Hello, and welcome to the forums!
I can see that this could possibly be handy in certain mission-critical applications, it's something which would likely see limited appeal though, especially given that several lamp technologies to behave in such a way anyway. Also, most mission-critical applications will generally have at least one backup lamp there in reserve for the very reason that lamp failures can be unpredictable.
Fluorescent lamps (CF's failing due to a duff ballast excepted) will tend to dim at least somewhat visibly in their last couple of hundred hours, and blackening at the tube ends tends to give the game away to an extent too, they have caught me off guard a couple of times though.
Metal halide lamps often show a *very* distinct colour shift if operated beyond their rated lifetime (generally an extremely rapid increase in colour temperature) prior to actual lamp failure. Obviously (especially in compact designs) this is an ill advised proceedure anyway, due to the potential for the arc tube to rupture under such circumstances. If you're using a properly designed fixture, that won't cause any danger...but you'll still have a few zillion bits of finely atomised quartz to clean up!
Mercury vapour lamps don't tend to show quite such a huge shift in CCT, but if the lamps are run past the rated end of their life (which often they will - mercury vapour lamps are well known for the fact that they'll often just keep on going!), the output will tend to drop very visibly after a time before the lamp actually fails.
Which sort of technology were you intending this to be applied to? The only type of technolgy which really gives no indication of impending lamp failure whatsoever would be incandescent lamps. The drawback of introducing new technology here, is that one of the virtues of incandescents is that they're so cheap - complicating them would obviously increase the price - rendering that value pretty much invalid. Hence, I can't see a huge market for it - I'm no expert though, so could well be wrong (would be far from the first time! - don't let me put you off the idea!). I would be most curious to see what this idea of yours is though, purely from an engineering point of view if nothing else.
