subset of English:  just four sentence patterns.

John



--
Jack

_________________________________________________________________
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    (01)

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