Re: Why DOCTYPE Declarations for XHTML?

"W. Eliot Kimber" wrote:
> 
> I appreciate what Murray is trying to do and I completely understand his
> intent, but it is misguided and will only lead to pain. It is not
> necessary for the XHTML user community to understand the subtlty of my
> argument in order to use some mechanism other than DOCTYPE declarations.
> It is only required that XHTML do the appropriate thing.

Pain? What pain? We're doing the appropriate thing within the world of
XML, as defined by the XML 1.0 Recommendation. Your suggestion would be
inappropriate within this, and would be invisible to current XML parsers.
The use of DOCTYPE is completely conformant with the XML 1.0 Recommendation.
 
> Why should XHTML require the the feature of SGML that XML correctly
> worked so hard to eliminate (DOCTYPE declarations) when the use of it is
> not reliable as a way to define types, for all the reasons I've stated
> many times?

Simple answer: there's not a conformant XML parser alive today that does
anything other than use DOCTYPE to declare what document type definition
a document conforms to, and as you well know (being to my knowledge part 
of the reason you left the XML WG), the emperor doesn't like PIs. We've 
been told repeatedly to avoid use of PIs for anything fundamental. 

The current use of a PI to declare an XSL stylesheet (from "Associating 
Style Sheets with XML documents" by James Clark) is to my knowledge an 
interim measure:

   "The use of XML processing instructions in this specification should
    not be taken as a precedent. The W3C does not anticipate recommending
    the use of processing instructions in any future specification. The
    Rationale explains why they were used in this specification." [1]

If XML parsers expect something (as defined in the XML 1.0 Recommendation)
we will follow that expectation. It does little good to suggest alternatives,
regardless of whether you are right or wrong, whether this is a religious
issue in the markup world, etc.  
 
> There exist at least two standardized facilities for unambiguously
> binding documents to their types: architecture use declaration PIs as
> defined in ISO/IEC 10744:1997 (as ammended) and name-space declarations
> coupled with normative meanings in their governing specs (i.e., as done
> for XLink).  XHTML could use either of these or, as Arjun points out,
> trivially define a purpose-built PI. It also seems likely that the XML
> Schema group will at some point realize it needs a similar type of
> declaration and provide it.

I've heard nothing of this activity and given the dislike of PIs in 
the W3C this seems extremely unlikely. If such a change happens within 
the life of the HTML WG we could certainly take advantage of it. Otherwise
the notion is irrelevant to our activity.
 
> The advantage of the PI is that it does not require a DOCTYPE
> declaration yet still enables validation because any processor can still
> apply the XHTML DTD to the document *as though it had declared it as its
> syntatic DOCTYPE declaration (this is essentially what the 10744
> architecture mechanism does).
> 
> I think it would be a serious mistake for XHTML, probably the most
> important standardized XML application, to require thte use of DOCTYPE
> declarations to define types. It would not serve XHTML well and it would
> serve to institutionalize what could be the worst single bad practice
> the SGML and XML community suffers under.

Until we have an alternative endorsed by the W3C in a W3C specification we
must continue to use the tried and true methods as defined in SGML and XML,
ie., the tools we have. We can't even use XLink in our specs until it 
reaches Recommendation.
 
I honestly don't know why you're arguing with me about this. It seems 
pretty silly. Neither I nor any of the other HTML WG members has any
power to do anything other than take what we've been given (XML 1.0)
and work with it. If you think there should be another way I suggest you
propose it to Tim Berners-Lee and his crew. Our working group is bound
to use features that are politically and technically available to us.
Until there is a recognized alternative all this discussion is moot.

Murray

[1] http://www.w3.org/TR/xml-stylesheet/
...........................................................................
Murray Altheim                                   <mailto:altheim@sonic.net>
Member of Technical Staff, Tools Development & Support
Sun Microsystems, Inc. MS MPK17-102
1601 Willow Rd., Menlo Park, California 94025  <mailto:altheim@eng.sun.com>

   the honey bee is sad and cross and wicked as a weasel
   and when she perches on you boss she leaves a little measle -- archy

Received on Monday, 17 January 2000 15:17:45 UTC