Mixing collect-expression and collect-formula instructions

[rakhimov]

It's not certain how to handle mixing expressions with formulas
in event-tree branches.
Some ideas/approaches:

• The Ugly:
  Make it a special case across all the analyses,
  and multiply the expressions with the probability expressions
  of the collected formulas to find the sequence probabilities.

• The Bad:
  Create 'phony' basic-events for event-tree functional event path states
  that are described by the collected expressions.
  However, these 'phony' events cannot show up in reports
  or fault tree analysis products.
  Moreover, since the functional event states are not constrained,
  the actual semantics of these basic events cannot be deduced.
  In other words, the analysis can end up with a non-coherent fault tree
  without even being capable of deducing it from the input alone.

• The Good:
  (a) For qualitative analyses, force users to be explicit
  and to use concrete Boolean formulas or basic-events
  for functional-events and their states.
  (b) Perform only quantitative analyses assuming (testing?)
  that the collected formulas are conditionally independent in parallel paths;
  however, put a giant warning in the report.