[r-t] Gangnam (or whatever) etc.

[Don Morrison]

On Wed, May 14, 2014 at 6:51 AM, Simon Gay <Simon.Gay at glasgow.ac.uk> wrote:
> Consider the rows of the form 56xxx1 and 65xxx1.
>
> There are 12 such rows, and they occur 4 per lead, with lead heads
> of the form 15xxx6 or 16xxx5.
>
> In a multi-extent block in which each row occurs n times, we
> therefore have 3n leads which between them contain all the
> occurrences of 56xxx1 and 65xxx1.
>
> Within one such lead, all occurrences of 56xxx1 or 65xxx1 have the
> same parity. Therefore half of these leads need to have positive
> lead heads, and the other half need to have negative lead heads.

This argument only applies in the absence of half-lead singles, right?
Can it be extended to one counting suitable rows in a half-lead, from
lead heads and lead ends, with some suitable constraint on how those
lead heads and lead ends match up or something?



-- 
Don Morrison <dfm at ringing.org>
"Spam is bad. The amazing degree of unanimity that greets such a
simple declaration is, paradoxically, the biggest impediment to
progress in anti-spam standards."
  -- Nathaniel Borenstein, presentation NIST/FCC, 2004.