[Bug 1804] [FS] editorial: E.1.4.2 Erases

http://www.w3.org/Bugs/Public/show_bug.cgi?id=1804





------- Comment #12 from jmdyck@ibiblio.org  2006-04-16 09:09 -------
Okay, let me see if I can clarify some things.

Re my use of "derive": Formal language theory defines the "derive" relation on
sequences of symbols; A derives B (within a grammar G) iff (roughly speaking)
you can get from A to B by applying productions from G. I'm using exactly that
sense when I say that 'Value' does not derive 'SimpleValue': you can't start
with the 'Value' symbol, apply productions from the Formal grammar, and end up
with a 'SimpleValue' symbol.

I also use "derive" in a closely allied sense: a (non-terminal) symbol
'derives' its derivation trees (i.e., the syntax trees that have that symbol at
their root).

Which brings me to my next clarification. As I said in comment #4, by
"syntactic object", I mean 'derivation tree' (or 'syntax tree'). Whereas it
looks like you're taking it to mean the end result of the derivation, i.e., a
"word" (sequence of tokens or characters) in the language generated by some
symbol. So I'll avoid using the term "syntactic object".

Given those clarifications, we go back to the question: What things can legally
be bound to an italicized non-terminal (e.g., SimpleValue) in a judgment?

My answer is: any syntax tree that is derived from the (SimpleValue)
non-terminal.

Your answer is: any sequence of tokens/characters that can be derived from the
(SimpleValue) non-terminal.

My answer says that there is nothing that can bind to both italicized words
'SimpleValue' and 'Value' (and so those two italicized words cannot unify);
your answer says that there are lots of things that can bind to both (and so
the two words *can* unify).

My objections to your answer are:
-- The FS is fairly clear that it's dealing with syntax trees, not strings of
tokens or characters. See 3.2.1 and 3.2.3.
-- With your answer, I believe it's harder to prove that the rules are complete
and consistent (not that anyone is trying to do such a thing), and also harder
to be certain that the rules are saying what you want them to say.

Received on Sunday, 16 April 2006 09:09:16 UTC