RE: RFC: White Space Handling In XML Parsing

Some general thoughts: 

1. It would be really, really nice if the DEFAULT behavior were 
deterministic, i.e., guaranteed to produce the same DOM tree
under different parser implementations.  If one is concerned
about parser efficiency (i.e., not unduly constraining parser
implementations), then an option could allow the parser to
produce its preferred DOM tree for maximal speed.
In other words, it would be nice if the specifications supported
an application development model of: "FIRST make it correct, 
THEN make it fast".  This would be best supported by making
the default behavior be the most deterministic, and the non-default
behavior be the fastest (assuming one must choose between the two
objectives).

2. It seems to me that the most appropriate 
default should be to NOT preserve whitespace, since:

	a. This is likely to be the behavior that is
	desired by most applications; and

	b. The people writing the applications that wish to
	PRESERVE whitespace will be a somewhat more select group
	of programmers; thus it is more reasonable to require
	them to know a little more than others, rather than
	the other way around.

Personally, I like to think of XML as a very direct
representation of an abstract syntax tree.  Under this
kind of thinking, syntactic details such as whitespace 
are irrelevant.  This viewpoint may reflect a bias of my
background, but I suspect I am not alone in this view.
This is another factor in favor of making the default
behavior be to NOT preserve whitespace.  However, I am
certainly open to other viewpoints.

David Booth
Bluestone Software, Inc.      +1 609 727 4600   ext. 1740


> -----Original Message-----
> From: Joel A. Nava [mailto:jnava@Adobe.COM]
> Sent: Tuesday, May 18, 1999 8:19 AM
> To: www-dom@w3.org
> Cc: www-dom@w3.org
> Subject: RE: RFC: White Space Handling In XML Parsing
> 
> 
> xml:space=default implies that the application should use it's
> default whitespace handling mode, which could be what ever the
> application has decided it's default mode. If xml:space is not
> defined on the root element, then the application has to decide
> for itself what it will do. There may be some circumstances where
> is default is expressed the application handles things one way,
> and if it is not defined, then the application might handle things
> a different way. I agree that it is clearly permissible to say
> that parsers and application conforming to your RFC will have
> the same behavior whether default or undefined.
> 
> --
> Joel A. Nava                  (408)536-6209
> Adobe Systems, Inc.         jnava@adobe.com
> 
> > -----Original Message-----
> > From: Arkin [mailto:arkin@trendline.co.il]
> > Sent: Monday, May 17, 1999 6:41 AM
> > To: John Cowan
> > Cc: Joel A. Nava; www-dom@w3.org
> > Subject: Re: RFC: White Space Handling In XML Parsing
> >
> >
> >
> >
> > John Cowan wrote:
> > >
> > > Joel A. Nava wrote:
> > >
> > > > 3) Elements that do not specify a value for the 'xml:space'
> > > >    attribute inherit that value from the element in which
> > > >    they are contained up to the root element. If the root
> > > >    element does not specify a value for the 'xml:space'
> > > >    attribute, the value 'default' is assumed.
> > > >
> > > > [This is what the XML REC requires.]
> > >
> > > Actually not.  The last paragraph of clause 2.10 says:
> > >
> > > # The root element of any document is considered to have
> > > # no intentions as regards application space handling,
> > > # unless it provides a value for this attribute or the
> > > # attribute is declared with a default value.
> > >
> > > So there are really three possible states of "xml:space" as
> > inherited:
> > > default, preserve, and clueless.
> >
> > That pretty much depends on how you define "xml:space=default". Is
> > 'default' a specific behavior, or is 'default' really "no
> > intentions as
> > regards application space handling". My understanding of the
> > XML REC is
> > that there are only two possible states: preserve and default, where
> > default and clueless are pretty much the same.
> >
> > In no place does it really say what 'default' should be, 
> which is what
> > the WS RFC attempts to clarify.
> >
> > Arkin
> >
> > >
> > > --
> > > John Cowan      http://www.ccil.org/~cowan
> > cowan@ccil.org
> > >         You tollerday donsk?  N.  You tolkatiff scowegian?  Nn.
> > >         You spigotty anglease?  Nnn.  You phonio saxo?  Nnnn.
> > >                 Clear all so!  'Tis a Jute.... (Finnegans 
> Wake 16.5)
> >
> >
>

Received on Wednesday, 19 May 1999 18:24:35 UTC