Actually, the question of absolute phase is a very big consideration in recording and mastering.
First, every professional microphone specifies its wiring polarity referred to sound-pressure polarity - this is crucial for any kind of consistency in application of microphone techniques. Also, keep in mind that a large percentage of microphone models are used in both a recording and a sound-reinforcement context, and in the latter case it's extremely important to keep track of absolute phase as the sound of the instruments (or instrument amps) themselves interact directly with the front-of-house and monitor loudspeakers. (Anybody who's worked with older JBL stuff should be familiar with these phasing issues . . . the driver labelling is reversed.)
Second, in either the live or studio context a great number of sound sources and equipment processing loops is ultimately mixed together, at least the relative phase inarguably crtitical. So in practice, the correct connection of all equipment, wiring, and patchbays (observing individual TRS and XLR pinouts, etc.) is a cornerstone of good workmanship in professional practice.
So unless somebody's made a mistake, absolute phase should indeed be preserved all the way through the studio recording chain, and also through the chain at the mastering studio, especially if it's digital. For record lathes, I know that the Neumann and Ortofon cutting amplifiers are very clearly specified as far as their absolute input phase, and of course the cutting-head MUST be properly phased to the amplifier or it will oscillate and destroy itself.
Now there are generally three places in the recording chain where phase can be deliberately manipulated - when tracking, on mixdown, or in mastering . . . through the use of a phase-inversion switch on the mic preamp, console channel strip, or mastering console. In practice, all of the switches start out "non-inverted", and the phase of a particular channel/microphone is inverted only when necessary for specific interactions . . . for example, when two mics are used for the top and bottom of a snare drum. Inverting a single microphone during tracking is generally frowned upon; keeping the audio as un-molested as possible until mixdown is the usual goal. For the overall absolute phase, the mastering engineer usually makes a final decision.
As for Al's question, most of the guys I've met who record on-location with minimal microphone techniques pay very close attention to absolute phase, especially with M/S and Decca Tree configurations. If there are any distant "hall" omnis, they'll usually phase these to preference, while keeping the primary mics uninverted. They also almost always track directly to digital, making it easy to keep everything the same through mastering.
Now after all this blathering, I agree with Atmasphere as to the general utility of an absolute-phase switch in a home reproduction context, if maybe for different reasons. I feel that there are very few reasons not to design equipment or wire a system so that absolute phase is maintained. But how much it matters is one of those old, unsolveable audio debates.
The absolute phase of any loudspeaker through the midrange and treble can be thought of as arbitary, as connection polarity, driver response, and crossover designs vary considerably. And in lower frequencies, the interaction between the cabinet/driver/port tuning and room trumps everything else. Maybe for headphone listening there's some validity for phase absolutism, but through loudspeakers, it's a "flip to your taste" kind of thing, if you care to take the time . . . which I personally don't.
