[bojan]

Tony,
The problem (apart from diffraction, which can't be avoided - it is the same thing as placing the fishnet in front of of the lens - no matter how thin it is, it WILL create diffraction patterns and destroy the resolution) is also in colour response - LCD screen "generates" colours by appropriate dosage of individual primers (RGB) to create the *perception* of colour in our brains.
This perception has nothing to do with colour index, because the same or very similar perception can be achieved with different primers... so we have no standard colorimetry here.

Transmission is another problem - it is generally very low for most LCD's (bright back light takes care of this.. but in astronomy we have to try to catch every photon available whenever possible).

Spectroscopic analyser is even less feasible - because spectroscope does not deal with colours (which are our physiological response to stimuli and very often the totally different stimulus will result in the same or very similar response)... It deals with wavelengths, which are physical properties of light (EM radiation)
Filters used in LCD are very wide, passing comparatively huge bandwidth, while spectroscope bandwidth resolution is and must be very narrow - couple on nm and less.
Any Wikipedia article on the subject will clarify those points for you.
This idea simply won't work.

So, as colour (astro)photography is concerned, stick to the standard colour sensors with Bayer matrix - the principle is the same as this LCD screen idea, but it has no diffraction problems (because filter is in front of the sensor), transmission is high, it's been done already by both industry and individuals and it's available off the shelf and it is known how well/bad it works.
For any serious and scientifically usable work, the combination of b/w CCD and standard colour wheel is un-avoidable.
For spectroscopy, the diffraction grating and/or prism are the only practical/economical ways (at the moment.. maybe in the future there will be optically active materials available to do the job) to achieve usable/required resolution, while keeping absorption (photon efficiency) at acceptable levels.