[r-t] Method extension

[Andrew Johnson]

"ringing-theory" <ringing-theory-bounces at bellringers.net> wrote on 
28/04/2015 04:33:55:
> From: "Philip Earis" <pje24 at cantab.net>
> PABS:
> "What limitations? Can you give an example of something that ought to be 

> considered an extension that is not covered, and does not introduce some 

> obvious feature that is not present at the lower stage?"
> 
> I am not disputing that your algorithm can generate methods at different 

> stages with close similarities.  I agree that such an algorithm has use 
in 
> an advisory capacity if someone wants to generate methods.
> 
> But it is clear that no algorithm can (or should) produce a 1:1 mapping 
for 
> (method at stage a : method at stage b).  You acknowledge this yourself, 
eg 
> with Cambridge Minor possibly extending to Yorkshire, Lincolnshire, etc.
> 
> This, to my mind, makes the case for mandating the use of an extension 
> algorithm very weak.
> 
> Why?  Well, say that you start with Oxford TB minor, and no 8-bell 
methods 
> have yet been rung.  A band wishing to use your algorithm to extend this 
to 
> 8 could (with good reason) end up with the method known as Barry 
Peachey's 
> Dog. A separate band, starting with Kent TB Minor, might chose to extend 
to 
> what we know as Kent Major. Before long there would be a clamour to 
> hard-code some arbitrary preferences (like the length of the lead must 
> remain constant, or must extend with stage, or whatever) into the 
extension 
> rules.  And before long you'd end up with a heavily parameterised mess, 
like 
> the current Decisions, that would continually introduce new loopholes 
and 
> problems (like with extending the method where the treble plain-hunts to 

> n-1th place).
> 
> You will also end up under fire from the ringer in the street when you 
say 
> that the method known as London actually shouldn't be as it no longer 
forms 
> a valid extension, and myriad other examples of historical baggage now 
and 
> in the future.
> 
> I don't really get Richard Johnston's point that a more liberal approach 

> would lead to "poor extensions".  There's no shortage of terrible 
extensions 
> at the moment that comply with the current Decisions, nor indeed or 
terrible 
> new methods rung.  Why is this an argument not to trust bands?  I agree 
that 
> a liberal approach could result in terrible methods being rung, but to 
my 
> mind the status quo often mandates terrible methods being named thus. It 

> hardly seems like a better situation.
What should happen if a band rang Double Bob and Reverse Bob Fourteen, but 
using the opposite place notation? Should the CCCBR let that stand as the 
band is to be trusted? What about not giving an extension the same name as 
a parent?

As ever the problem seems to be that some things obviously fit a pattern 
and are easy to categorise, some obviously do not and never should be 
counted as being similar, and there is a grey area between.

If you don't have a list of desirable and undesirable characteristics of 
extensions and some sort of consensus as to what is important then giving 
two methods the same name in different stages has no significance and you 
might as well prohibit two methods at different stages with the same name. 
That must be the simplest extension rule!

We aren't starting from a clean slate, so we need to examine what we have 
so far.
Most ringers would understand the current extensions for
Plain Bob
Grandsire

Cambridge
Yorkshire

Double Court/Double Norwich

Kent TB
Oxford TB
St Clement's College Bob

Little Bob
Stedman
Erin

More complex
Superlative Surprise
Bristol Surprise

How many extensions don't fit the simple patterns above?

Extensions that don't quite work, but why?
London No.1,2,3 Surprise Royal
Superlative No.1,2 Surprise Royal

Here are some thoughts on characteristics - others should add to or argue 
about these..
I don't think a right place method should extend to a wrong place method - 
though that is tricky for extensions by one stage as technically odd 
numbered stages have wrong places.
Similarly a treble bob method with symmetric sections shouldn't extend to 
one with asymmetric sections - though perhaps this causes trouble for 
double methods.
Conventional reflective symmetry should be maintained.
Should double methods stay double? I guess they should to maintain the 
name 'double' otherwise it gets very confusing!
Should reverse methods extend in the same way as the base method, so the 
'reverse' qualifier still applies?
If the method has already been named at more than one stage then any 
extension should fit all of the already named methods?
An extension that works on more closely spaced stages is to be preferred 
e.g. 8,10,12,14,.. rather than 8,14,20,...

Retaining plain bob lead ends seems desirable.
Maintaining truth - a true method shouldn't extend to a false method.
Maintaining same number of hunt bells or not? E.g. Double Norwich Royal.
Lead-end order - is that important, or desirable?
What is the reason PB minor should not extend to 
x18x18x18x18 LE 14 (bob course of PB major)
or Gainsborough?
x18x18x16 LE 12
Does that mean that external places are different to internal places?
There doesn't seem to be a problem with having more places in a change 
than in the parent (e.g. Cambridge minor x36x14x12x -> x38x14x1258x ) but 
why?


Andrew Johnson





Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU