I would like to raise an objection to using the term `2-group' as on nlab and elsehere since for the group theorists this has a specialised meaning: See the following wiki entry, especially the first 2 words: "In mathematics, given a prime number p, a p-group is a periodic group in which each element has a power of p as its order. That is, for each element g of the group, there exists a nonnegative integer n such that g to the power pn is equal to the identity element. Such groups are also called primary." I feel we should try to avoid and even to reduce confusion, especially as there are claims that crossed modules, for example, can be thought of as `2-dimensional groups' (I agree with this, of course!); there are nice crossed modules M \to P in which M and P are 2-groups in the group theoretic sense! My favourite example is \mu: Z_2 \times Z_2 \to Z_4 in which Z_4 acts by the twist (of order 2), and \mu maps each factor Z_2 injectively into Z_4. This crossed module has non trivial k-invariant. I think Johannes Huebschmann first observed this. So an example oriented approach to crossed modules could well need the term p-group in its standard group theoretic usage. Some examples of finite crossed modules are in R. Brown and C.D. Wensley, `Computation and homotopical applications of induced crossed modules', J. Symbolic Computation 35 (2003) 59-72. However I think one can be happy with the well established term 2-groupoid. I would just like this point to be discussed: terminology is important, and confusing an established use might raise hackles unnecessarily. Ronnie