Unify nomenclature: splitting methods vs. homomorphism method vs. ...

[fingolfin]

The newly added HOWTO talks about "leaf methods" and "splitting methods". Yet in chapter 3, we just talk about "homomorphism" (and sometimes this may include "leaf methods", sometimes not). In the paper "A data structure for a uniform approach to computations with finite groups" by Ákos Seress and Max Neunhöffer, these are also called "non-leaf methods/nodes", and also implicitly "reduction method" is used. In "A practical model for computation with matrix groups" they talk about "reduction epimorphisms" and "non-leaf nodes". Other sources may also use other names

In recog itself, often we only talk about "homomorphisms", which produce either leaf nodes or non-leaf nodes. And indeed, strictly speaking, there is no distinction a-priori between methods/homomorphisms/... producing either kind of of nodes -- indeed, it would be possible to write a method which sometimes produces a leaf node and sometimes a non-leaf node. And while this is right now never the case, AFAIK (I might be wrong, though!), I think we should be careful to make it clear in the manual (including the HOWTO) that this is not a strict separation.

Indeed, methods/morphisms correspond to edges, not nodes... A "lead method" then is one where the image is a leaf. Note that there still can be a kernel!

I am not yet sure to what exactly this all should amount, but I certainly believe that we should revise everything, esp. the new HOWTO. Since no other source uses the terminology "splitting method", I'd change it to "reduction method", and/or "non-leaf method"