[Grace-core] Some comments from Tijs on parsing Grace

[Prof. Andrew P. Black]

> On 22 Jul 2016, at 23:55 , James Noble <kjx at ecs.vuw.ac.nz> wrote:
> 
> - parse tree can then be more general, because those distinctions don't really 
>     matter semantically, so can abstract them away

I’ve independently come to the same conclusion.  For example, I’ve been confused (in switching from old to new class syntax) when I accidentally put a method inside a method; the parser isn;t looking for the method keyword, so you get a confusing message.  The grammar would be simpler and the error messages better if the parser just accepted it, and then there was a structural check along with the other checks on the parse tree.

	Andrew