Re: Spyglass HTML Validator

Joe English (joe@trystero.art.com)
Mon, 28 Oct 1996 08:58:58 PST


Message-Id: <9610281658.AA03533@trystero.art.com>
To: www-html@w3.org
Subject: Re: Spyglass HTML Validator
In-Reply-To: <327474BB.6DAD@imc.exec.nhs.uk>
Date: Mon, 28 Oct 1996 08:58:58 PST
From: Joe English <joe@trystero.art.com>


Chris Ridd <C.Ridd@imc.exec.nhs.uk> wrote:
> David Perrell wrote:
> > which worked fine in both NSN and MSIE. What I'm hearing is that the
> > URL string is an SGML attribute and to be legal the & needs to escaped
> > as
> >  <A HREF="oop&amp;d.html">
>
> No, I don't believe that's correct. The DTD (certainly the Cougar DTD
> and the HTML 2.0 DTD) declares the HREF's value as CDATA, which by
> definition isn't parsed for SGML entities like &foo;.


David is correct.  Attribute value literals are always parsed
as replaceable character data, regardless of their declared value.
The kewyord "CDATA" in the attribute list declaration doesn't
affect how attribute values are initially parsed.

This is Confusing SGML Issue #3.  See

    <URL:http://www.art.com/~joe/sgml/cdata.html>

for more details...


--Joe English

  joe@art.com