Projecting image without manual focussing I was wondering if it was possible to project a magnified image on a wall without the need of focusing, so just by dimensioning the lenses right.
I know I have to use the principal of Maxwellian View for the illumination of the slide. However, there are a lot of parameters left and I can design a system that requires a very minimal (say 1mm) displacement of the projection lens when projecting on 0.5m and, say, 2.5m. However, I was wondering if I oversee something here. I've seen that it is possible however I was also wondering if some special kind of lens is needed here.
 A: You can do it by correctly choosing the focal length of the lens(es) you're using. Is that what you mean? The drawback is that you can only use it in one specific setup. (I'm not sure if I understand correctly as this seems a bit trivial).
A: 
I was wondering if it was possible to project a magnified image on a
  wall without the need of focusing, so just by dimensioning the lenses
  right.

Of course You can!
The real problem is that You do not know the exact position of the slide 
in its holder. 
Second problem is, that it is cheaper to use a standard commercial 
projector including manual focus, as opposed to a custom made 
projector without manual focus. 
I do not know abot "principal of maxwellian wiev" , but I know why You 
shoud not rely on the method of fixed focus often used in cheap cameras:
This would be wasting a lot of light shone on the slide. 
A: The only possibility to do so is using point light source, illuminate slide directly and you'll get projected image in 'focus' at any distance without any focusing. 
The problems are obvious - point-sources cannot be very strong, especially as you might want 0.5mm size and lower (achievable on LEDs and especially lasers with aspheric optics)
A: You can build a projector that doesn't need any lens.
The idea is to deflect a laser (or 3 lasers) on a mirror with maybe 1mm diameter:
http://www.youtube.com/watch?v=Z448L94DVt4
You can buy those MEMS devices.
http://www.ipms.fraunhofer.de/content/dam/ipms/common/products/SAS/Systeme/laserprojmeasure-e.pdf
The best ones I've seen had the a two-axis mirror oscillate with its eigen frequencies. The beam traces a complex Lissajous pattern and by switching three lasers (red, green, blue) very fast, you can draw a picture.
You can easily run into problems with speckles, though.
