Your answer enabled me to find the better solution:
HVNorm := A -> LinearAlgebra:-Norm(A[() .. -2], 2, conjugate = false)
HVNorm stands for Homogenious Vector Norm.
BTW The intent of the extra dimension is that Homogenious transforms have an extra degree of freedom so all the variables can be scaled with the transform remaining invarient. By normalizing the last row of the Metrix to 0 .. 0 1 and the vector to 1, all division is removed from all the calculations except for division by 1. This means no division by zero or losing percision because of scaling problems brought on by a string of transform multiplications. Everything is automatically scaled so that results in the the denominator before and after all calclulations is always 1.... you can build a whole CAD system and never even use division per se.
---end of "aside"
I should not have mixed metafores. I know bettter.
VectorCalculus and LinearAlgebra have things in common, but they have their own courses, and this will always be the case unless everyone in the World talks to everyone else on all matters. I don't see that happening any time soon.
Although, a good approach to fixing such differences is by backing off to look at the larger picture to see where they fit in relationship to each other. I don't know enough about Maple to have any context on this. I've only been in the envrionment 3 days so far. I am on trial version and thinking about getting an individual license.