Characterizing a class of simply presented modules by relation arrays (Q1293303)

Let \({\mathfrak H}\) be the class of \(R\)-modules \(M\) over a discrete valuation domain \(R\) with torsion submodule \(tM\) a direct sum of cyclics and \(M/tM\) divisible. The authors study simply presented and more generally Warfield modules belonging to this class \({\mathfrak H}\). The first class can be identified with direct sums of modules of a certain subclass \({\mathfrak H}^1\subset{\mathfrak H}\) where \(M/tM\) has rank 1. This is based on a theorem which is also of independent interest: A non-split module \(M\in{\mathfrak H}^1\) can be dccomposed into \(G\oplus T\) where \(tM=tG\oplus T\) and \(G\) can be `controlled' by an indicator of \(M\). (Indicators are natural generalizations of characteristics of rational groups, known for a long time.) The reader interested in this useful characterization should check what I called `controlled' from the paper. The result naturally extends to Warfield modules in \({\mathfrak H}\). Examples and `realization theorems' like the old Zippin theorem (also shown by Ulm) clearly indicate the complexity of \({\mathfrak H}^1\) which still can be classified.