W3C home > Mailing lists > Public > w3c-sgml-wg@w3.org > June 1997

Re: KISS (was: Parameter entity references in WF docs)

From: Joe English <jenglish@crl.com>
Date: Wed, 04 Jun 1997 11:17:51 -0700
Message-Id: <199706041817.AA17635@mail.crl.com>
To: w3c-sgml-wg@w3.org

Eve L. Maler <elm@arbortext.com> wrote:

> At 12:31 PM 6/3/97 -0700, Joe English wrote:
> ...
> >CDATA and RCDATA declared content can be replaced with (#PCDATA),
> >as this is semantically (though not syntactically) equivalent.
>
> Since XML has no inclusions, I think (#PCDATA) results in a model whose
> instances have the same constraints as RCDATA models.  Right?  CDATA is
> another story.


Depends on how you look at it I suppose...

Neither ESIS nor the SGML grove plan distinguish elements
with declared content from those with mixed or element content
(unless you look up the element type declaration in prlgabs1),
which is why I feel that they're semantically equivalent.
That inclusions cannot appear in RCDATA declared content
nor data entity references in CDATA declared content I
consider to be an incidental consequence of the parsing
process; that is, there's no semantic restriction against
them, it's just that the parser will refuse to recognize
them should they appear.

In any case, replacing CDATA declared content with (#PCDATA)
yields a superset language, and changing RCDATA declared content
to (#PCDATA) yields an equivalent language (modulo delimiter
recognition modes), like Eve said.


BTW, in my previous message I left out: If an element has
a #CONREF attribute, and its content model does not allow
empty content (i.e., is not "inherently optional"), you have
to add an OPT occurrence indicator to make this possible.

(I also completely forgot about entity declarations, but
Eve's message covered those pretty thoroughly.)


--Joe English

  jenglish@crl.com
Received on Wednesday, 4 June 1997 14:31:34 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 10:04:39 EDT