Coax length.

The great debate.
There appears to be a great debate about whether coax length is important. Some swear by set lengths, while others swear at them.
The short answer to the question is that, as long as you set up the antenna correctly, coax length is not important. It only becomes an issue if there is a problem with the antenna or coax. If trimming your coax appears to change the SWR, then you have a problem that needs fixing.
However, if you have come this far through this section, having read the other pages, then I guess you may want to know why this happens.

Previous notes.
If you have read the other pages, you may remember that coax impedance will only match the characteristic impedance in certain conditions: Also, unless one of these conditions is met, the coaxial will act as a tuned circuit, with values that depend upon the length.

Length.
As long as the source and load match, one of the above conditions applies. The coax can do it's job as mere transportation and, more importantly for this issue, length will not matter.
If the load does not match the source, then the coaxial becomes like a tuned circuit, with reactances and impedance changing along the length. Different lengths of coaxial will now offer different impedance values, changing the total impedance seen at the source end of the cable.

Cutting coax.
By adjusting the length of the coax, you can make the reactance cancel out the antenna reactance - showing a 1:1 SWR at the transmitter end. However, to get the correct reactance, you may end up cutting large amounts off, ending up with it too short to reach between the transmitter and the antenna.

Hiding the painful truth.
It must be noted that nothing you do with the coax can tune the antenna. The best you can hope for is to hide the problem from anything placed at the source end of the coax.
As well as hiding the problem from the SWR meter, you may also be able to hide it from the transmitter - avoiding damage. I say may be able to... it is never good to rely on such tricks, you are better off tuning the antenna correctly.

Half wavelengths.
People often talk of requiring a multiple of half wavelength of coax. This is not needed, but it can help you to get more accurate readings. Not totally accurate (see below), but possibly more accurate than before.
Multiples of half wavelength satisfy condition 3 (above). The impedance at the source end should be the same as at the load end, so any SWR reading taken at the transmitter end should now show the match between the transmitter and load.
Unfortunately, due to coaxial losses, any SWR reading at the transmitter end can never be totally accurate (see below).

Velocity factor.
It is important to remember that a half wavelength of coaxial cable is not the same as a half wavelength in free space. After working out the half wavelength for the frequency used, you must multiply this by the velocity factor of the coax you are using.
Most coaxial used for CB installations will have a velocity factor of around 0.66.

Loss.
Any loss in the coax is doubled for reflected signals, as reflected signals have made two trips along the coax. When you measure SWR at the transmitter end, you are reading the full power out without any loss and reflected signal after two trips along the coax.
Losses in the coax can increase when there is a mismatch, so the problem can be even worse that due to the specified loss of the coaxial in use.

Accurate readings.
The only real way to get a true reading is to measure at the load end of the coax. The meter will then be measuring both forward and reflected after the same amount of loss in the coax. Also, any effects of the coaxial impedance will not show at that end.