Why do you have to begin with the largest angle or side when using law of cosines Explain why you should always start with the largest angle or the largest side when using law of cosines. I don't understand why but my professor says so.
 A: The law of cosines is true no matter which sides you take as $a$, $b$, $c$.  The only thing that can "break down" is that in some cases when solving for one of the adjacent sides given the angle and the other two sides, there may be two solutions. 
A: The modern law of cosines applies to any side of any triangle, and the formula in no way breaks down.
I can only think of one scenario in which what your professor claimed makes sense. In the classical Euclidean geometry of Euclid's Elements(which was developed well before invention of trigonometric functions), the precursor of what we'd later call the "law of cosines" had to be split up into separate cases. Specifically, Proposition 12 of BookII states:

In obtuse-angled triangles the square on the side subtending the obtuse angle is greater than the squares on the sides containing the obtuse angle by twice the rectangle contained by one of the sides about the obtuse angle, namely that on which the perpendicular falls, and the straight line cut off outside by the perpendicular towards the obtuse angle.

The above theorem only applies to the obtuse angle in an obtuse triangle (which is of course always the largest). Proposition 13 then handles acute triangles. This is almost certainly the context in which your prof made the statement she did.
That or she is in idiot.
A: There is one case where I can make some sense of your prof's advice: suppose you know the three sides of the triangle, and are trying to compute the angles with the formula
$$C = \cos^{-1}\left(\frac{a^2+b^2-c^2}{2ab}\right) = \cos^{-1}(X).$$
If you have a triangle that is very acute (has an angle close to $0$) you will want to solve for the angle first with $X$ closest to zero. This is because the formula above becomes numerically unstable for $X$ close to $\pm1$, where the slope of arccosine diverges (in fact, if due to numerical error $X$ is slightly larger than 1, the calculator is likely to give an error.)
Better yet, calculate the angle using
$$C = 2\operatorname{atan2}(4A, (a+b)^2-c^2)$$
where $A$ is the triangle area, as computed using a robust version of Heron's formula.
A: The law of cosines says $$a^2=b^2+c^2-2bc\cos A$$ where $A$ is the angle opposite the side $a$. If $bc\neq 0$ we have:
$$\cos A=\frac {b^2+c^2-a^2}{2bc}$$If the triangle inequality holds we have the fraction between $-1$ and $1$, and this gives a unique angle $A$ between $0$ and $180^{\circ}$. If you know all the sides you can calculate the angles in this way in any order. The largest angle is the only one which may have a negative cosine.
If you know two sides and an angle, you get a quadratic for the third side giving two solutions. If you have $b,c,A$ then the two solutions are the positive and negative square root and only the positive one is admissible. 
If you have $a,b,A$ then there may be two, one or no real solutions for $c$. In the case of two solutions, both may be positive, or one may be negative and the other positive. In the case of no solutions, no triangle is possible.
To my way of thinking there is no ambiguity in any of this.
A: Someone above did touch on the correct answer about obtuse triangles. If you know three sides of a triangle and want to use law of cosines to begin solving it, you must make c the largest side unless you know that it is not an obtuse triangle. If you find an acute angle first in an obtuse triangle, then when you use law of sines to solve for the obtuse angle, you will find the acute angle with the matching sine value. If you were to continue solving the triangle with law of cosines, then it would not be an issue.
For example:
A triangle has sides a = 8.8, b = 12.1, and c = 19.7.
If you set up law of cosines to have:
19.7^2 = 8.8^2 + 12.1^2 - 2(8.8)(12.1)cosC
the measure of angle C will be about 140.5 degrees.
But if you set it up like this:
12.1^2 = 8.8^2 + 19.7^2 - 2(8.8)(19.7)cosB
the measure of angle B will be about 23 degrees. If you then use Law of Sines to solve for angle C:
(sin23)/12.1 = (sinC)/19.7
...it would now equal about 39.5 degrees, despite it being the same triangle, because the sine of 140.5 and 39.5 is the same.
A: Let side $a$ = $1.82$.  Let side $b$ = $2.31$.  Let side $c$ = $3.30$.  If we find angle $A$ first, using the law of cosines, we get $32.11^\circ$.  Then, if we use the law of sines to find angle $C$, we get $74.56^\circ$.  But, if we use the law of cosines to find angle $C$ first, we find that angle $C$ is $105.46^\circ$.  So, I advise my students to find the angle opposite the longest side first.
A: The key word in understanding the original question is "start".
The poster isn't running the Law of Cosines to find just one value.  His professor is referring to solving a SSS triangle.
If you "start" by finding the smallest angle (across from the shortest side) you will be creating an ambiguous situation for yourself if you try to use the Law of Sines to find your second angle.
However, if you "start" by finding the largest angle (across from the longest side), you are guaranteed to not have an ambiguous situation when you run a Law of SInes for your second angle.
If you are solving a SAS triangle, you're forced to start with the angle you have.  Once you find it's opposite side, you will be faced with 1 of 3 situations as you go to run a Law of Sines for your second angle.

*

*The new side is the largest.  You will not have an ambiguous situation.

*The new side is the middle side.  Find the smallest angle next to avoid an ambiguous situation.

*The new side is the shortest.  You will have an ambiguous situation.

