I know where you are coming from! Understanding the theory is very easy, but applying that theory to your benefit can be a bit hit-and-miss.
Before I start, I've just had a quick google this looks like a good reference page for the stuff I don't explain well or get wrong
...
http://www.recordingeq.com/articles/321eq.html
When you visualise 'two waves cancelling', I'm fairly sure you are probably picturing two perfectly superpositioned sine waves that completely null each other. This is fair enough when trying to explain the basics of wave superposition and constructive/destructive interference, but in reality things are rarely as simple as perfect sine waves
Also, just for transparency I'll point out that phase is an offset in the time domain. Polarity (which is often slightly confused with phase) is related but not techincally the same. With a sine wave, a 180' phase shift and an inversion in polarity will have the same effect, but as things get more complicated its important to remember the difference. Judging from you talking about offsets in wavelengths I believe you already understand this though.
Right, there are a number of things to think about:
Firstly (though from reading your post I'm sure that you already understand this) there are varying 'degrees' of phase (no pun intended!)... going back to our perfect sine waves, yes they will completely cancel if in antiphase (completely out of phase, 1π radian or 180' out of phase, separated by ½λ, etc), but consider the result if they are, lets say, 90' out of phase. Waves don't always have to completely cancel - there can be varying levels of interference depending on how incoherent the waves are.
Now I'm going to try (yet again) to get away from the simple sine wave descriptions
... Next, you mention your understanding of the sound hitting one mic "half a wavelength behind the other", but remember that the wavelength is proportional to frequency (λ = v/f). If you have a complex sound covering many frequencies, the half-a-wavelength thinking goes out the window.
Really, whenever two complex audio signals are summed there is more that likely going to be some descructive as well as constructive interference. Again, this doesn't have to mean complete cancellation and it doesn't have to be audible either.
So, on to how this applies in the real world.
Looking at your drum mic example with the snare to begin with; the close mic and the overheads generally pick up a very different sound of the snare. You are not picking up anywhere near the same signal from the two different mics and you would never be able to get them to null. If the snare mic is badly out of phase with the overheads, its usually the low-mids that are affected and and the snare looses the 'punch' of its bottom end. If the signals are
that badly out of phase then its possible a wiring problem that put the polarity out to begin with and flipping the polarity would probably sort a lot of the problem. Anywhere inbetween though and you need to be looking at your mic positioning.
When setting up overheads, I usually use a coincident stereo pair (where the mic capsules are close to each other). Setting them up like this is an easy way of (almost) guaranteeing more coherency between the two signals. In an XY setupd, both mic capsules are likely to be more-or-less the same distance from each source of the kit.
As soon as you move the mics apart into a spaced pair setup, you introduce differences in distances between kit pieces and the two mics. Its not uncommon to use a piece of string to check that both overheads are at least the same distance from the snare as each other, even if not with other parts of the kit. Don't get me wrong, AB overheads can sound great, but you can face more problems.
Whilst there is no "right or wrong" time to use different mic positions, I usually use spaced-pair overheads when doing more heavy rock / metal stuff, where the overheads become more like 'cymbal mics' and most of the snare sound comes from the close mic and/or triggers. In this situation the spaced pair can give a nice wide stereo separation between each side of the kit.
However for many situations you'll want to get a lot of the snare sound from the overheads - this is where phase coherency between the overheads and the spot mic is more important.
And when positioning mics, the three-to-one-rule is a good one to follow:
This rule states that for every unit of distance away from the sound source, your mics should be at least three units apart. For instance, if your mics are six inches away from the source then they should be eighteen inches apart. If they're 1 foot from the source they should be three feet apart. This will keep you out of phase problems when close miking.
Right, I've run out of steam. I've missed loads out of that answer and not a lot of it is written well or makes sense, but its a start and I'm sure someone else can chip in
