[r-t] Delta-based transposition representation

[Richard Smith]

Alan Burlison wrote:

> ... and you trivially know if a change is valid by summing 
> the deltas - the result must be 0.

Picking up on a point Don didn't mention, this isn't true. 
Certainly if it does not sum to zero then the change is 
invalid.  But you're arguing the converse, that if the 
detals sum to zero the change is valid.  That's not true: 
invalid changes can have deltas that sum to zero too. 
Consider +1 +1 -1 -1.

RAS