One of the big plusses of MathCAD is its ability to predict the units and dimensions of results, and to detect unit and dimension errors that are within the SI system, and the additional special mathcad currency dimension.
The SI system still has many dimensions that are declared as simply "Unity" "1" (that's in bold), but as far as the dimensional consistency checks go they have no value at all. There are many places where users would benefit from the provision of spare dimension to allow the detection of scaling and kind errors within their analysis.
In particular the radian (1) and steradian (1) have a long history of requiring some mechanism for spotting errors in calculations, such as distinguishing between Torque (Nm) and Work (Nm).
In terms of useful extra dimensions there are at least three different views on radians, such as being the mechanical spatial angles where the single 3d dimension of "length" has had a cancellation Lx/Ly from different directions (which is undetected); there is the phasor angle where Fourier repetition phase angle along a line (e.g. Time) is being mapped to a circular aspect; and there is the mathematicians trig function angles in radians where the mathematicians deny the existence of named dimensions anyway.
A Dimension symbol can also be used to indicate some specific reference such as an 'input voltage' measure, rather than an output voltage, so that results are not confused.
There are many mechanical dimensionless numbers (see Buckingham Pi theorem http://en.wikipedia.org/wiki/Buckingham_%CF%80_theorem) such as Mach number, Reynolds Number, and others http://en.wikipedia.org/wiki/Dimensionless_quantity.
By allowing users to attach a special user dimension to these numbers, where they feel appropriate, they can spot many calculation errors that are otherwise almost impossible to detect.
One of my regular performance assessment sheets for a camera system performance (fixed size target with a bar pattern in cycles, as percieved at a distance) has results in Cycles per milliradian, which is totally unitless in the current SI system, no matter how hard one tries, so many scaling errors are easily missed.
Mathcad should provide an additional set of User dimensions for the user to apply as required.
