Cramped style change (#164) meeting discussion

Our meeting on Monday spent most of the time discussion #164 -- what should
be done about cramped style.

The beginning of the discussion agreed that because checking for
embellished operators was hard and that having an embellished operator in
an accent is extremely rare, we should change the rule from looking for an
embellished operator (which is basically impossible in CSS) to just looking
for an operator. That was tacitly agreed to without a request for consensus.

Then the discussion veered off onto accents on <mo> in general and other
rules before coming back to cramped style. After 50 minutes of discussion,
we agreed that the 1/2 point difference in height between cramped and
uncramped for mover/munderover is not much in the grand scheme of things
and so we should do the simple thing and ignore this case.

I volunteered to write up the discussion which led me back to the start --
just don't worry about embellished operators in the accent . That seems to
have gotten lost at the end of the discussion.

Question: since the goal was to find a simple solution, why:
1. Don't we check the accent property on mover/munderover
2. Don't we check for that property on an (unembellished) mo in the accent
position

Both these seem like simple CSS rules to write and catch virtually all
cases that were missed. I put the examples into codepen
<https://codepen.io/nms/pen/KKpYJMP>.  Firefox does this correctly, which
you can see if you view the codepen in Firefox.

I've sent this to the mailing list because several people aren't subscribed
to the issue. I'm putting this email into the issue
<https://github.com/mathml-refresh/mathml/issues/164>, and people should
reply there with their thoughts rather than via this mailing list so that
it is tracked properly.

    Neil