I was reading a solution of Exercise Ch I. 2.11 b) from Hartshorne's book that uses the following fact: if $S:=K[x_{0},\ldots,x_{n}]$ and $f_{1},\ldots,f_{m}\in S$ are linear polynomials then $\dim S/\langle f_{1},\ldots,f_{m}\rangle=n+1-m$. The proof begins with
"Since $S/\langle f_{1},\ldots,f_{m}\rangle\approx (S/\langle f_{1},\ldots,f_{m-1}\rangle) \, / \, \overline{\langle f_{m}\rangle}$ it suffices to prove the case when there is a single $f_{i}$, say $f$."
From this one constructs a homomorphism $S\to K[x_{1},\ldots,x_{n}]$ using $f$ in order to conclude that $S/\langle f\rangle \approx K[x_{1},\ldots,x_{n}]$.
I have some questions:
Shouldn't it be $S/\langle f_{1},\ldots,f_{m}\rangle\approx (S/\langle f_{1},\ldots,f_{m-1}\rangle) \, / \, (\langle f_{1},\ldots,f_{m}\rangle/\langle f_{1},\ldots,f_{m-1}\rangle)= S/\langle f_{1},\ldots,f_{m-1}\rangle/\langle \, \overline{f_{m}} \, \rangle$ ?
From this isomorphism, how does one conclude the induction step? I guess it would be an application of Theorem Ch I. 1.8A b) from Hartshorne's book, namely, $\mathrm{ht(p)}+\dim B/p=\dim B$ for a prime ideal $p$ and an integral domain which is a f.g. $k$-algebra. But do such hypotheses actually hold?
Could this inductive process work to show that $\langle f_{1},\ldots,f_{m}\rangle$ is prime whenever the $f_{i}$'s are linear polynomials. If so, how? I see no reason as to why $\langle \, \overline{f_{m}} \, \rangle$ should be prime.