I think the difference here is one of definition. If you look at section data (plotted from wind tunnel data where the wing extends from wall to wall simulating an infinite span wing) you see that as the angle of attack is increased from the zero lift angle, there is a drag component that increases as lift increases. The actual force on a wing section developing lift is always tilted rearward resulting in "induced drag". This is the penalty nature extracts for producing lift. The total drag of a wing section is the sum of "form" drag (the parting of the air and friction as the body passes through it), and the "induced" drag (caused by the rearward tilting of the resultant force on the wing). This induced drag is always present whenever the wing is generating lift.

There is another factor to be considered for a wing of finite span that is the result of spanwise flow. This is also frequently referred to (even in some texts) as "induced" drag. This factor would more properly be called "additional induced" drag. This is the factor that appears in the equation:

Cdi=CL (squared)/(pi)(aspect ratio)

As A.R. becomes infinite, this additional induced drag (Cdi) becomes zero.

Total drag = form drag+induced drag(due to lift)+Cdi(due to finite span)

Cdi is affected by planform and tip design but even in high angle of attack situations is seldom the major source of drag.

I believe if this confusion of terms is cleared up we will find that the several opinions are indeed saying the same thing.