Re: Pragmas Reqs 9-10

> My intuition is that the annotating literals wouldn’t be too hard; annotating individual character classes or hex tokens (or any of the other smaller “atomic” parts) might be possible, but complex. The more complex it is, I think the more impact it’s going to have on succinctness and readability so the use cases have to be progressively more robust.

Right. Given acceptance of point 9, you could annotate a literal by defining a new, suppressed, nonterminal for it, and annotating that nonterminal. But I can also see the case for being able to annotate literals directly. I can also see potential cases for annotating some other constructs. Point 10 doesn't preclude the possibility that these should be added to the list in point 9.

Point 10 just kicks the can down the road a little - it essentially says that adding to the list of annotatable constructs should be a matter for further discussion. I think this is probably the right approach, because it means that the discussion can happen once a concrete proposal for adding pragmas is on the table. I suspect that constructing use cases will become easier under those conditions.

BTW

___________________________________________________ 
Dr. Bethan Tovey-Walsh 

linguacelta.com <http://linguacelta.com/>

Golygydd | Editor geirfan.cymru <https://geirfan.cymru/>

 <https://geirfan.cymru/>Croeso i chi ysgrifennu ataf yn y Gymraeg.