ENRAC One-Year Anniversary
The News Journal for Reasoning about Actions and Change started exactly
one year ago - the original monthly webpage appeared on July 31, 1997.
As most of you know, we switched to the system with direct-mail
Newsletters in September, and that's clearly the major reason why the
interactions have been very lively since. The present one is the 97:th
Newsletter, so another reason for celebration is coming up soon.
ETAI Volume 1 Completed
Just in time for this anniversary, we have now completed the official
version of ETAI 1997. (Please remember that each annual volume contains
those accepted articles that were submitted during that year, except
of course those requiring major revision, so the delay of a number of
months is quite normal). The completed volume is available, each paper
in its own file(s), on the Linkoping E-Press server, and the ETAI webpages
for the Journal have been modified accordingly. In particular, responding
to a suggestion by Rob Miller in an earlier Newsletter, we now have an
overview page for the annual volume that gives parallel access to the
article itself and to the discussion about the article.
At this point, I feel that the basic idea with this electronic journal
has become very clear in the webpage structure. While in the midst of
Newsletter activities - reading and contributing to them - one may
sometimes have the impression that this is the ETAI, or more precisely
one of its sections. At the end of the year, however, we have the
articles that have been discussed, revised, and refereed through the
Newsletter-based interaction process, and for posterity ETAI will of
course be the collection of those articles and their commentary.
Please take a look at the ETAI journal page (accessible from the main
ETAI page) to see things in this perspective.
All of this has of course been produced by you the subscribers/ peers/
participants/ colloquium members. One other key notion in the ENRAC,
in particular, is that debate contributions ought to be produced with
a look-and-feel that matches their quality and the work that got into
them, or in brief, that everyone who contributed to the discussions
shall be able to feel pride for it and be able to refer to it. This is
one of the reasons why the News Journal (containing the accumulated
contents of Newsletters, one News Journal issue per month) is produced
not only in HTML form but also in postscript via latex. I feel that much
of what is written in our debates has a lasting value, and that it is
worth referring to. Please feel free, therefore, to cite ENRAC debates
just like you would cite any other journal. It does have official status,
and its contents can be characterized in the same way as for other
journals, in terms of ISSN number, volume, page, and so on. In particular,
in order to refer to the discussion about an article, the appropriate
citation is to the News Journal issue where the comment was made.
The News Journal is organized with separate segments for separate
articles being discussed.
Today's debate contribution
Today's Newsletter contains a contribution to a `panel' discussions:
Jixin Ma's answer to the recent contribution by Erik Sandewall in
the debate about the ontologies of time.
|
Ontologies for time
In Newsletter ENRAC 27.7 (98059), Erik wrote:
| Jixin has observed that sometimes it is necessary to assume an
intervening point between two intervals, for example, for being able
to say that a ball that has been thrown vertically up into the air
will be motionless (zero velocity) at a certain point in time. For
some other actions or events it is desired not to have an
intervening (time-)point, as is the case for two successive
intervals, one where a switch was "on" and the next where the switch
was "off". Sergio Brandano has already observed (ENRAC 23.4) that
the timepoint domain will then be agent-specific. It will also be
local to each scenario, so that if additional actions are added to
the scenario description, one must change the timepoint domain
accordingly. In fact, it even becomes necessary to revise the
timepoint domain each time a query is asked for a scenario, which
seems a bit odd.
| First of all, it is important to note that, by taking both intervals
and points as primitive, the time domain is general enough for
various scenarios, and hence does not need to be revised when a
"query" is asked for a special scenario. In fact, although the
time line consists of both intervals and points, it is not required
that between any two intervals there must be a point standing there.
Therefore, the time domain itself allows the case of two successive
intervals over some parts of the time line, that is an interval meets
another interval without any point standing between them (this is the
case 1 as I classified in ENRAC 13.3). But this does not
necessarily exclude the case where a point "Finishes" the first
interval or "Starts" the second (Please bear in mind that an
interval "Meets" a point is treated as different with the case the
interval is "Finished-by" the point; similarly, a point "Meets" an
interval is different with the case that the point "Starts" the
interval. See below). Therefore, if needed, it can also cope the
other two cases, i.e., case 2 and case 3, by imposing futher
information that a point "Finishes" the first interval, or "Starts"
the second, respectively). In addition, it supports the case that an
interval meets a point which in turn meets another interval (this is
the case 4 which is needed in modelling the motion of the a ball
being thrown vertically up into the air thrown). All these four cases
are allowed by the same time theory, without the need to revise the
time domain at all. In fact, one can have some of these scenarios
appear somewhere over the time line, meanwhile with other scenarios
appear over somewhere else. Specially, as Pat pointed out, one may
even have the case where over somewhere actions or events are
modelled as discrete, while over somewhere else (of the same time
line) actions or events are modelled as continuous.
