ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Financial Industry Business Ontology (FIBO)

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: William Frank <williamf.frank@xxxxxxxxx>
Date: Sun, 14 Dec 2014 17:59:21 -0500
Message-id: <CALuUwtB-uKfqP-zW=+ZzN+Fpsx7M3zpXBCzy09LSTW3m=L_SJA@xxxxxxxxxxxxxx>


On Sun, Dec 14, 2014 at 5:13 PM, Matthew West <dr.matthew.west@xxxxxxxxx> wrote:

Dear William,

Well it is probably just as well then that the definition of “thing” does not matter much at all for FIBO. For sure the text (probably arrived at by a committee) will have no weight at all. Only the formal constraints will actually matter.


Ah, thank you Mathew.  A return to sanity from my flame.   You are of course completely right.  This, indeed, is my point about UML. Fortunately, nobody CARES about its absurd separation of things into incontroverable levels, each with its own 'language', even using different names for categories of things (messages, classes, dogs) and their instances (signals, objects, (but, gee, what are we going to call dog instances??)).  

We ULM lovers just need those great and easy to draw with greawt tools hierarchical state transition diagrams, sequence diagrams, activity diagrams, that we call all foot to the same class diagrams, and the fact that the mothers of uml don't like to let you put objects and classes is the same diagram (because Fido and Dog are in different languages!) is just a minor annoyance.

Same here,  the benefit of FIBO is not going to be a confused rehashing of the ULM meta model, but what is actually the financial services **content.**  

However, with a richer, yet far more simple, more rational 'upper' ontology, based as John suggests, on logic, especially, I think, on Common Logic, so that we would be able to bounce around those different 'languages,' talking about properties, colors, red, and the red of my sweater as easily as most 5 year olds can do, (at least before they are taught UML or OWL, for that matter), their contribution would be more directly usable in different languages, data modelling paradigms, and they would be able to express a richer set of logically necessary relations (for example, a trading TE, event with buyer B and seller S and traded asset A results in a trade settlement commitment, which is a relation representing a commitment with B playing the role of buyer and S as seller and A as exchanged asset, which in term results in a settlement EVENT ES. ....   And, a repurchase agreement is a type of trade, with the added wrinkle that repurchase agreements are based on standing agreements which are in fact types of contracts, and a trade settlement commitment is a type of contract.  Meaning that all of a sudden we are talking about 'a' repurchase agreement, as if it were a type, and trade itself is being treated as a type.   And, the transformation from standing repurchase contract to repurhcase execution to settlement commitment to settlment event is in fact a key part of the very meaning of these concepts.

I think this is at least as awkward to say in OWL as it is in UML.  I think OWL is just old wine in a new Klein bottle. without all the other kinds of pretty pictures I need.  On top of which, at least UML gives strong support for the use of part-whole and attributive relations, which are fundamental to financial services.  (For example, a "sushi bond' or a zero coupon bond are NOT effectively modelled as TYPEs of bonds.  They are more effectively understood as bonds, in which one of their parts has a relation to another thing with another attribute.  (A U.S. sushi bond being a bond issued by the U.S. agency of a Japanese company, denominated in dollars, and a zero coupon bond being a bond in which the periodic payment terms of the bond is 'no periodic payments' -- i.e., the classifier depends on a characteristic of one of the constituents of the bond, not the bond itself).  Now, given that the OWL tutorial tells us that Chardonnay is a type of wine, instead of a type of grape that can be used as a classifier of wines, and that we should draw a box called Chardonnay that means wine type, as well as another box that says White wine, and another that says Burgundy, the OWL tuturial would have you treat a zero coupon Sushi bond and multiply inheriting from Sushi bonds and zero coupon bond.   With all the inventiveness of financial engineers, this gets impossibly ugly very fast. At least new Wine 'types' need 10 years for the vines to mature.     

OTOH, I guess that one can't expect to be making advances in an application of ontology, like FIBO, while at the same time advancing ontology itself.   Yet, this just shows how poor a foundation weakly thought out upper ontologies like those of UML and OWL give for necessary and valuable efforts like FIBO.


