Re: Syntactic variability

> I don't think removing these alternatives is a necessary change, then.
> But I could live with it.

I concur that it isn’t necessary. Aesthetically, I’m not hugely pleased
with requiring a space before the “:” and that requirement is going to
generate parsing errors that users will have to learn and correct.

Following on something John said, I suppose “:=” or even “::=” could be
used instead of just a bare “:”. Those are, in fact, more common in
(E)BNF grammars in my experience than just “:”.

It feels like we could make users lives easier: prevent confusion that
might arise from different grammar authors using different punctuation,
and remove both an awkward rule and the error that arises from failing
to satisfy it.

I’m not really sure how we wound up with two sets of delimiters anyway,
except perhaps that the CG couldn’t decide so it did the thing that
groups always do when they can’t decide: pick both.

I’ll also repeat myself and observe that there are relatively few
punctuation characters in US ASCII, so chewing up four of them when we
could use only two seems like a choice we might later regret.

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica