- From: Steven Pemberton <steven.pemberton@cwi.nl>
- Date: Mon, 28 Aug 2023 10:07:29 +0000
- To: "Norm Tovey-Walsh" <norm@saxonica.com>
- Cc: Dorothy Hoskins <dorothy.hoskins@gmail.com>, "Liam R. E. Quin" <liam@fromoldbooks.org>, graydonish@gmail.com, ixml <public-ixml@w3.org>
On Sunday 27 August 2023 15:18:56 (+02:00), Norm Tovey-Walsh wrote: > Steven Pemberton <steven.pemberton@cwi.nl> writes: > > Nothing needs to be added to the language. > > *Shrug* > > You’ve missed the point, I think. The goal isn’t that I should be able > to define my own nonterminal to mean empty. The goal is to prevent > grammars from using the absence of any token at all to mean empty. > > I’ve tripped over the empty choice in rules like this one > > rule: thing1; thing; thing;. > > enough times to think it’s a bug not a feature. > > Several other readers on the list expressed sympathy for this position. Oh yes! I forgot about your empty-blindness ;-) I shall try to improve. I don't know if was thanks to my early introduction to Algol, or English, but "; ." is so visible to me that I forget others don't see it. I blame C for the barbarism of making semicolon a terminator rather than a separator. Steven > > Be seeing you, > norm > > -- > Norm Tovey-Walsh > Saxonica >
Received on Monday, 28 August 2023 10:08:27 UTC