SUO: Re: Lifecycle Integration Schema
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
LIS. Discussion Note 2
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
Matthew,
New comments unindented below.
Re: "thing"
http://www.tc184-sc4.org/wg3ndocs/wg3n1328/lifecycle_integration_schema/lexical/thing.html
LIS: | A <thing> is anything that is or may be thought about or perceived,
| including material and non-material objects, ideas, and actions.
JA: A statement that begins "An X is any X that ..." is problematic.
JA: I will pass it by for now, and treat it as a comment that tells me
something about how you plan to use the word "thing", in some actual
or potential relation to the things that you call "perceptions" and
"thoughts", and applicable inclusively to all of the things that you
call "actions", "ideas", "material objects", and "non-material objects".
MW: It is normative text in the standard, but you are right not to worry too
much about it. Thing is nearly equivalent to John Sowa's TOP, but within
a 4D ontology, so a continuant is not an ISO15926 thing, since it doesn't
fall within the paradigm. In a lattice of theories it would be the root
for this 4D ontology, and fall somewhere under John Sowa's TOP.
It's possible we use words like informative and normative differently.
Let's see if there's some way to cut to the chase and say what we are
really doing when we spend time talking about words like "thing" and
"things". The only reason for spending time on these sorts of words
is that there is this sadder-but-hopefully-wiser history of crashing
whole systems of thought when people tried to take on too much at
these initial stages of the game.
I understand you to be saying that you intend to apply the word "thing" to
every object that arises in a particular discussion, and for my part I am
content to work within one context of discussion at a time, so let us not
spend too much time indulging in the speculation that we might just have
some concept of anything whatever, independent of all practical contexts
of discussion. If the instruction manuals for our logical formalisms
do not come with formalizations of these "needless to say" sorts of
understandings, then it is likely that we will need to write them,
if not for our sake, then for the sake of computers yet unborn,
Taken in a particular context of interpretation, then,
there is a universe of discourse X, and there are terms,
for example, the word "thing", that denote every element
in X, or, said another way, whose denotation ranges over
the whole universe of discourse X that is pertinent to
a particular discussion.
At this point, at least for my part, interest shifts to the many different
universes of discourse that one may consider, the sorts of structures that
they may have and the sorts of relationships that may exist among them.
That is more or less why category theory got invented, to help us keep
track of the ranges of variable words like "thing", which ranges of
variation we now call objects in a category.
MW: For the purpose of the SUO we would need to
reword this definition at some point.
Yes, but ...
I still cannot see it as a definition,
but it's okay to begin with undefined
primitives, and statements of their
relations to other terms, defined
or otherwise.
LIS: | Every <thing> is either
| a <possible_individual>,
| or an <abstract_object>.
JA: This is an informative statement.
MW: But normative in standards terms.
By "informative", I just mean that it conveys information.
That is, the statement expresses a constraint among the
things to which it and its component terms apply.
JA: It specifies a mutual constraint on the applications of the three terms,
"thing", "abstract_object", and "possible_individual". Moreover, it's
a constraint that can be captured in zeroth order terms (boolean algebra,
monadic predicates, propositional calculus, sentential logic, etc,) and
expressed very succinctly in my favorite language for same, so I will
capture it immediately in the following form:
JA: ( thing ,( abstract_object ),( possible_individual ))
JA: Roughly speaking, taken as an assertion, this says that every thing
is either an abstract_object or a possible_individual, but not both.
MW: Correct. If had had gone a couple of lines down
you would have found the following EXPRESS:
MW: ENTITY thing
ABSTRACT SUPERTYPE OF (ONEOF(possible_individual, abstract_object));
MW: ABSTRACT means it is only instantiated through its subtypes, and
ONEOF means one or the other but not both. ABSTRACT is indicated
in the diagram by the (ABS) in the entity type name, and the ONEOF
is indicated by the "1" by the subtype/supertype tree.
MW: We often did not bother to give these statements informally
if the information was contained in the formal EXPRESS.
Good, there are many interesting things here,
that I hope to get back to at another time.
LIS: | NOTE 1. Every <thing> is identifiable within a system.
| System identifiers created by other systems and received
| as part of a data exchange may be stored for future reference
| as an identification, referring to the originating organisation
| or system.
JA: You are saying that you are really only thinking about things
that have identifiers in given system, or names in a certain
context of discussion. This is very significant and needs to
be elevated to the level of an explicit principle, instead of
being left implicit in a note that is likely to be neglected,
in other words, relegated to a hidden axiom or constraint
whose consequences are not critically reflected on.
MW: Firstly NOTEs and EXAMPLEs are informative in standards terms.
I probably need a definition of what you mean by "informative".
It sounds like you are saying "not definitive" or "optional".
The way I see it, if it's not avoidable, it's not optional.
However we gloss it, I am trying to indicate a potential risk of
misinterpretation, of being taken to mean something that is very
different from what one really intends, and there is a potential
danger of, well, glossing over this point if we don't keep all
the constraints and intentions in plain view. There is also
an efficiency loss that comes from trying to do more than
one really needs.
MW: A data model is to support the holding of data about things.
Here we are noting that different systems that might wish to
communicate about some thing may have their own identifier
for it, and that this is not a problem.
LIS: | NOTE 2. Every example provided for other entity data types
| declared in this schema is also an example of <thing>.
JA: Okay for now, as long as it's just a comment.
MW: It is, see above.
This is beginning to worry me more.
We may need a way of formalizing
how seriously comments need to
be taken.
JA: There are some generic issues already arising here,
that come up in every discussion of this type, but
I'll need to think about some new ways to say them.
Jon Awbrey
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o