Re: Suggestions for FF[5
- Hyoungsoo Yun (Mar 7 1999) wrote:
>I'd like to propose that we include more *rigorous*I think that this proposal, combined with Hyoungsoo's draft set of
>definition of SGF (e.g. in EBNF) into FF.
EBNF productions, form an excellent initiative which will be
welcomed by anyone who sets about writing a new SGF-processing tool.
I would like to table two related suggestions:
1. That the formal set of EBNF productions to be included in
FF5 be numbered (to facilitate reference and discussion).
2. That the spec. adopts an EBNF notation which uses the regular
expression symbols *, +, and ?, to facilitate and abbreviate
indications of whether or not particular elements are
mandatory/optional and repeatable/non-repeatable.
(i.e. the familiar meanings of:
* "0 or more" (optional, repeatable)
? "0 or 1" (optional, not repeatable)
+ "1 or more" (mandatory, repeatable)
(no symbol) "exactly 1" (mandatory, not repeatable) )
Apart from the inherent power and economy of this notation
(see for example its use in the official XML spec) it would
facilitate development of a related DTD, since the syntax of
SGML/XML DTD's requires this format.
3. (viz HyoungSoo's comment in http://www.vilab.com/bgml/sgf4.html#notes
>it seems too tedious to include all restrictions on property values: That even if the listed EBNF productions do not reach all the way down
>for the predefined properties, and it has been left out.)
to the data format for each property (personally I think it would be
useful, if onerous, if they did), that the "lexical" entries for the
grammar (the detailed discussions of each property) do at least include
a formal regular expression pattern for each data format.
Again, this will be required for a full DTD, and would be a very useful
reference for anyone writing a parser, especially in a language that uses