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

Re: DTD Fragments and XML

From: Harvey Bingham <hbingham@ACM.org>
Date: Sat, 03 May 1997 23:42:28 -0400
Message-Id: <>
To: w3c-sgml-wg@w3.org
At 13:54 1997/05/03 -0700, Terry Allen wrote:

>How can you tell what's wrong with the following document 
>without the DTD that says that <cat> has nine <life>s?  This
>kind of insidious data corruption will become common:  docs will
>be WF but will not work right in their intended apps because they
>don't obey certain constraints designed into those apps.   In
>SGML we cover a lot (not all) of this ground with the code portion
>of the DTD and much of the rest (we hope) in the prose portion
>of the DTD.
I believe most applications would leave counting to 9 up to the
semantics, rather than explicitly requiring 9 lives. There are
many places where syntactic freedom is constrained by semantics.
I doubt we'll ever be free of that for non-trivial DTDs.
Those same semantics would seem required to do an adequate style
sheet, without attempting all possible syntactic cases, such as
means to limit recursion, particularly when we have no exclusion
exception in XML, not even self-exclusion. 

My doc would have expected
<scale note="C">meow</scale>
<scale note="B">meow</scale>
<scale note="A">meow</scale>
<scale note="G">meow</scale>
<scale note="F">meow</scale>
<scale note="E">meow</scale>
<scale note="D">meow</scale>
<scale note="C">scat!</scale>

What does the name "life" do differently than the name "scale"
without a DTD.

>But then, more work for consultants is not something to complain
>about, I suppose.
Life is too short, I'm noticing. There are too many ephemeral people
and flowers to enjoy.

>  Terry Allen   

Regards/Harvey Bingham 
Received on Saturday, 3 May 1997 23:42:14 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:26 UTC