Re: SUO: Re: TLC in KIF
Jon,
I completely agree:
>But I will go ahead, anyway, and state the lesson that I, for one,
>learned from my honest, if naive, attempts, and that is just that
>formalisms like the "logic of deductive inference" (LODI) and the
>"theory of classes and sets" (TOCAS) are perfectly -- well, almost
>perfectly -- good forms of representation for succintly summarizing
>what you have discovered, at any rate, once you have discovered it
>usually by any other means possible, in what is often called the
>"context of justification and review" (COJAR), but that these
>brands of formalism are darn near useless and a positivistic
>nuisance when it comes to the discovery phases of the overall
>"context of inquiry" (COI) itself.
I made a similar point in one of my notes about the language
issue: a formal language in which every symbol has exactly
one definition is useless for the initial stages of design and
analysis when the specifications have not yet been pinned down.
One of the major advantages of NL is that it lets people
negotiate. Any two people who begin to talk about any topic
whatever are almost certainly going to have different purposes,
intentions, needs, desires, points of view, backgrounds, etc.
Even when they think that they are using the same words in the
same senses, they almost certainly will have divergent meanings
(in whatever way you want to define the meaning of "meaning").
It is only after some discussion and negotiation (sometimes
very prolonged and painful) do their meanings begin to converge.
Eventually, they might become as close as an old married couple
who can finish each other's sentences. But that stage requires
a long and very painful negotiation.
Those axioms that I presented in Chapter 2 of the KR book are
an example of how the final results of the analysis might
be presented. At this stage of the SUO project, a detailed
comparison of the KIF representations of the Lenat axioms with
the Sowa axioms with the Guarino axioms, etc., will only reveal
one major point: It's a long way to Tipperary.
John Sowa