Re: Why Doesn't IE5 use the DTD to Validate?

Subject: Re: Why Doesn't IE5 use the DTD to Validate?
From: Chris Lilley <chris@xxxxxx>
Date: Sun, 04 Apr 1999 20:00:22 +0200

Didier PH Martin wrote:
> [I said]
> If, in the documentation for the aircraft I am flying on or the nuclear
> plant that generates my electricity or the maintenence manual for a
> blood gas analyser whose results my doctor is depending on, the document
> is invalid because
> 
> <!ATTLIST failure-condition severity
> (silent-redundant|minor|fixat6months|fixin24hours|urgent|emergency|criticalh
> alt)
> #REQUIRED>
> 
> and the required severity attribute is missing, and the style sheet has
> 
> diagnostic fault failure-condition[severity=criticalhalt]:before
>   {content: "Shut this thing down NOW before she blows"'
>    color: red; background: black;
>    font-size: 8em; text-decoration: blink }
> 
> I certainly want to know about it and I want a validation check each
> time the document is viewed.

> And just to take your scenario. The receiving browser equipped with a
> validation parser detect structure error, the document is not displayed

Correct, which immediately shows that there is an error. So, the first
time the documentastionis delivered, it is immediately apparent thatthe
documentation is not valid.

Your way, the document is displayed and the only sign that something is
amiss is that there is no red warning notice on page 576 when there
should be. Easy to miss. And then, later, when some smallpart fails,
there is nothing to indicate that this failure is actually tremendously
serious. 

> and
> the nuclear central creates a huge ecological disaster. If the document is
> not valid a validating parser cannot fix the situation. 

True, but it makes the invalidity very obvious, each time the
documentation is inspected. So, it will get fixed; hopefully, the
documentastion will have been viewed at least once before someone needs
to consult it in detail about some failure mode.

> It can just tell the
> nuclear operator "run for your life dear". And a browser using a non
> validating parser will just end up with the same result :-)))

> > All browsers are equal, but some are more equal than others. And, some
> > things are more important than electronic commerce.

> <Reply>
> Sorry Chris, I don't follow you there. I guess you just got me with an April
> foul :-)))
> </Reply>

Sorry. The first comment was a reference to the George Orwell satirical
novel, Animal Farm. The second comment was taking exception to what you
seemed to be asserting - that document presentation could always get by
with invalid documents, and ecommerce could *never* get by with invalid
documents, thatthat there were no other categories.

Incidentally did anyone tell you that your <YourComment> and <Reply>
elements are  not scalable (because the meaning of the referent "Your"
is lost) so cannot be used for multiple passes or response? Also, you
should have a single top level element?

Perhaps

<mail>
<writers>
<participant name="Chris Lilley" id="chris@xxxxxx">
<participant name="Didier PH Martin" id="martind@xxxxxxxxxxxxx">
</writers>
<posting>
<statement msg-id="3704E841.BF612F3@xxxxxx" author="#chris@xxxxxx"
id="f12345">
Yes, but

a) they will only take that responsibility if it buys them something
b) if it seems to usually work anyway without validating, they don't
think it buys them anything
</statement>

<statement referent="#f12345" author="#martind@xxxxxxxxxxxxx"
id="g987654">
Simple scenario: The emitter sends a document having its structural
integrity not valid....
</statement>
</posting>
</mail>

Except that email addresses and internet message ids are not valid
NAMEs, but hard to see how to get unique IDs otherwise. And you do need
to parse (but not to validate ;-) to find out that author and about are
declared as IDREF.

And, to not restrict ourselves to single documents nor to the structural
units provided by the participant one is commenting on, we should use
XLink and XPointers

And, the expectation of trimming down referred to material is hard to
reconcile with the concept of uniquely named statements.

And lastly, that everyone has to play or it doesn't work. Whhich needs a
flag day on email clients; alternatively, a parallel non-mail discussion
forum.

Now, if everyone had access to their own personal web server, this could
work...


--
Chris



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread