Re: semantical annotations

  Ron,

  You are correct in saying that type information really needs to have
  embedded math objects. A simple example, in Axiom notation, where
  a math object is part of the type is SquareMatrix(2, Integer). Here the
  "2"
  is really a positive integer, although it could obviously be part of a
  string.

  Here is a more sophisticated example (I hope you like the type name!):

  FiniteFieldNormalBasisExtensionByPolynomial(Prime Field 3, X^^3+2*X^^3+2)

  The second argument is a polynomial whose coefficients come from the
  first.

  Contexts, in the OpenMath sense, are not parametrized, so a string will
  suffice there.

  Bob