AW: Re: Using content-MathML for computation and analysis in Science and Engineering

I did not actually intend to detract from the previous discussion on assignment operators. I just noticed that a piece of common knowledge would have been important to know for newcomers to MathML Content. 

That said, I'm afraid that I need to burst a little bubble here. MathML Content brings with it what you need to describe (the semantics of) side effect free algorithms. 

Any clean Content markup language would need to add a language primitive for something that has a semantics that falls outside of the range of its existing semantic constructors. Since assignments are not context free, you would probably have some serious trouble to define an OpenMath dictionary with a clean semantics for an assignment operator.

In my 2003 dissertation I argued that this compositionality principle for content markup languages implies, in a similar logic, that typing is another semantic primitive missing from OpenMath (and by implication, MathML 3).

Now as we all know implementations are free to ignore this, as is Peter. As always, my comments are just for the language purists. 

-Andreas null