[r-t] Yorkshire Surprise Minor, etc

[Iain Anderson]

Mark

I liked your email.  Well balanced.  I have one question about this comment:

"In summary, I think there is a balance to be struck between two much 
classification (harms extension, applies bias to method types) and too 
little (not enough namespaces)."

At the current rate, when do you think we will run out of method names?

IJA

On 2017-03-19 10:59, Mark Davies wrote:
> I guess the history of it is that method classification is a just a 
> way of providing a tag which alerts people when a method falls into a 
> particularly popular category. This is why we've ended up with more 
> fine-grained classifications for the more popular types of methods 
> (despite this, there are more methods in these categories than the 
> wider, unpopular ones). Often we don't (or didn't) even have a name 
> for the general case.
>
> From some perspectives, this is a strange way of doing things, and 
> downright ugly if you're one of those people who thinks popular 
> methods should be less popular. However, it can of course be very 
> useful at times, such as when your village six augments to eight, you 
> want to christen the new ring with a peal, and you discover some other 
> band have already named Lower Snotscommon Surprise Major. But Lower 
> Snotscommon Delight is free!
>
> When it comes to extension, classification once again both helps and 
> hinders. Where there are two reasonable-looking extensions, and one 
> falls into the same class as the parent and the other doesn't, well 
> perhaps there's a good argument that you should go with the first one, 
> because there's a shared structural property. In the case of 
> "Yorkshire D Minor" if you were to try and extend it to Major, 
> wouldn't you want to keep the cross in the 3-4 section, and the 
> external places in the 4-5 section? Why would an extension introduce 
> places in every cross-section? I haven't looked, but it seems likely 
> there are false Major methods which look more similar to the Minor 
> than Y8 does. Much as we'd like a Yorkshire on six, does it really exist?
>
> But often the only reasonable-looking extension doesn't share the same 
> class. Sometimes, to my mind, this means that the classification is 
> just plain wrong. For instance, currently where a Major method extends 
> to a short-course Royal method, the latter is classed as 
> "Differential" and so is not accepted as an extension. Unlike the 
> Yorkshire case, this is a problem which may prevent an otherwise 
> perfect-looking extension applying on an infinite number of stages. 
> Why is the number of leads to the course deemed to be such an 
> important structural property? The leads still fall into the same 
> group (the PB leadheads).
>
> In summary, I think there is a balance to be struck between two much 
> classification (harms extension, applies bias to method types) and too 
> little (not enough namespaces). To my mind, the best classification 
> system is one that works well and helps out method extension as far as 
> possible. The worst failing of the current system is not 
> TD/Delight/Surprise, but the separation of short-course methods into a 
> separate class.
>
> MBD