I'd do it as follows (which is, basically, one of the most usual proofs of the CRT): we have the coprime (primes, in fact) integers $5,13,17$:
$$(1)\;\;\;\text{solve}\;\; 13\cdot 17y_5=1\pmod 5:\;\;\;\color{red}5\cdot(-44)+\color{red}{13\cdot 17}\cdot1=1\;\;\;\implies \color{blue}{y_5=1}$$
$$(2)\;\;\;\;\text{solve}\;\; 5\cdot 17y_{13}=1\pmod {13}\;\;\;\color{red}{13}\cdot (-13)+\color{red}{5\cdot17}\cdot2=1\;\;\;\implies \color{blue}{y_{13}=2}$$
$$(3)\;\;\;\;\;\,\text{solve}\;\; 5\cdot 13y_{17}=1\pmod {17}\;\;\;\color{red}{17}\cdot23+\color{red}{5\cdot13}\cdot(-6)=1\;\;\;\implies \color{blue}{y_{17}=-6}$$
and now define
$$x:=\color{green}2\cdot\color{blue}1\cdot13\cdot17+\color{green}1\cdot\color{blue}2\cdot5\cdot17+\color{green}5\cdot\color{blue}{(-6)}\cdot5\cdot13=-1338=-233=872\pmod{5\cdot13\cdot17}$$
You could, of course, have chosen $\,11\,$ instead of $\,-6\,$ in (3) above, but you'd have obtained a much bigger solution, $\,4187\pmod{5\cdot13\cdot17}\,$ , so a negative solution here and there isn't usually a problem but, in fact, an advantage.
The biggest pro in the above method, in my opinion, is that it serves you in any case, whether you have 2, 3 or more congruencies to solve, and is very mechanical.
Equation 2: $x = 3 + 4 × s$ (mod 4)// Equation 3: $x = 1 + 5 × u$ (mod 5)// Solving the 3 equations simultaneously, we get $x = 11 + 60k$, where $k\in\mathbb{Z}$
– Poseidonium Apr 16 '13 at 10:08