Dear John, (01)
> The kinds of applications I have in mind involve connecting a database
> or knowledge base to de facto standards, such as the Amazon.com schema
> for selling, shipping, etc.
>
> When they're selling a book or camera to some individual named
> 'Matthew West', they never dream of getting involved with issues
> about 3D vs 4D ontologies. They just assume an underspecified
> entity type Person, which is associated with identifiers for other
> underspecified entity types: Address, Profile, CreditCard, etc. (02)
[MW] Interestingly, I would describe this sort of thing as over constrained.
One of the characteristics of data models is just that they are not
permissive. You can only hold the data that they allow, nothing else is
possible. You are only a person to them if you can provide an acceptable
data set about yourself. Quite possibly the same person can provide more
than one acceptable data set. (03)
And what you are talking about in this situation is a mapping between an
enterprises own database and the Amazon schema. That will need to take
account of any ontological differences, something that will often be done
without even realizing what is happening - which is why this can be an
expensive and error prone affair. During the '90s one part of Shell
estimated that 70% of the cost of new systems was in constructing such
interfaces...
>
> MW> When I have looked at that I have been forced to a different
> > conclusion, at least as far as 3D and 4D is concerned. The problem
> > is that most of these things are classes of individual, but in 3D
> > and 4D the individuals are different things, so the memberships
> > are different, so the meaning must be different. Now if the lattice
> > can accommodate varying meanings, then you may be ok. There are
> > clearly at least analogous objects.
>
> If systems based on your ontology cannot interact with the Amazon.com
> databases, then you have a very serious problem. (04)
[MW] Interaction requires a mapping, not a common upper ontology. Knowledge
of the upper ontologies of the two systems will make this much easier and
less error prone.
>
> That is why I have said that upper level ontologies, when axiomatized
> with too much detail can be *barriers* to interoperability. (05)
[MW] Who said anything about axiomatising lots of detail? Not me thank you! (06)
> I am
> all in favor of well designed upper-level ontologies. But they
> have very few axioms. The axioms for detailed reasoning belong
> where the reasoning is done: in the low-level microtheories. (07)
[MW] I have never said anything else. Getting things at the right level is
critical to long term success.
>
> If you listen to the philosophers, they will tell you that identity
> conditions are essential to ontology. But for interoperability, you
> have to *ignore* the identity conditions. When you're linking your
> DB or KB to the Amazon.com schema, you *never* want to worry about
> whether a human being is a 3D or 4D entity or whether a vase is
> identical to the lump of clay from which it is made. (08)
[MW] Yes you do, or your mapping won't work. Any differences have to be
catered for in the mapping for the interoperation to work consistently.
>
> MW> The problem is that any lower ontology must have an implicit
> > upper ontology, and it at least can't be both 3D and 4D or neutral
> > to them.
>
> No. Many, if not most, ontologies that are important for
> interoperability are neutral: The ontologies for units of measure,
> for example, are completely neutral to 3d or 4d issues. (09)
[MW] That would be very interesting if true. Can you prove it? (010)
> The
> typical databases of most businesses, including the Amazon.com
> schema, don't require detailed reasoning about identity conditions. (011)
[MW] It is not just a matter of reasoning (in fact it isn't generally a
matter of reasoning at all).
>
> I would agree that identity conditions can be very important for
> certain kinds of reasoning. For example, banks and other financial
> institutions lose billions of dollars to cases of identity theft
> and fraud. Those systems will have to extract a great deal of
> data from typical business databases, but they will have extremely
> detailed and complex identity conditions that go far beyond what
> typical business databases use. (012)
[MW] Indeed. But this is more to do with the level of confidence you have of
two records representing the same person or not, based on how many of the
attributes match or are not inconsistent with each other.
>
> MW> If they are not joined at a lower level (i.e. include some
> > same concepts) they would not become joined by an upper ontology,
> > except in the sense of being different parts.
>
> They will certainly have many of the same concepts, but the
> detailed definitions of Person, Place, Animal, Vegetable, etc.,
> at the upper levels will be extremely underspecified. (013)
[MW] What do you mean by "underspecified"? Can you give an example? (014)
If you mean that it does not say much about them, then I agree, but you how
you would say certain things at a detailed level would depend on e.g.
whether a person was 4D or 3D. (015)
> The
> detailed reasoning, including the detailed axioms and definitions
> will be done in the microtheories. (016)
[MW] Indeed. The detail should always be at the detailed level. In human
things this is often intentional, so the Amazon process is the Amazon
process because they say so. Your choice is to comply or not to use it.
>
> Just think of all the DBs and KBs that have been implemented over
> the past 40 years and are still in daily operation. They have
> been interoperating with each other very well -- the entire
> world economy depends on their interoperability. (017)
[MW] Quite. And that is because the mappings between them successfully
transform between any ontological differences. It is not that they do not
have ontological differences, or that these are unimportant. Indeed, as I
said above, the cost of these interfaces has been extremely high - largely
because of the ontological differences, and generally over constrained data
models.
>
> Those systems interoperate successfully *because* they don't
> have detailed axioms and definitions. If you insist on putting
> those details in the upper levels, you will destroy the world
> financial system (or whatever is left of it). (018)
[MW] Please stop putting words in my mouth. I have NEVER suggested putting
detailed axioms in upper levels. But there are things that do belong there,
and it is just as dangerous not to recognise those as it is to include
things that do not belong.
>
> MW> None the less there will be an implicit upper ontology. It cannot
> > be otherwise.
>
> I agree, and we have that already. I would suggest that you look
> at a few legacy systems, such as the Amazon.com schema and other
> widely used business databases. Their upper levels are almost
> free of detailed axioms and identity conditions. (019)
[MW] I have been for more than 25 years thank you, as well as designing some
of them. Some of their ontological commitments are truly frightening, and
often make it almost impossible to do business, with many systems being used
in ways that are somewhat different from the intent.
>
> And that is no accident. An underspecified upper level is a
> *prerequisite* for interoperability with a wide range of
> systems. (020)
[MW] This is simply rubbish. What you need is a correctly specified upper
level. (021)
Regards (022)
Matthew West
Information Junction
Tel: +44 560 302 3685
Mobile: +44 750 3385279
matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx
http://www.informationjunction.co.uk/
http://www.matthew-west.org.uk/ (023)
This email originates from Information Junction Ltd. Registered in England
and Wales No. 6632177.
Registered office: 2 Brookside, Meadow Way, Letchworth Garden City,
Hertfordshire, SG6 3JE. (024)
>
> John
>
>
> _________________________________________________________________
> 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
> (025)
_________________________________________________________________
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 (026)
|