The first two lines from the article:

"Suppose for a moment that you're an engineer designing a new 8051-based product. Not unexpectedly, the application's code size will greatly exceed the 64KB architectural limit of the 8051's program memory." (my emphasis)

This means only one thing: You are a lousy engineer that has chosen the wrong controller.

While the proposed solution is kind of interesting, from a purely academic point of view, it would require a custom assembler/compiler, as Jan already said.
And more importantly, there is no such thing! The article merely describes a possible solution, but presents it in a way that, at first glance, you'd think it is real.

In my ~15 years of working with the 8051 family, I have yet to exceed 32K code size, let alone 64K. If, in the future, I run into a design that I expect would "... greatly exceed the 64KB architectural limit of the 8051's program memory", I would choose a different type of controller. With ARMs available for less than $10,- the choice would be obvious.


Rob.