>World simulation
C.J. Appelo
Phil Goets talked about objects that are openable or transparent.
Well Phil, I have some teasers for you. I like your idea of allowing
sight through an object when it is opened of transparent. However,
the model isn't finished yet. I thought of a birdcage that has to
be closed in order to keep the bird in. Still, it is possible to
see the bird in the cage. You could say that the cage is transparent,
but this gives rise to another problem. When the bird is in a glass
bottle, it will survive (for a while) when the bottle is thrown in
a swimming pool. When this is done with the birdcage, the bird will
die (unless it is rescued by Superbird). So transparent and ever-
opened objects must be treated differently.
I was thinking of modelling light, sound and air (or liquids), so you
can specify how they "go through" (? what is that English word again?)
objects. So, sound, for example, spreads among links in the semantic
network. The volume reduces when it passes a link with a filter
(a door, a container). For the birdcage the filter for air/liquids
(or just substances) is 0 and for sound and light too. For a glass
bottle the filter for light is (nearly) 0, for substances it is 100
and for sound it is somewhere in between. (When a filter is 0 it lets
everything through, when it is 100 it blocks.)
This could be an answer to te problem Jamie mentioned, where walls
should direct sound only in a single direction. It also solves
another problem I had with the idea of an area of effect. The area of
effect for sounds should be different for different sounds. You can
hear a bomb explode on the other side of an island, but the singing
of the bird will have much less influence (especially when he's in
the swimming pool). The sound of the bomb will have a much higher
starting level. Also light and sound travel in different ways,
and should have different areas of effect. You could easily model
a recording studio, with a sound-proof glass window.
I only came up with this idea of filters a few hours ago. It will
certainly have many drawbacks as well I suppose. Increasing complexity
will be one of them. But this is the case with every reality supporting
idea. What I am thinking of right now is: should objects detect
themselves wether a sound can be heard from somewhere of should they
be told so by the noise making object? The difference between them is
the backtracking that takes place in opposite directions.
I just thought of something else. When light comes from different
sources, the resulting levels of light should be added up when light
is needed in an object. How about sound and substances? (Oh my god,
what have I started?)
Another subject: (This one is for Phil again.)
Phil mentioned that he had difficulties in having a key under a rug,
for there was no place for a rug on a key. The first thing that came
into my mind was that with a ball under a cupboard you seem to have
the same situation. I thought that, apart from having On/Under
connections, we also need Above/Beneath connections (or whatever
you want to call them). These have other restrictions on size and
no restrictions on weight. This should cure the key-under-rug syndrome.
My happiness about this solution lasted for only a few minutes. With
"Beneath" you should be able to put anything under an object that has
room for this connection. But what if you put your mothers applepie
under the rug? I bet she won't be pleased.
Now I have three possibilities when object A is under object B.
1. The weight and size of B both influence A.
A has to have room for B to be put on top of it.
(table under book)
2. Only the weight of B influences A.
(pie under rug)
3. No attribute of B influences A.
B has to have room for A to be put under it.
(ball under cupboard)
I cannot think of examples in which only the size of B influences A.
If there is an example then a fourth possibility should be added to the
list.
Do you believe these nasty things almost kept me awake last night?
It is horrible, but very fascinating.
Jurgen.
===================================================================
| Jurgen Appelo | |
| Delft University of Technology | When Technology Is Master, |
| The Netherlands | You Reach Disaster Faster. |
| app...@dutiag.twi.tudelft.nl | |
===================================================================
Adam Justin Thornton
make for a very realistic setting. Sure, it would be large
to encode, since the effect would be something like raytracing
(though each wall is, I guess only one point), but with today's
memory sizes...neat! If anyone implements this, I'd love to see it.
I'd also like to see the ALAN Floyd code.
Adam
--
"And in the heartbreak years that lie ahead, |++| ad...@rice.edu |++| Cthulhu
Be true to yourself and the Grateful Dead." --Joan Baez | 64,928 | fthagn!
"Very often, a common stone, thrown away and despised, is worth more than
a cow." -- Paracelsus | If these were Rice's opinions I'd shoot myself.
Phil Goetz
>
>I was thinking of modelling light, sound and air (or liquids), so you
>can specify how they "go through" (? what is that English word again?)
>objects. So, sound, for example, spreads among links in the semantic
>network. The volume reduces when it passes a link with a filter
>(a door, a container). For the birdcage the filter for air/liquids
>(or just substances) is 0 and for sound and light too. For a glass
>bottle the filter for light is (nearly) 0, for substances it is 100
>and for sound it is somewhere in between. (When a filter is 0 it lets
>everything through, when it is 100 it blocks.)
>
>This could be an answer to te problem Jamie mentioned, where walls
>should direct sound only in a single direction. It also solves
>another problem I had with the idea of an area of effect. The area of
>effect for sounds should be different for different sounds. You can
>hear a bomb explode on the other side of an island, but the singing
>of the bird will have much less influence (especially when he's in
>the swimming pool). The sound of the bomb will have a much higher
>starting level. Also light and sound travel in different ways,
>and should have different areas of effect. You could easily model
>a recording studio, with a sound-proof glass window.
The idea of spreading activation for sound using filters is brilliant!
However, it requires a formalization of the connections between
locations. The connections that {Peter Weyrauch? Dave Bagget?
I can't remember} talked about in Oz might be like this. But then you
also have to add containing as a "connection"; in this case the filter
between A and C is the object B which contains C and is in A.
Phil Goetz
go...@cs.buffalo.edu
Disclaimer: My opinions do not reflect those of
the KGB. University of Buffalo.