There are two recent approaches to modern
temperament design that have interested me:
Bill Bremmer's EBVT temperament, which has undergone some revisions in recent
years. It started
with a compromise between temperaments of the "victorian" style
and equal temperament, to suit an aesthetic that favors variation in dissonance
across the circle of fifths, but is still comfortable to the modern ear in the distant
keys. More recently it has focused on the contribution of temperament design to beat cancellation. That is, focusing on beat rates of interval tests in isolation
does not take into account the effect of temperament on beats in chords and larger
combinations of intervals.
Another is the "pure twelfths" approach exemplified by the work of Bernhard
also focuses on the beat cancellation idea, but comes at it from a mathematical approach.
If the question is concerning experimentation in temperament design, then historical temperaments don't really apply since those such as Werkmeister or meantone have
accepted definitions which are not changing. People experiment with their
applicability to certain situations like spinet scales or period performance, but they
are not experiments with temperament design.