Commutative Algebra, IX, p. 50
Let J be the set of integers ⩾1. For every n∈J, let us denote by Jn the set of integers d⩾1 which divide n. Let A be a ring. We make AJ into a product ring. We define the mappings Φ,fn,vn(n∈J) of AJ into itself by the following formulae. For a=(aj)j∈J, put
Φ(a)=(Φn(a))n∈J,whereΦn(a)=d∈Jn∑dadn/d, vn(a)=(nam/n)m∈J,fn(a)=(amn)m∈J; it being understood that
am/n=0 if
m/n∈/J. Note that
Φn depends only on the
ad (
d∈Jn). We shall sometimes write
Φn((ad)d∈Jn) in place of
Φn(a).
For every prime number p, we denote by wp the p-adic valuation of Q. These notations will be retained in the exercises that follow in § 1. a) fn is an endomorphism of the ring AJ, and fn∘fm=fnm for all n,m in J. b) vn is an endomorphism of the additive group of AJ, and vn∘vm=vnm for all n,m in J. Let n,m be in J and d=pgcd(m,n) c) One has fn∘vm=d⋅vm/d∘fn/d. In particular fn∘vn=n.Id, and fn∘vm=vm∘fn if d=1. d) For all a,b in AJ one has
vn(a)⋅vm(b)=d⋅vnm/d(fm/d(a)⋅fn/d(b)) ((q)p/wf)p/u⋅(v)ua=((q)wf⋅v)uaup dnvn(fm(b))=vn(1)⋅vn/d(fm/d(b)). (The second formula follows from the first, and the third from the second by taking
a=1.) In particular, one has
vn(a)⋅vn(b)=nvn(a⋅b),
and
vn(a)⋅vm(b)=vnm(fm(a)⋅fn(b)) if
d=1.