a) Let F be a nonzero element of Z((T)) such that there exists r∈Z, with (1−T)rF∈Z[T,T−1]; then F can be written uniquely in the form F=(1−T)−d.Q, where Q∈Z[T,T−1], Q(1)=0 and d∈Z. If F≥0, then one has Q(1)>0 and d≥0.
b) Let Q,R be in Z[T,T−1], d,d′ in Z with Q(1)>0. If
(1−T)−d.Q≤(1−T)−d′.R,
then either d<d′, or else d=d′ and Q(1)≤R(1).
a) We may write F=(1−T)−rTnP(T) with r,n∈Z and P(T)∈Z[T]. By euclidean division, we may write P(T)=(1−T)pR(T) with R(T)∈Z[T] and R(1)=0. Hence F=(1−T)−(r−p)Q(T), where Q(T)=TnR(T)∈Z[T,T−1] and Q(1)=0. This proves the existence of d and Q. Moreover, if (1−T)rQ(T)=(1−T)sR(T) with r>s and Q,R in Z[T,T−1], we have R(T)=(1−T)r−sQ(T), hence R(1)=0; this proves the uniqueness. Suppose that F is ≥0; if we had d<0, then we should have F(1)=0, which is impossible since F is nonzero and all its coefficients are positive; thus we have d≥0. If d=0, then Q=F≥0, hence Q(1) is positive. If d≥1, then Q(1) is positive by formula (3). This proves a).
b) Suppose d≥d′. Then (1−T)−d((1−T)d−d′R−Q)≥0; since S(T)=(1−T)d−d′R−Q belong to Z[T,T−1], this implies S(1)≥0 by what precedes. If d>d′, we have S(1)=−Q(1)<0, whence a contradiction; if d=d′, we have S(1)=R(1)−Q(1), whence Q(1)≤R(1).
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.
Nothing.