[r-t] Exhausted search spaces

[Richard Smith]

Alexander Holroyd wrote:

> I'm a touch sceptical that anyone has searched the whole space of extents of 
> plain bob minor.  I'd be more inclined to believe it if I saw the data.

Now you mention it, I'm sceptical of this too.  I left a 
search running over night, and in 9h it found over 30m 
extents modulo rotation and had only covered 0.04% of the 
search space.  Admitedly my search program progresses 
through the search space superlinearly -- i.e. it speeds up 
as it progresses due to the way it prunes rotations.  Also 
the code I used is very general purpose and doesn't include 
a number of optimisations that are specific to PB6 (or lower 
numbers in general).  So I reckon that this search is now 
with reach of a custom-written search program running 
constantly for a week or two.

But I cannot believe that this was done "many years ago" and 
the suggestion that the results were printed out is clearly 
wrong.  In a 6pt font, printed doubled-sided on A4 paper in 
two columns, you could get about 500 extents per sheet of 
paper.  The 30m I found last night alone would require 120 
reams of paper to print and, even on lightweight 70gsm 
paper, would weigh quarter of a tonne.

RAS