What I would be much more concerned about is the definition of something specific to the financial industry that will actually be acted on, so something like Allotment Right. Even then, of course, a textual definition does no more than give guidance as to the intended interpretation. I’m sure that there will be multiple models possible. I expect that almost all the terms defined in FIBO are primitive.

The main purpose of this ontology will be that different organizations will use the terms with the same meaning, the value will come not from reasoning but from reduced time and cost from data not having to be translated between the different and independent classification schemes used by different organizations.

Good text definitions at the operational level will be critical for this since it will be people making the mappings.

 

Regards

 

Matthew West                           

Information  Junction

Mobile: +44 750 3385279

Skype: dr.matthew.west

matthew.west@xxxxxxxxxxxxxxxxxxxxxxxxx

http://www.informationjunction.co.uk/

https://www.matthew-west.org.uk/

This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.

Registered office: 8 Ennismore Close, Letchworth Garden City, Hertfordshire, SG6 2SU.

 

 

 

From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of William Frank
Sent: 14 December 2014 21:48
To: [ontolog-forum]
Subject: Re: [ontolog-forum] Financial Industry Business Ontology (FIBO)

 

And, to follow on John's remark, of course the authors of this ontology standard never studied, or at least never learned, any axiomatic mathematics or theory of definition, whereby all defined terms need to be rooted in primitive, undefined, terms, in a partial order with the undefinable primitive terms as the roots. 


The 'meaning' of primitive terms can't be provided in a definition, for obvious reasons.  They are only given p 'meaning' in terms of their use with respect to each other, as specified by the axioms.

As a result, they do not know the difference between:

a definition, whereby a term to be defined, a definiendum, is defined using an _expression_ (the definiens) with which the denifiniedum can be replaced.

 

an explication, whereby a primitive term is loosely explained in terms of some examples of its use, some of its likely interpretations in other languages, just to give a little more human weight to the intent of the primitive terms.

 

Clearly, they intended 'thing' to be a primitive term, which means, that ***in their context***, and most of the likely useful ones I can imagine, 'thing' will be undefinable.   But it is not undefinable in and of itself.  It is only in the context of a given axiomatic system that a term used in the system is primitive or defined.  For example, 'or' and 'plus' are primitive in some axiomatizations of Boolean algebra and arithmetic, and not in others.  (I have been told too often that 'or' is 'really' a bunch of nands, whatever really is supposed to mean here.)

 

I would imagine, though, that they intended to be using other primitive terms, in addition to 'thing'.   Since 'things' are what they want to show in boxes.  I imagine there are some arrows flying around somewhere, too.   

Sadly, after all this time, a new 'standard' that does not regard what you put in a box or what you put on an arrow as a matter of a modelling decision, or a matter of the syntax you use to name the concept (gerunds, lamda operators, etc., anyone?) that is, how you want to CAST the concept, depending on the focus of the domain, but instead still imagines that these syntactic choices and accidents are something **essential** about the very concept.

So, according to standards like this, we can't REcast anything.   Marriage can't be cast as either an event or as a relation, so any relationship between marriage and divorce ceremonies and marriage relationships is purely coincidental.  Moreover, whichever of those two marriage 'really' is, I guess a marriage could certainly never be a 'thing.'   (Before UML, I doubt anyone had invented an artificial language in which it was actually *impossible* for human thought to occur.  Fortunately, people who actually use UML, like me, use the pretty pictures, and try their best to forget about its ontological absurdity, in the unlikely event that they ever were exposed to that dark glass in the first place.) 

Wm

 

 

On Sun, Dec 14, 2014 at 3:56 PM, William Frank <williamf.frank@xxxxxxxxx> wrote:

I wish words failed me, too.    But my outrage needs some words to go with it.

I am **most** concerned not with how egregiously AWFUL this is, but what it means about the engineering culture in which this could occur. 

There are two disfunctional characteristics that seem to me to be at play here. 

