I am trying to show the Dihedral group $D_{2n} = \langle r,s ~|~ r^n s^2 = 1, rs = sr^{-1} \rangle$ can be generated by $s$ and $sr$.
I found a relatively brief proof here, but if I am not mistaken it presupposes the following theorem:
Let $G = \langle x_1,x_2,...,x_n ~|~ R_1 \rangle$, where $R_1$ is a collection of relations on the generators $x_1,..,x_n$. If $x_i \in \langle y_1,y_2,...,y_k ~|~ R_2 \rangle$ for all $i=1,2,...,n$, then $G = \langle y_1,y_2,...,y_k ~|~ R_2 \rangle$.
Is this right? Although it seems obvious, I am having a little trouble proving it. Here is my shot at part of the proof:
If $g \in G$, then $g = f(x_1,...,x_n)$, where $f(x_1,...x_n)$ is some expression involving the generators. But the generators $x_1,...x_n$ can be written in terms of $y_1,...,y_k$. Therefore $g = f(y_1,...,y_k) \in \langle y_1,y_2,...,y_k ~|~ R_2 \rangle$.
However, this seems a little flaky and non-rigorous. So, my questions are, is this the theorem the Project Crazy Project is presupposing; and how would one prove it rigorously?
Perhaps easier and simpler:
$$r=s\cdot sr\implies \langle s,r\rangle\le\langle s,sr\rangle$$
and since ther other inclusion is trivial we're done.