[TAG] Max-spread algorithm

[Neil Youngman]

On or around Thursday 05 July 2007 15:17, Ben Okopnik reorganised a bunch of 
electrons to form the message:
> On Thu, Jul 05, 2007 at 08:28:19AM +0100, Neil Youngman wrote:
> > I would try something like
> >
> > assert( pork_cubes > 1 );
> > gaps = pork_cubes - 1;
> > fill = tomatoes/gaps;
> > toms_used = 0;
> > for( g = 0; g < gaps; ++g )
> > {
> >   toms_in_gap[g] = ((g+1) *fill) - toms_used;
> >   toms_used += toms_in_gap[g];
> > }
> >
> > I think this still biases a little towards one end, but it shouldn't be
> > too hard to add a little tweak that fixes that.
>
> It's a nice approach - it creates a list of gaps and details how many
> tomatoes should be inserted into each gap. However, it still has a
> problem - both the tomatoes and the pork chunks are integers (cutting a
> tomato to 0.666666666666667 of its original size is a little difficult -
> and recombining the remainder with the next tomato down the line is even
> more so. :) Besides, I'm struggling a bit to see how I'd convert that to
> an actual interleaved array (it may just be because I'm a bit slow on
> the uptake this morning...)

Yeah. I didn't really make the mixing of integers and floating point numbers 
clear enough. Variables toms_in_gap  and toms_used were always meant to be 
integers :-) Would something like

  toms_in_gap[g] = round(((g+1) *fill) - toms_used);

fix it?

Neil