1.Standardize everything first, then use it later, if at all.   No more try before you buy.  Back in pre-history, things like SQL and C were introduced, adopted, then finally turned into standards.  Now, every time somebody has a new idea (or even, as in this case, seems to wish they were having an idea), the FIRST thing they do is to define a 'standard'.

2. Talk (or code) first, learn (or reuse) never.   This subject, as we here all love to yammer on about, has a history more than 2000 years old, about which lots has been learned.  So, the modern day experts on how to effectively stipulate what an ontology will mean by the term 'thing', as well as how to effectively construct defintions,  have a great deal of help from the shoulders on which they stand.   This so-called standard, on the other hand, is like a nightmare of jumbled words that might have popped up in the 2000 years of incremental improvement in our understanding of what is useful to cast as a 'thing', and when.

 

I am sure the roots of these patterns are economic, as are the roots of most things, but  in the past few millenia of science and engineering, there have been many celebrated endeavors motivated by the desire to increase our understanding of things.  (or maybe, our understanding of  "sets of individuals which are defined according the facts (properties) given for that kind of thing. ")

Although unnecessary, I can't help but repeat Pat's quote from the standard and attack a few of its more obvious failings.   Here is what these people sponsored by the OMG and propose as a standard for the Financial Industry:

"A Thing is defined as the set of individuals which are defined according the facts (properties) given for that kind of thing. "

Well, in general, circular definitions are undesireable. 

So, they have the concept of individual, as a primitive from which Thing is constructed. So, presumably, individuals are NOT things.


Second, they do not distinquish between facts and properties, suggesting that they will have a really hard time when it comes to dealing with the different meta types in the ontology.

The capstone of the absurdity, though, is that in defining thing, they rely on, as part of the definition of thing, the concept of 'KIND of thing.'  How are we to know what a kind of thing is, before we know what a thing is? 

Looking at this some more though, alot of the nonsense in the official semantics for the Unified Modelling Language seems to be embedded here, in that they seem to be making the same fallacy, that individuals and types are so entirely different in nature that we can't even talk about them in the same language.    Of course, the so-called 'semantics' of UML itself has the same disfunctional roots as this stuff does.   Lets make some money for our companies that make tools for O-O languages.

No wonder ontology is not more widely respected, if this is the kind of stuff that happens.


Wm

 


 

 

On Sun, Dec 14, 2014 at 2:17 PM, Pat Hayes <phayes@xxxxxxx> wrote:

This is hopelessly confused. The technical part of it is nonsense. Just as a sample:

"6.3.3.1 Thing
A Thing is a set theory construct. This is shown on the diagrams as a box with a name. On some diagrams, additional
textual entries in the box show the Simple Properties about that thing.
A Thing is defined as the set of individuals which are defined according the facts (properties) given for that kind of thing. "

Words fail me at this point. How is it possible for educated adult human beings to get themselves so unbelievably muddled over what should be one of the simplest ideas ever stated?

Pat



On Dec 14, 2014, at 12:40 PM, John F Sowa <sowa@xxxxxxxxxxx> wrote:

> A report on the FIBO project:
>
>    http://edmcouncil.org/view/reports/20141121_FIBO_Report_to_Members.pdf
>
> See below for excerpts from the FIBO Semantics Repository Home Page.
>
> John
> ___________________________
>
> Source:  http://www.edmcouncil.org/semanticsrepository/index.html
>
> This website provides a partial report of sections of the Financial
> Industry Business Ontology (FIBO). This is being submitted to the Object
> Management Group (OMG) as a set of proposed standard ontologies under
> the FIBO umbrella. These FIBO OMG specifications are optimized for
> semantic technology applications.
>
> Alongside these we are working to release the full canonical reference
> ontology (as seen in these pages) as RDF/OWL...
>
> _________________________________________________________________
> 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
>
>

------------------------------------------------------------
IHMC                                     (850)434 8903 home
40 South Alcaniz St.            (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile (preferred)
phayes@xxxxxxx       http://www.ihmc.us/users/phayes



_________________________________________________________________
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
 

_________________________________________________________________
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>