Joseph A. D'Appolito is best known as the developer of the "D'Appolito Configuration" (the MTM design) which was first described in his "A Geometric Approach to Eliminating Lobing Error in Multiway Loudspeakers"[1]. This configuration is recognized by loudspeakers which have two midrange or woofer drivers arranged vertically above and below the tweeter[2]. Author of "Testing Loudspeakers" and a contributing editor to AudioXpress magazine[3]. Owner of consulting firm, Audio and Acoustics, Ltd. whose clients include Snell Acoustics; HECO/Recoton, GmbH and Usher Audio Technology. D'Appolito has electrical engineering degrees from Rensselaer Polytechnic Institute and Massachusetts Institute of Technology (MIT).
All that is nice and rather impressive, but I agree that in theory the MTM center speaker is flawed. The MTM configuration projects a controlled dispersion in the plane going through the tweeter and 90 degrees from the axis created by the woofers. It does this by the additive effect of the in phase signals in the plane equidistance from both woofers and the subtractive effect when the signals are not in phase (anywhere the signals are not equidistance from the woofers.) With this dispersion pattern the floor and ceiling reflections (effects) from the woofers are reduced. In theory, at some distance/frequency the signals from the woofers blend and appear to come from the same position in space as the tweeter, thus providing a point source for both low and high frequencies. Personally I'm not a fan of MTM as it (in theory and from my limited experience) lessens any vertical imaging you might get, the blending doesn't occur, and the types of dispersion pattern from woofer and tweeter don't match (unlike what happens in the "real world").
So in a normal sideways application of a center channel, the theory says that more of the sound goes along a vertical plane straight out the center of the speaker. I'm not into HT (heck I bought a small MTM for center channel use and are doing exactly as you suggest

) but the THX acoustical standards for HT specifies controlling vertical dispersion (the exact opposite of the horizontal MTM).
The biggest problem with what you propose (again in theory) is that vibrations from the center channel will transmit into the DVD player (and also the rest of the electronics) which could induce microphonic vibrational feedback issues. Another issue is the excessive bass reinforcement from partially enclosing the center channel versus providing "free space" as standmounting would provide.
All this aside, HT is a low priority for me so I've not noticed any of these issues being a problem.
Note that two woofers will go louder/deeper than one and that in many (most?) cases tweeters are more efficient that woofers, so the overall efficiency of the speaker is raised. It also allows for nominally low or high impedance woofers to be combined to produce a more typical (and amp friendlier) 8 ohm impedance load. But all this works regardless of the driver configuration.