| I suppose one could get used to this mode of thinking, and to accept
that time is in the mind of the beholder. However, here are two
examples where the punctuated time approach leads to absurd
conclusions.
A. Jim fires a model rocket and observes its flight. At the moment
when it reaches the top of its trajectory, he turns a switch. If the
flight of the rocket is modelled like Jixin proposes to model the
throwing of a ball, then it requires that there exist a point for
the clocktime when it is at its apex, but at the same time the fact
that Jim turned the switch implies that there must not be any such
point. Therefore, if the scenario description includes the statement
that Jim turned the switch at the same (clock-)time as the rocket
reached the apex, then it is semantically inconsistent.
| I think this example is similar to the following one as raised by
Galton:
Suppose two lights, Green light and Red light, are both switched
at the same time point P. By commonsense, we have that
(C) "GreenOff Meets GreenOn" and "RedOff Meets RedOn".
In addition, assume that by some reason we impose that the Green light
is, for instance, On, at the switching point P, how are we going to
express the situation for the Red light? In other words, is the Red
light "Off" or "On" at the switching point P? Galton was afraid that
this would lead to the Deviding Instant Problem again.
Let me sort out this problem first and then show that Erik's example
can be dealt with similarly.
Since we have no information about the the state of the Red light at
the switch point P, we may just use two successive intervals, I1 and
J1, to express the scenario as:
|
Holds(RedOff, I1),
| (R) |
|
Holds(RedOn, J1),
| |
|
Meets(I1, J1)
| |
However, we do have the addition information that the Green light is
on the switching point P, therefore, the expression for the situation
of Green light could be:
|
Holds(GreenOff, I2),
| (G) |
|
Holds(GreenOn, P),
| |
|
Holds(GreenOn, J2),
| |
|
Meets(I2, P),
| |
|
Meets(P, J2)
| |
And hence, we can express the assumed whole scenario as:
|
Holds(GreenOff, I2),
| (GR) |
|
Holds(GreenOn, P),
| |
|
Holds(GreenOn, J2),
| |
|
Holds(RedOff, I1),
| |
|
Holds(RedOn, J1),
| |
|
Meets(I2, P),
| |
|
Meets(P, J2),
| |
|
Meets(I1, J1).
| |
where I1+J1 = I2+P+J2
Since (GR) allows us to express P+J2 , i.e, the ordered union of P and
J2 , as a single interval J2' , such that Holds(GreenOn, J2') , it
therefore provides a satisfactory expression consistent with (C).
However, if we in addition specially impose that, for instance, the
Red light is Off at the swiching point P, we can still express the
whole scenario as:
|
Holds(GreenOff, I2),
| (GR*) |
|
Holds(GreenOn, P),
| |
|
Holds(GreenOn, J2),
| |
|
Holds(RedOff, I2),
| |
|
Holds(RedOff, P),
| |
|
Holds(GreenOn, J2),
| |
|
Meets(I2, P),
| |
|
Meets(P, J2).
| |
Again, since (GR*) allows us to express P+J2 as interval J2' , such
that Holds(GreenOn, J2') , and also to express I2+P as interval I2' ,
such that Holds(RedOff, I2') , again, we reach a satisfactory
expression consistent with (C) for the specified scenario.
Now, let's come back to Erik's example A:
First of all, I don't think the time theory criticized is
semantically inconsistent. Yeah, for the modelling of the throwing of
a ball, it requires that there exists a point referring to the apex.
