W3C home > Mailing lists > Public > www-xpath-comments@w3.org > July to September 1999

XPath (Section 3.7) Lexical Structure

From: Denys Duchier <Denys.Duchier@ps.uni-sb.de>
Date: 12 Jul 1999 22:48:46 +0200
To: www-xpath-comments@w3.org
CC: jjc@jclark.com
Message-ID: <isoghhppap.fsf@grizzly.ps.uni-sb.de>
This is a comment concerning (Section 3.7) "Lexical Structure" in the
XPath specification.

James Clark was kind enough to send me some code clarifying the intent
of the specification and I now realize that Section 3.7 plays an
important role in disambiguating the grammar.  I believe that it would
be extremely beneficial to strengthen the wording in that section.

1. state explicitly that the tokenizing rules serve an important role
   in disambiguating the grammar.

   This may be obvious to those involved in the elaboration of the
   specification, but it is easy to overlook or misappreciate for an
   outsider.

2. identify the exceptional tokenizing rules (they are informally
   listed in the 3rd paragraph of the section) and clarify their
   precedence: e.g. is the 1st identifier in "and(" an operator or a
   function name?

A more general comment: the specification does not make it clear (tout
du moins, pour un lecteur non-averti) that the start symbol of the
grammar is Expr (or, at least, that this is the most general phrase
recognized by the grammar, or something to that effect).  Instead, it
starts in the "middle" of the grammar, with location paths, and then
introduces larger phrasal types; I found that confusing.  I am not
suggesting a different organization, but again a strengthening of the
wording so that the game plan becomes clearer.  Here are some specific
suggestions:

"The primary syntactic construct in XPath is the expression"

This should be followed by "and is described by non-terminal Expr",
with an appropriate link to the corresponding production.

"One important kind of expression is a location path"

This should be followed by "and is described by non-terminal Location
Path", with an appropriate link to the corresponding production.
Furthermore, the specification should say that "due to their
importance, they will be described first in Section 2", or words to
that effect.

In other words, the introduction should state clearly: _The_Grammar_
_Really_Starts_There_ (pointing to expressions) ... but
_This_Is_The_Interesting_Bit_ (pointing to location paths).

Sincerely,

-- 
Dr. Denys Duchier			Denys.Duchier@ps.uni-sb.de
Forschungsbereich Programmiersysteme	(Programming Systems Lab)
Universitaet des Saarlandes, Geb. 45	http://www.ps.uni-sb.de/~duchier
Postfach 15 11 50			Phone: +49 681 302 5618
66041 Saarbruecken, Germany		Fax:   +49 681 302 5615
Received on Monday, 12 July 1999 16:48:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 3 October 2007 16:05:53 GMT