Re: adjacency info

[Seth Teller <teller@csail.mit.edu>]

vitaly et al.,

is this necessary?  i would not be in favor of
adding this complexity to the design, unless it
is absolutely necessary.


note that if we can support coincident
(but oppositely directed) portions of contours,
we can support "holes" or "nested contours."

consider this example:

    AAAAAAAAAAAAAAAA
    A              A
    A   AAAAAAAAA  A
    A   A       A  A
    A   A       A  A
    A   A  (c)  RRRR
    A   A       LLLL
    A   A       A  A
    A   AAAAAAAAA  A
    A              A
    AAAAAAAAAAAAAAAA

where the "R" and "L" segments overlap and are
oppositely directed.  this defines a closed region
inside "AAAA..." that excludes the center region
(c).

isn't this sufficient for what we want to do?
if not, please explain why.

prof. t.


Vitaly Kulikov wrote:
> Hi guys,
>
> I've successfully got to Ukraine. I am planning to stay here for some 
> time and go back to Boston on January 9th.
>
> I have a question. As far as I understand, the current adjacency format 
> does not support spaces specified by more than one contour. This is a 
> problem if we want to allow spaces with other spaces inside. What do you 
> think, can we modify the format to support more than one contour per 
> space, the way UG format works?
>
> Vitaly
>
> _________________________________________________________________
> Enjoy the holiday season with great tips from MSN.  
> http://special.msn.com/network/happyholidays.armx