[FLD/BLD/PRD] PS for variables

All (and Harold and Michael in particular),

As mentionned in several comments and review already, the current PS for 
variables in BLD and FLD, which is reused in PRD as well:
  Var ::= '?' UNICODESTRING
is problematic.

In the case a rule is authored using the PS, what will be the content of 
the XML <Var> element: '?' UNICODESTRING, or UNICODESTRING?

If the rule has been authored using another syntax, the XML Var element 
should quite obviously contain the string that represents the name of 
the variable as chosen by the author. But then, what will be its PS 
representation?

Suppose the author called a variable "dummy". The XML will show:
<Var>dummy</Var>. Will the PS show "?dummy" or "dummy"?

If the author followed the common practice and started the variable name 
with ? and called it "?dummy", will the PS be "?dummy", like the XML 
content, or "??dummy"?

As a consequence, I suggest to change the Var production to Var ::= 
UNICODESTRING (and to add a comment that it is good practice to start 
the name of a variable with a ?).

Cheers,

Christian

Received on Friday, 4 July 2008 12:47:51 UTC