Re: XML Editors

At 10:48 AM 9/13/96 CDT, Paul Grosso wrote:

>I.  Which of the following statements comes closest
>(1) the majority of existing SGML editors can create XML;
>(2) in theory, an ideal SGML editor could create XML;
>(3) new tools will have to be created or existing tools modified 
>    to enable the creation of XML.

(1) - no doubt.  Although since XML abandons useful SGML features in
a trade-off for simplicity and network delivery, one might ask: "If you're
willing to use a real SGML editor, why don't you use real SGML?"  There are
some interim problems if we do things like NET tricks to emulate EMPTY, 
(because the SGML editor will want to write it out differently) but
with luck these will go away assuming WG8 makes a distinct EMPTY syntax.

>(a) given that the editing tool has access to an SGML
>declaration and DTD that exists (or is generated) above and beyond that
>which is part of the XML definition itself

This has to be the way it is.  Some have envisioned an XML processor
writing out an SGML Declaration and DTD based on the XML markup declaration,
and in fact I have software that does this for MGML.  You can't ask for
help from an SGML editor without this stuff.

>II. Under the category of (b) above, assuming we continue down the road of
>defining XML in such a fashion that it does not include a complete DTD

No.  This is also the crux of what's bothering Len, and I think it's really
a non-issue.  I,and I think most people involved in this, absolutely definitely 
do envision having the equivalent of an SGML Markup Declaration.  (Although 
since XML markup syntax is frozen, the term "Markup Declaration" is a bit 
misleading - I prefer Schema Declaration or Grammar Definition or some such.)  
However, there is some sentiment in favor of a couple of differences between 
SGML and XML markup declaration, specifically:

 a) XML MD might have a different syntax, specifically that of an XML (SGML)
    document, and
 b) XML MD might be defined in such a way that it useful only when partially
    present.  (e.g. for definining entities, #FIXing attributes, and so on -
    think of a DTD subset without a DTD)

But your questions highlight a point that I don't think has been carefully
enough considered.  Namely, our design goals say XML shall be designed such
that it can be used in the absence of a DTD.  But with incomplete or absent
markup declarations, is it "really" XML?  Probably the standard has to have,
formally, two flavors of XML compliance.  For example, you could say "Valid 
XML" is an instance for which a markup declaration is available and that meets
the structure constraints of that MD; and that "Unvalidated XML" has 
proper markup syntax throughout but that in fact either no MD is available
or that validation has not been accomplished.

Cheers, Tim Bray
tbray@textuality.com http://www.textuality.com/ +1-604-488-1167