[GAP Forum] "Knuth-Bendix Rewriting Confluent" hangs

[Robert Eckert]

On Mon, 3 May 2004, Derek Holt wrote:

> ... I guess KBMAG should have an
> option to essentially do nothing and to use just the existing rules for
> reduction - I have occasionally wanted to do that myself. But you can
> essentially achieve that aim by setting the maximum number of rewrite rules
> allowed to be something less than the initial number.
This sounds like it will do what I want it to, thanks.
> It might, however,
> insist on tidying up your rule set slightly - for example, the rewriting
> algorithm in KBMAG gets confused if two different rewriting rules have the
> same left hand side.
This can be avoided in this case.
> A number of orderings of words are available in KBMAG, some of which will
> result in rewriting systems which can make words longer. For example, in the
> wreath product orderings (which are the ones on which reduction to normal
> form in polycyclic groups is based) generators have levels, and one
> occurrence of a generator at a higher level will be regarded as larger (in
> the ordering) than any number of occurrencs of lower level generators.
>
> But again, I don't know whether this is the kind of thing that you
> are looking for!
Very similar.  I want everything of "high grade" pushed in one direction.