I once saw a Vision Engineering booth at a trade show and I have to admit I did fall in love.
Later on , I made a deal on ebay and received two Vision engineering scopes. Earlier models than the Mantis which is still current.
I discovered that the way these work is by projecting into space a HUGE exit pupil. This does make them easier to use for people that wear glasses or who don't like to look into eyepieces. They also are very unusual and it is the unusual optical experience which causes them to be attractive.
However, having read PierreH 's posts I think that he should remove the Mantis from his list of toys that he will ask Santa Claus for.
I think these instruments are for electronics manufacturing factories, not for perfectionist microscopist.
They need so much light that they had models with TWO 270 watt metal halide illuminators. To get a good image on the stereo versions you need to put so much light onto the subject that there has to be a grey plexiglass "sunglass" shield to protect your eyes from being dazzled. (Similar to the orange shield on a fluorescence microscope) The stereo model I had had FOUR 75 watt quartz halogen lamphouses. Two for the eight point ring light and two with large diameter spot lenses.
I sold my Stereo Dynascope to a guy who is an electronics engineer. I had no trouble fusing the stereo images but he did initially until we discovered that there were to beam paths and the interpupillary distance is adjusted by a screwdriver adjustment, not labeled on the outside. Then he was very happy with it.
Even though for me it was not a useful purchase, I had another opportunity and I could not stop myself. They have a 30mm pair of eyepieces which works on the same principle. These oculars are called ISIS. They will fit any standard stereo and are a more sensible purchase (in my opinion ) than the Mantis.
http://www.visioneng.com/isis-microscop ... erview.php
I still have one of the two scopes which is a Vision Engineering head on a compound incident light microscope. It has a set of Nikon MPlan objectives and both incident and transmitted light. I find it more satisfactory than I did the stereoscopic version.