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

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





------- Comment #8 from jmdyck@ibiblio.org  2006-04-15 08:16 -------
(In reply to comment #7)
> 
> The interpretation I have for a grammar is that it derives a (possibly
> infinite) language of "words" that are accepted by that grammar.

Yup. (You could also call them "sentences".)

> Here each of those "words" correspond to a data model value.

Yes, if you start with the Value non-terminal. (For "correspond to", one might
also say "denotes", "represents", or "means".)

> For instance:
> 
> (1,2,3) is a word that correponds to a sequence of 3 integers 1, 2 and 3, and
> is a valid instance of the formal value grammar.

Actually, you can't derive
    (1,2,3)
from Value, because of the parentheses, but you can derive
    1,2,3
*provided* you assume that the Decimal symbol can derive 1, 2, and 3. (There's
no production for Decimal, so the grammar would appear to be treating it as a
terminal symbol. However, examples like
    42 of type xs:integer
are clearly treating it as a non-terminal that can derive 42 (etc).)

But yes, I agree that the Value
    1,2,3
corresponds to the value you described.

> Can you give an example of a simple value which won't be a formal value?

Nope. Given the definitions of "value" and "simple value" in Section 2.3.1,
it's clear that every simple value is a value.

(However, that's an assertion about values in XS value spaces, whose only
connection to Formal syntactic objects is via the (not actually defined)
"corresponds to" relation.)

Received on Saturday, 15 April 2006 08:16:27 UTC