On Feb 16, 2013, at 2:12 AM, John F Sowa wrote:
1. Implicit: The only ontology about the subject is in the mind
of the programmer, and not in any form that can be examined
and used by any other agent -- human or computer.
2. Stated in the requirements. The ontology is represented,
formally or informally, in some document that specifies the
domain of the application and requirements for what it does.
3. Object level. The ontology is represented in a form that can be
processed by programs that use it to reason about the subject.
4. Object level + Metalevel. The ontology is explicitly represented
in a form that can be processed by tools that reason about the
ontology, translate it to object-level operations on the data,
and use it to reason about and process the data.
In our discussions in Ontolog Forum, we have focused on levels 3
and 4. Levels 1 and 2 also use ontology, but not in form that is
formally represented and processed.
Fair enough. Excellent peeling the onion.
I focus on #1.
I'd be more than happy to argue that #2 is very, very suspect.
What do you calculate the distance is between #4 & #1?
I would argue that to a large extent what "formal requirements" exist are only the code since it's what runs in production tonight. The formal, written documentation that may have existed initially has been allowed to lapse.
The sort of information systems I deal with are heavily skewed to one-off, 100% custom, singular works. That is... IBM's accounts payable system/process is entirely unique. There is no such collection of artifacts anywhere else. Every business is a collection of such artifacts. How it all works is beyond me (spreadsheets I guess), but somehow the world seems to continue to lurch forward on this assembly of off-round wheels.
As I think I remember in the VivoMind examples, "the system" is also a combination of the rules as formally expressed in the code plus how the data works. Isn't there an example of how VivoMind uncovered using the data—"overloading"—to pay a computer to express overtime for a person? [I'm sure I've jumbled the precise findings.]