Another possible view is that there is no "post process".
The distinction between the rewriting and the scripting aspects of the earlier
draft was short-sited and I welcomed its elimination from the current version.
Since a rewriting system, in general, can be sufficiently powerful as to act
as the basis for a general programming language, the distinction "behaviour"
v/s "style" brings no inherent benefit.
There are publications from several years back on constraint-based programming
environments, (Wm Leler, ...Constraint Programming Languages...) which
demonstrated this for term rewriting systems. Where the "styling" forms
describe rewriting operations all the way to "machine code", there is no need
for a "scripting" language.
If you're curious, a quick search turned up what appears to be an active URL
for him (http://wm.simplenet.com/wm/) and a pointer to a source for the
language he described (http://www.cirl.uoregon.edu/constraints/systems/bertrand.html)