Where do the "default renderings" in spec chapter 4 come from?

Dear MathML developers,

  as some of you are aware of, our group is doing research on mathematical
notations.  In the MathML 3 spec, chapter 4, you give a "default rendering"
for each Content MathML symbol or constructor.  For our research it would now
be interesting to learn from you:  Where do these default renderings come
from?  How did you agree upon one rendering when you had to choose between
more than one?

For example: Why did you choose (0,1) for an open interval instead of ]0,1[?
Why ø for the empty set but not {}?  Why \subset for the proper subset but not
\subsetneq?  Why f' but not df/d? -- Note: Of course I don't expect an answer
for every individual case, but some hints on your general intuition or
strategy would be helpful.

Thanks in advance,

Christoph

-- 
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701