Empty PE - was Re: XDB 0.3 available (XMLized Docbook 3.0)

Norbert H. Mikula (nmikula@edu.uni-klu.ac.at)
Mon, 05 May 1997 08:26:35 -0700

David Durand wrote:
> I want to ask what justification there is, if any, for ruling out empty
> PEs? I don't remember discussion of this point clearly (though I do
> remember shock when the rule was pointed out).
>
> This seems like a very bad idea, as Terry's desire for add entities in
> is very reasonable.

I do agree, that Terry's desire is very reasonable. However, if we can
find not a formal and concise way to express it we have problems. One of
the
objectives of XML was, that it should be "easy" to implement and it
should incorporate "contemporary" disciplines of computer science like
formal languages etc. As you probably know, I use JavaCC, a Lex/Yacc
like
approach, to build NXP. I really had difficulties to transform this
production to LL(1), and I am still not sure if there is a clean way to
bring it to LL(n) (a way that I could live with).

Furthermore it violates the general idea that %a should actually
satisfy S? a S?. With an empty PE this would not work.

<FN>As much as a first
liked the %a idea, when I had to implement it, and I still not satisfied
with my current solution, it caused a lot of headache (ouuuch).</FN>

So in short, I don't mind the idea of having empty PE's, if it is
possible
to implement/express it in a reasonable way. Any ideas would be
appreciated !

-- 
Best regards,
Norbert H. Mikula


=====================================================
= SGML, XML, DSSSL, Intra- & Internet, AI, Java 
=====================================================
= mailto:nmikula@edu.uni-klu.ac.at 
= http://www.edu.uni-klu.ac.at/~nmikula
=====================================================