P&I are terms of an control loop called PI(D) Controller google that.
It is not so hard to understand. In a control loop you have a current state and a state you want the system to be ("set point"). So that difference describes an error. The P term is the proportional response to that error and is basically an multiplicator to the actual error. A high P will produce a strong and fast correction but an overshoot as well. A too low P can lead to slow response or never hitting the actual set point ("undershoot"), that's why the idea of an "I" was born. The I helps the "P" out of its undershooting misery and so indirectly helps the overshooting you would produce by an too high P to actually reach and surpass the setpoint. The I accumulates the error overtime. So in that undershooting situation the P is too low to reach the wanted setpoint because the error is constantly too low BUT with the "I" the constant little error accumulates and "lifts" the result to the setpoint (and generally increases reactiontime). This error collecting "I" therfor also helps when holding an angle in acromode because it increases the reaction(and speed) to small changes from outside (wind) or inside (pilot). (NOTE: That is a rough description of "I" because it can be also negative... but it is easier to understand it in the "positive" way..)
So P & I work hand in hand. Too high values for "I" or "P" will produce overshoot and wobble. When dailing in an "I" value you will normally have to decrease the "P" value.
Play with those values and you will see in flight their meanings.
There must be some guides online concerning the KK boards (I don't have one).
The limits describe the maximal influence of each component on the result (tiltangle).
The "gain" describes the multiplicator of the error ("P") or the accumulated error over time ("I").
You can not fly with a "P" of zero but with an "I" of zero.
Generally you increase the P till a constant oscillation appears, than back off that value a good portion (not more than the half of that oscillationvalue, some theory suggests to take 60% of the oscillating "P" as the baseline "P" for further tuning - but from my experience with other FC that theoretical value is too small.), so you have headroom for your "I". Now dail in the "I". If the copter becomes oscillating again or it feels bad on the sticks, it was too much
.
Cheers Rob