Re: Notations are useful

On Sat, 5 Feb 2000, Murray Altheim wrote:
> Arjun Ray wrote:
> > 
> > the HTML DTDs don't use [NOTATIONs], and it seems they may never.
> 
> Well, only because the part we need in XHTML is missing in XML.
> For now, anyway.

Yup.  Notations are much more useful with attributes.  (Even so, what
limited use they have without attributes is still being pretty much
ignored in XML, I think.)

> > For instance, something like this isn't really meaningful:
> > 
> >    <!ATTLIST ...
> >          href   CDATA    #IMPLIED
> >          ... >
> > 
> > and hiding the CDATA declared value in a suggestively named PE,
> > like  %URI.datatype; - as in the new xHTML 1.1 DTD [1] - is
> > really just all handwaving too, since the *essential*
> > information is inside a comment:
> 
> Only because that is our only option. At least now they're
> regularized.

Sure.  This wasn't a criticism of the DTDs.  It was a more or less
paradigmatic example of what we sort of routinely do, without much
satisfaction, lacking better means.

> As you know, I'm quite interested in use of WebSGML's 'DATA'
> attributes feature, and have lobbied within the HTML WG to begin
> work on figuring out exactly what all of the data types (ie.,
> notations) currently used in XHTML are, and come to some
> determination on how they can be declared in a way that is the
> same between XHTML DTDs and Schemas. 

For now, it could be kept as a disabled option, using some more PE
voodoo:)  I'm thinking of a marked section with a different set of
definitions at the top of the -datatypes.mod, something like this

  <!ENTITY %Use.Data.Atts  'IGNORE' >
 <![%Use.Data.Atts;[
  ...
  <!NOTATION Uri "-//IETF RFC 2396//NOTATION ..." >
  <!ENTITY %URI.datatype 'DATA Uri' >
  ...
 ]]>
  ...
  <!-- a Uniform Resource Identifier, see [URI] -->
  <!ENTITY % URI.datatype "CDATA" >

which would leave the rest of the current modularization structure
unaffected.  (Alternatively, all of these new definitions of the
declared values could be in a separate file, included via a switchable
PE in -datatypes.mod.  This might be the better idea for now, while
the set of NOTATIONs is being determined; later on, they should be
integrated into one file to avoid PE name-coordination problems.)

> We're kinda in uncharted territory here since this would require a
> change to XML in order to really use it, but it's nevertheless not
> wasted activity. [...] If the 'DATA' attributes feature proves
> that valuable, then perhaps we can lobby for its inclusion in a
> future version of XML.

I don't think it needs drastic changes in non-validating parsers, so
I'm not sure that we need to wait for a new official spec.  The point
I think is to get the word out, and see (from feedback) whether it's
useful.  

> In line with a 'me too' comment: write up a proposal, submit it to
> the W3C Core WG. 

Sorry, proposal for what?

> They may ignore you, 

Whoops, the 'briar patch' strategy...?;)

> Your input on architectural forms wasn't wasted. (Thanks!) 

There are three interrelated points about AFs, I think, that I tried
to bring out in my recent post to this list:

  http://lists.w3.org/Archives/Public/www-html/2000Jan/0217.html

  1.  The distinction between 'encompassing' and 'enabling', i.e.
      that it's not necessary for a document as a unitary entity
      to validate.

  2.  AF mapping is a purely formal, mechanical procedure, requiring
      no new syntax - and, as a matter of fact, helped tremendously
      if wellformedness is guaranteed to begin with.

  3.  Validation of architectural projections *is* useful, and I 
      dare say mandatory.

(I'll add that much of the namespaces 'debate' on the old xml-sig
group was taken up by misunderstandings on these issues.  For
instance, arguments about #3 - notably by Eliot Kimber - were dimissed
as irrelevant, and David Durand's proposals for attribute based
processing - the essence of #2 - was accused of being #3 in disguise
and likewise ruled out of court.)

> I don't in the end know what will happen with it, but I gave a
> short presentation on the functionality of AFs at the last F2F
> meeting, so that's progress in some people's book.

I wonder what they would think of DAFE (Data Attributes For Elements)?

  http://www.ornl.gov/sgml/wg8/docs/n1920/html/clause-A.5.3.html


Arjun

Received on Saturday, 5 February 2000 22:40:28 UTC