From: Kurt W. Hirchert					     J3/00-295r1
Subject: BNF Analysis					     Meeting 154


Like many other programmers, I couldn't resist enhancing the BNF
extraction and analysis script that I wrote during meeting 154.  In
addition to better commenting it and improving the appearance of some of
its output, I added a whole new section to recognize problems like rules
that are not defined in the section matching rule number and rules that
are not consistently defined in different sections.

I have the divided the electronic representation of this document into
three files:

Part A is this message.

Part B contains the extracted BNF, without constraints or duplicate
occurrences of rules.  (This was produced by invoking the script with
option -1.)

Part C is the cross-reference of the BNF non-terminal symbols.  (This
was produced by invoking the script with option -10.)  It now handles
the implicit BNF rules somewhat better.  Symbols of the form XXXX-name
are now treated as defined by the rule for name.  References to symbols
of the form scalar-XXXX or XXXX-list are now displayed as part of the
references to the underlying symbol XXXX.

I have given the script itself to Richard for inclusion in the scripts
directory of the 007 distribution.

If there is enough demand, I will work with Richard to adapt this script
to read MIF and generate MIF so the output could be included in 007, but
my initial look at the MIF suggests it is sufficiently ugly that I am
unlikely to switch scanning it unless there is such demand.