Room-based vs. coordinate-based

[S001GMU at nova.wright.edu]

Date: Thu, 12 Jun 1997 19:54:49 +0000 (PST8PDT)
From: clawrenc at cup.hp.com

>Change the scenario to:
>
>  Shoehorn is in a room,  
>  There are two other players A and B in the room as well.
>  A is near shoehorn.
>  B is some distance away.
>  There are also 15 tiny or invisible objects in the room 
>    near shoehorn (lint ball, a dropped cherry stone, etc).  
>  The extra objects create the following neighborhoods:
>
>    #1{Shoehorn,...A}
>    #2{A,...B}  
>
>  Shoehorn pulls a rabbit out of his hat.
>  The rabbit is too distant to be a member of #2.
>  #1 is too full to add the rabbit, resulting in:
>
>    #1{Shoehorn,...A}
>    #2{A,...B}  
>    #3{Shoehorn, rabbit}
>
>All the objects tiny, invisible, A, B, etc should see Shoehorn pull
>out the rabbit.  As far as A & B are concerned, it is the only thing
>happening in the room, and as such is not drowned in detail (or 50M
>rabbits).

I'd be tempted to add a scale to each neighborhood.  The scale would 
be a range of object sizes allowable within that neighborhood.  anything
too big or two small gets shuffled into a neighborhood of the appropriate
scale, creating a new one if none exists.   Neighborhoods of differing
scale can overlap eachother with no real problems.

Are there any glaring problems with that?  I've had some difficulty 
following exactly what you guys are talking about, but that probably
stems from not really reading the first couple of messages too thuroughly.
Will have ot go back and re-read those.  ;)

-Greg