Is the link between bowls of a lowercase /g/ really a terminal?

The description cited refers to the ‘ear’ of the g, not the connection between the bowls. The ear is the bit that sticks off the top right.

I disagree with the terminology as used on the MS site. The ear of the g is most often, well, an ear, not a terminal. It doesn't terminate anything; it is its own thing (in certain styles of g, the ear does form a terminal, i.e. if the top of the bowl flows into the ear independently of the right side of the bowl, in which case the ear forms the terminal of that stroke, but in most styles of g the ear is an appendage stroke, not a stroke terminal).

I would say that a terminal is the termination of a stroke, and hence a terminal can take many different forms. There are ball terminals and spike terminals and curled terminals and blunt terminals and, yes, serifed terminals. What determines that something is a terminal is that it ends a stroke. This is the only criterion by which you can determine, in a given situation, whether something constitutes a terminal or something else, e.g. an appendage. Structurally, a terminal is part of the stroke, i.e. something that if the stroke were drawn would be a continuation of the movement of the stroke, not a new stroke (a blunt terminal, of course, is a non-continuation of the stroke). From this it follows that serifs are sometimes terminals and sometimes not: it depends how the serif is constructed.

From this it follows, of course, that in type terminals are parachirographic constructs. It only makes sense to talk about terminals in the context of a stroke model of letter description (which does not necessarily imply a stroke model of letter making). If there are no strokes, there are no terminals.

Funny, I was reading this exact page only a couple of days ago, and I disagreed with it too. To me it seems strange to talk about the arm of a C (it says the arm is the portion across the top) and so the 'part before it hits the arm' makes no sense at all.