The scheduling in Squeak is very flexible and open, and it would be
wrong to think that because something doesn't come implemented "out of
the box" it means that Squeakers in general are not aware of the idea.
It probably means that enough applications can do just fine without it
that the necessary effort was never spent to make it a standard feature.
See the simpler "blocks are true closures" issue for an example of this.
It isn't that we don't know about closures. It isn't that we don't like
them. They are just not the highest priority right now.

Back in 1994 I added a "real time" thread scheduling system to Smalltalk
V/286, which has an infrastructure almost exactly like Squeak's. It was
great to have "first deadline first", "rate monotonic" and other odd
kinds of threads running side by side with the fixed priority ones and
even the cooperative threads. All this for about a page of code.

A friend did something similar for his PhD thesis, where he expanded the
scheduler for Self (a bit different from Squeak's, but comparable) so
that simulated hardware objects (and eventually real hardware) could be
mixed with the normal software objects and Morphic.

-- Jecel