[Fwd: Re: dejunk, etc]

[Seth Teller <seth@graphics.lcs.mit.edu>]

fyi.

-------- Original Message --------
From: willtor@mac.com
Subject: Re: dejunk, etc
To: "Allen Miu" <aklmiu@mit.edu>
CC: Seth Teller <seth@graphics.lcs.mit.edu>

Hey Allen,

Dejunk was the first of my projects over the summer, and I hadn't ever 
really
programmed in C, before, so I think my code isn't the most efficient.  
But here's
the basic idea:

The quarter-circular portion of the door has been identified, already by
acad2ug (the special, modified version).  Since that's been moved into 
the
door layer, already, I can find any wires that connect to the door 
(using my
handy-dandy data-structure).  I just check both ends to fit certain 
requirements
such as the equation that best represents it.  Once the door, in its 
entirety is
located, the door jambs lie off of each side.  If the door-jamb wire is 
actually
connected, then it's easy enough.  If it's not, Seth tells me that 
there's an nlogn
way to do this, but my algorithm goes in n^2, merely searching every 
possible
wire for each door.

As for hacking apart my code, go right ahead, but don't forget to modify 
the
documentation in /walkthru/BMG/end_to_end/documentation/

Sorry it's taken me so long to correspond.  Last semester was pretty 
crazy.

-wml