Easy task: Formulate the theoretic proofs of isomorphism for Loewy vector \((1,k,1,\ldots,1)\) for the special case of Singer algebras, remove the conditions on characteristic and fact that alg. closed. (How do the b_i map to the x_i, y_j in the proof in the paper?)
Loewy length formula for \(n = 4\)? (If the bound is not attained then \(e = 2e_1\), and ...) Look at more examples and hope for a counterexample, or try a theoretical proof?
Is there an infinite series of parameter pairs \((q,z)\), \((q',z)\) with the same Loewy vector such that \(A[q,z]\) and \(A[q',z]\) are nonisomorphic? (The examples with \(z = 117 p\) for small primes \(p\) could look promising.)
Classification of \(|\{ s_q(ke); 1 \leq k < z \}| = 1\)? (Under this condition, we have \(s_q(ke) = n(q-1)/2\) for all \(1 \leq k < z\). Is this condition equivalent to \((q^{{n/2}}-1 \mid e\)?)
Does the bound hold for \(e\) dividing \((q^6-1)/(q-1)\)?
Is there a closed formula for the Loewy length in case \(n = 3\), analogous to the one in ... for \(n = 2\)?
How many isomorphism types of algebras \(A(q,n,e)\) can exist if the Loewy length is \(4\)?
For which values of \(z\) do we have open cases in the classification by isomorphism types? (patterns?)
Is it possible that \(A[q,k n,z]\) and \(A[q^k,n,z]\) have the same Loewy vector but are not naturally isomorphic? Note that in this case, one direction is clear: No overflow in addition mod \(q\) implies no overflow in addition mod \(q^k\) for corresponding pairs of basis vectors. For the other direction, one would need that in case of an overflow mod \(q\) in one position where there is no overflow mod \(q^k\), there is another position in the coefficient vector where one has also overflow mod \(q^k\). (Look at example \(A[9,6,65]\) vs. \(A[3,12,65]\): The only pairs of coefficients mod 9 where one has overflow mod 3 but not mod 9 are (1,2), (1,5), (2,2), (2,4), (2,5).)
Is it possible that two Singer algebras are nonisomorphic over the ring of integers, but they have isomorphic reduction modulo some prime integer?
Why does the expression "Prod(Root0(3,1),Root0(2,1))"
not describe a valid invariant? (Is there an example in the database that documents this, that is, a pair of isomorphic Singer algebras for which the corresponding formal subspace has different dimensions?)
At the end of this manual, we reset the user preference DisplayFunction
of the package to its original value.
gap> SetUserPreference( "SingerAlg", "DisplayFunction", origpref );
generated by GAPDoc2HTML