[r-t] Jump change notation

[Don Morrison]

On Tue, May 19, 2015 at 10:11 PM, Don Morrison <dfm at ringing.org> wrote:
> With that emendation, that interior places that are being made must
> always be specified explicitly just as with ordinary place notation, I
> *think* I see an algorithm for turning such extended place notation
> into a permutation. It is rather more complex than I'd like, but I
> think it works.
>
> Given how helpful Ander's recent comments have been, I'm afraid I'm
> going to bother you all again, and ask anyone interested to pass their
> eyes over this and tell me if it seems right to you. It's complicated
> enough it seems unlikely I haven't gotten something wrong. And if
> anyone sees an easier way to do this, please do speak up!

While there appears to be little interest in this topic, so that it's
archived here for future reference here's the dénouement, anyway.

I'm now reasonably confident that this algorithm works, as I've implemented
and tested it. Not the same as proving it correct, of course, but it does
raise my level of confidence nonetheless.

In addition to parsing such notation, I've written some code to write it
out, and the heart of my testing has involved doing the round trip: start
with a change, write it out, read it back in again, and confirm that it's
the same. Of course, offsetting bugs in the two operations could....

Anyway, I've tested the round trip on the extents from 2 bells up to royal,
and on 10,000 randomly (well, pseudo-randomly, Mersenne Twister) generated
permutations at each of the stages from cinques up to twenty-four, and all
seems well.

For testing the flip side, whether it erroneously results in an incorrect
change from defective input I've not been as methodical; I've simply tried
a bunch of bogus input that I pulled out of my head and confirmed that in
all cases it correctly signals an error.

My revised place notation parser actually follows a suggestion made by
several folks in the recent, inconclusive poll: it supports two different
schemes for notating jump changes, this Holroyd notation, and simply
writing the permutation like a row, between square brackets. This parser is
currently irrelevant to anyone but me, but if all goes according to plan
one day it will show up in the grid/blueline thingie on ringing.org.



-- 
Don Morrison <dfm at ringing.org>
"We have so little leisure now that we have invented so
many labor-saving devices"
      -- Will and Ariel Durant, _The Age of Louis XIV_
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://bellringers.net/pipermail/ringing-theory/attachments/20150527/fc3f092e/attachment-0004.html>