Dear All,
In Paris we decided to add to the E2 Temporal Entity the well-known temporal
relationships from
J.F.Allen, 1983, which can be fairly regarded as a standard in Knowledge
Representation.
I terms of directed, bidirectional properties as we use in the CIDOC CRM, they
can be defined as:
before (after) : E2 Temporal Entity
meets in time (met-by in time): E2 Temporal Entity
overlaps in time (overlapped-by in time): E2 Temporal Entity
during (includes in time): E2 Temporal Entity
starts (started-by): E2 Temporal Entity
finishes (finished-by): E2 Temporal Entity
equal in time: E2 Temporal Entity
where the postfix "in-time" is used to disambiguate from spatial or
spatiotemporal relationships, that appear
lower in the IsA hierarchy of entities in the CIDOC CRM.
Here my proposal for a scope note:
The temporal relationships relate a Temporal Entity X, the domain, with a
Temporal Entity Y, the range.
Let us denote the real outer temporal bounds for the Entity X and Y by the
dates X-,X+ and Y-,Y+, where
"-" describes the lower and "+" the upper bound. These are the real narrowest
limits, in which the
Temporal Entity occurred, independent from our knowledge about them. The
relationships in parenthesis,
like (after), are the inverse ones, i.e. those which we obtain by interchanging
X and Y in the definition.
before (after) : X ended at a date before the one when Y started : X+ < Y-
meets in time (met-by in time) : X ended at the date when Y started : X+ = Y-
overlaps in time (overlapped-by in time): X started before Y started, but ended
within the duration of Y: X-<Y-, Y-<X+, X+<Y+
during (includes in time): X started and ended within the duration of Y:
Y-<X-, X+<Y+
starts (started-by): X started with Y but ended before it: X-=Y-, X+<Y+
finishes (finished-by): X started after Y but ended simultaneously: X-<Y-,
X+=Y+
equal in time: X started with Y and ended with Y: X-=Y-, X+=Y+
========================================================
My intuition about cultural documentation practice leaves me when to decide
about the application of these relationships
to spatiotemporal cases, or other dimensions of possible specializations of
Temporal Entities.
As we have decided not to attach these properties to the associated time spans,
there are two possible interpretations:
Given two "bubbles" in Space-Time, like Bronze-Age and Iron-Age, a term like
"overlaps in time"
may mean:
1) At any given point in space, at any given settlement, Bronze-Age overlaps
more or less Iron-Age.
2) We don't know if at any given point in space, Bronze Age overlaps with
Iron-Age, but at least when Iron-Age
first started in settlement A, Bronze-Age still prevailed in settlement B,
and when Bronze Age started, Iron-Age has
not started anywhere, and before Iron-Age ended in the last settlement,
Bronze-Age ended in its last settlement.
Both interpretations are equally valid, but I think one of those must be
culturally more relevant, i.e.
closer to the phenomena we regard worthwhile documenting.
If interpretation 1) is the relevant one, the relationship "falls within" of
E52 Time-Span is NOT synonymous
with the relationship "during" of the associated Time-Spans, which makes some
sense. Another aspect is,
that trivial relationships between known dates need not be expressed by
explicit temporal relationships.
Those are only needed for phenomena with nknown absolute dates.
If interpretation 1) is the relevant one,
"overlaps","starts","during","finishes" imply spatiotemporal overlap -
in contrast to interpretation 2). This may be wanted. This may make the "pure
spatiotemporal overlap"
advocated for in one of my previous messages unnecessary.
If interpretation 1) is the relevant one, "equal" becomes spatiotemporally
equal. This may be improbable for
any real case, except the cultural periods of Atlantis starting end ending with
its emerging and drowning in the sea...
A slight modification of interpretation 1) may be even more realistic:
At any given point in space, at any given settlement, where both, Bronze-Age
and Iron-Age occurred, Bronze-Age
overlaps....
This leaves space for spaces, where the one or the other period did not occurr.
This renders a "temporally equal"
different from a "spatiotemporally equal" (or "coocurrent"). Alternatively,
"temporally equal" may be defined only
on the associated Time-Spans.
As these operators were required by archeologists, we need a competent comment
on this issue.
Martin
--
--------------------------------------------------------------
Dr. Martin Doerr | Vox:+30(810)391625 |
Principle Researcher | Fax:+30(810)391609 |
Project Leader SIS | Email: [email protected] |
|
Information Systems Laboratory |
Institute of Computer Science |
Foundation for Research and Technology - Hellas (FORTH) |
|
Vassilika Vouton,P.O.Box1385,GR71110 Heraklion,Crete,Greece |
|
Web-site: http://www.ics.forth.gr/proj/isst |
--------------------------------------------------------------
