[EM] Regarding the districting solutions. My proposal in greater detail.

[Michael Ossipoff]

Ralph--

> On Fri, Jun 8, 2012 at 3:50 AM, Michael Ossipoff <email9648742 at gmail.com>
> wrote:
> 
> 
> 
> 	Starting from the top of the map, draw a horizontal line that
defines the
> 	bottom of the topmost band. Position that line so that that 1st band
will
> 	contain an integer number of districts--the population in the band
must
> be D
> 	multiplied by an integer.
> 
> 
> 
> This is pretty similar to shortest splitline.
> 


It has things in common with shortest splitline, but is less
computation-intensive.


> I presume for your first step you enclose the state is a bounding
rectangle?

No need to.

> 
> It seems that picking the line nearest the midpoint will give the best
ratio
> balance between the 2 sub-states.

That would be so if 1) the region being divided contains an even number of
district-populations; and 2) the population in that region is uniformly
distributed.

> 
> If you do that, then you get shortest splitline, but with the lines
restricted to
> being horizontal and vertical.

Band Rectangle Districting operates differently. It's less
calculation-intensive, because Shortest-Splitline requires trying every
possible orientation for the dividing line.

And of course Shortest Splitline successively divides regions into halves or
near halves, whereas Band Rectangle's divisions are not halvings. (but one
or more of them _could_ be, of course).
 
> It has the nice feature that you don't need to consider the State's
boundary at
> all.  You just need a list of positions of block populations.

Of course some of the districts are partly bounded by a state boundary or
coastline, no matter what the districting method is.

I don't know what is meant by a "block population".

Mike Ossipoff