RE: [xml-dev] Saxon and Sun Serializer problems?

["Jim Tivy" <jimt@bluestream.com>]

Hi Folks

I would be interested in the "exact reason" why this has happened in the
specs.  As well, I would like to be instructed as to why my concern is
obscure or misdirected.

In the meantime, on a practical level the Trax API in java is based on SAX
and handles both the LexicalHandler and the ContentHandler.  The
LexicalHandler has a method called:

void startDTD(String name,
              String publicId,
              String systemId)

which is used to serialize the DOCTYPE in the Sun provided serializer.

From what I can tell, XSLT 2.0 is based on XPath 20/XQuery datamodel which
does not include the DocType declaration that the infoset includes:

For example, when constructing XQuery DM from Infoset the text reads:

"If present among the [children], the document type declaration information
item is ignored."

from 6.1.3 of
http://www.w3.org/TR/xpath-datamodel/#const-infoset-document

Since XSLT is a "pass through" kind of technology I don't see the sense of
being lossy.  Especially when you can have Xml dialects like DITA where the
processing is off the class attribute and can handle multiple root elements
and corresponding DTDs.

Jim

-----Original Message-----
From: Amelia A Lewis [mailto:amyzing@talsever.com] 
Sent: Friday, May 29, 2009 5:43 PM
To: Jim Tivy
Cc: 'Michael Kay'; xml-dev@lists.xml.org
Subject: RE: [xml-dev] Saxon and Sun Serializer problems?

On Fri, 29 May 2009 14:40:38 -0700, Jim Tivy wrote:
> With regards to what a Transform Identity means - what does identity
means.
> If it means InfoSet identity then if I read Infoset right: 

But it doesn't.  The infoset specification came out after DOM and SAX 
and XPath, and attempted to unify the various different models.  XPath 
(1.0 in this case, I assume) has its own notion of what is and is not 
important (or even visible).  XSLT (1.0) builds upon XPath.

> If it is XPath/XQuery DM identity - then if as you say there may be no
idea
> of a DocType there, then that is a perhaps a flaw.

I dunno if they've addressed the issue in the XQuery Data Model (which 
is post-infoset, and an attempt to be still more formal and rigorous), 
as the doctype declaration or internal subset effectively exists at a 
"different layer" of processing (this is true even for bare XML 1.0, 
pre-edition-X, pre-namespaces, pre-infoset).  It is, roughly speaking, 
the same reason that an identity transform of é is considered 
correct if it produces é (the transformation does not operate at the 
processing level that exposes general parsed character entities; it 
just gets characters ... and likewise, the doctype decl and internal 
DTD subset are gone before it has a chance to look at them).

Amy!
-- 
Amelia A. Lewis                    amyzing {at} talsever.com
According to Business Week, in the 1990s the ratio between a chief
executive's salary and the takehome pay of the typical, feckless, 
whining grunt on the shopfloor rose from 85:1 to 475:1. (In the UK, 
which is seeing a vigorous popular backlash against "fat cat" pay 
packets, the ratio is 24:1).
               -- The Register