I am independently studying valuation rings and I am currently trying to prove that a DVR is a Euclidean domain using the definition here http://abstract.pugetsound.edu/aata/section-factorization-domains.html.
Here is what I have:
Let $V = \{a \in F| \nu(a)\geq 0\}$ where $\nu: F\rightarrow \mathbb{Z}\cup \{\infty\}$ is a discrete valuation. Then
1.) For all nonzero $a,b \in V$ $\nu(a)\leq \nu(ab)$ since $ab \in \left <a \right >$ (this is a theorem).
2.) Let $a,b \in V$ with $b\neq 0$. Then if $ab^{-1} \in V$, we can write $a =b(ab^{-1}) +0$. If $ab^{-1} \not\in V$, then $\nu(b) > \nu(a)$. Now we can write $\nu(a) = \nu(b +(a-b)) \geq \min\{\nu(b), \nu(a-b)\}.$
I would like to be able to conclude that $\nu(a) \geq \nu(a-b)$ so that I can write $a = 1(b) + (a-b)$. It seems that this follows from $\nu(b) > \nu(a)$, but I am not completely confident that this is the case.
Thus, my question is: does $\nu(b) > \nu(a) \Rightarrow \nu(a) \geq \nu(a-b)$? If so, why? Or is my approach flawed?
The definition of the link refers to a Euclidean valuation, while a a discrete valuation on $F$ is a surjective function satisfying $$ \nu(ab)=\nu(a)+\nu(b),\;\nu(a+b)\ge \min\{\nu(a),\nu(b)\}. $$ The DVRfor $\nu$ then is the ring $R=\{0\}\cup \{x\in R\mid \nu(r)\ge 0\}$. It satisfies the Euclidean valuation property (is a Euclidean ring), because we have $\nu(a/b)=\nu(a)-\nu(b)$, see here, page $3$.