However, the fact that Jim turned the switch does not necessary imply
that there must not be any such point, especially if one insists that
"at the moment (point?) when it (the ball) reaches the top of its
trajectory, he (Jim) turns a switch". I guess what Erik actually
means here is that we don't have any information about the state of
the object (a light?) being switched at the switching point. What we
do know is just that the "On" state is immediately after the "Off"
state. Therefore, similar to the treatment to the above "Two Lights
Problem", we can express Erik's scenario as:
|
Holds(BallGoingUp, I1),
| |
|
Holds(BallStationary, P),
| |
|
Holds(BallGoingDown, J1),
| |
|
Holds(LightOff, I2),
| |
|
Holds(LightOn, J2),
| |
|
Meets(I1, P),
| |
|
Meets(P, J1),
| |
|
Meets(I2, J2).
| |
possibly with the constraint: I2+J2 = I1+P+J1
| B. Tom and Bob compete for eating icecream cones. They start with
four cones each, and have to eat them as fast as possible, starting
at the same time. The referee rings a bell when one of them has
finished eating all his cones. If one of the contestants tries to
cheat by dropping icecream on the ground, he also rings the bell in
order to call off the contest. The bell sounds at time t . What
conclusions can be drawn from this?
Consider the following ontology for this domain which is plausible
if one uses punctuated time in order to deal with the DIP: 1. The
ringing of the bell is momentary, that is, it occurs at a particular
point in time. 2. The cone is supposed to still exist while it is
being eaten, it does not exist after it has been consumed, and the
clocktime at the end of the eating period does not have a
corresponding point. This convention is made in order to avoid the
dividing instant problem with respect to the existence of the
icecream cone that has just been eaten. With these assumptions, it
follows that one of the contestants cheated at time t , because in
the case of an honest game, a point both exists and does not exist
at the clocktime when the winner finishes his last cone and the
bells rings.
It is easy to generate additional examples of the same kind. I don't
see how one can get around them as long as punctuated time is
allowed. The only reasonable approaches seem to be either (a) the
transitory value approach, where one uses an additional fluent
value for "undefined or "transitory" for e.g. the state of a switch
in the moment it is being thrown, or (b) the multiple models
approach that I mentioned above. Both of these approaches are
sufficiently expressive for dividing instant situations even when
complete (unpunctuated) time is used. How else would you deal with
these examples?
| First of all, I don't agree with the claim that "a point both exists
and does not exist at the clocktime when the winner finishes his last
cone and the bells rings". Again, I think this claim was reached by
means of confusing the two cases, that is, the case that an interval
"Meets" a point, and the case that an interval was "Finished-by" a
point. This is in fact the approach taken by many researchers such as
Vilain (1982) and Parthasarathy (1990, 1994), and which actually
leads to the problem. However, as I shown in ENRAC 1.4, these two
cases should be taken as different since they play the critical roles
in characterising the closed and open nature of intervals, and
treating the corresponding problems.
Now, let me try to model this interesting example:
For Tom, we have:
|
Holds(f1, I1),
| |
|
Holds(f2, I2),
| |
|
Holds(f3, I3),
| |
|
Holds(f4, I4),
| |
|
Meets(I1, I2),
| |
|
Meets(I2, I3),
| |
|
Meets(I3, I4)
| |
where f1 denotes that all the four cones have not been finished by
Tom, and f2 denotes that just one cone has been finished by Tom, and
so on.
Similarly, for Bob, we have:
|
Holds(g1, j1),
| |
|
Holds(g2, j2),
| |
|
Holds(g3, J3),
| |
|
Holds(g4, J4),
| |
|
Meets(J1, J2),
| |
|
Meets(J2, J3),
| |
|
Meets(J3, J4)
| |
To express that Tom and Bob start at the same time, we have:
|
Meets(I, I1),
| |
|
Meets(I, J1)
| |
Also, we have:
If Tom finishes his last cone before Bob, that is:
|
Duration(I1+I2+I3+I4) < Duration(J1+J2+J3+J4)
| |
then we have: Meets(I4, t)
Similarly, if Bob finishes his last cone before Tom, i.e.,
|
Duration(J1+J2+J3+J4) < Duration(I1+I2+I3+I4)
| |
then Meets(J4, t)
As for the cheating:
If
|
Holds(TomDroppingIcecream, K) ^ During(K, I1+I2+I3+I4)
| |
then Meets(K, t)
Similarly, if
|
Holds(BobDroppingIcecream, L) ^ During(L, J1+J2+J3+J4)
| |
then Meets(L, t)
I feel that's it, though I wonder if I missed some of the
specifications of the example.
Jixin
|