Unfortunately, the groups obtained for already seem to be much harder.
While the results for
are still very good compared to Felsch or HLT
style methods it was not yet possible to compute
.
Here, a tuned version has to be used which uses less space.
However, the results for
show some similarities to the results
obtained for
.
The strategies NONE and I-R perform best for most orderings used followed by
I-ALL which is even the best for the syllable ordering syl-l-BbAa.
But I-ALL is considerably worse for the length-lexicographical orderings.
All other strategies perform even worse producing the best results for the
syllable orderings.
The best combination was I-ALL/syl-l-BAab which was not used for
.
Here, further investigations are necessary.
It should nevertheless be noted that
, and
are
all isomorphic to the generalized quaternion group of order 16 and that the
coset enumeration for the latter two is almost trivial!