xml-dev - Re: Feeler for SML (Simple Markup Language)

["Liam R. E. Quin" <liamquin@interlog.com>]

There was a "tinyxml" proposal, I think by Dan Connolly but I
might have misremembered, during the development
of XML.  I had hoped it would succeed, but the overall
feeling was that we had to have a single unified front,
no optional XML features.  I think that was the
right decision.

On the other hand, there are a number of features in XML that
at the time I opposed, or that I now wish I had opposed :-)

Water under the bridge collects in stagnant pools at times,
as len might say I suppose :)

I've contributed some ideas below, and I think in some ways they
are closer to whatthe XML WG might have come up with had namespaces
and schema proposals been around at teh time, and  had the concept of
getting rid of so many featuers been a feasible one.


On Thu, 11 Nov 1999, Don Park wrote:
> o No Attributes (ouch!)
almost, see below
> o No PI, Comments, Notations, or CDATA sections
agreed 100%
> o No document type declaration
OK..
> o UTF-8 encoding only
hmmm, Latin 1is useful thouugh.   Not sure on this one,
but I could be convinced.
Agreed -- they are pretty poorly defined anyway, if you read
the spec carefully.
> o No non-character entity references
> o No predefined character entities (I am iffy on this one)
You need a way to escape & < and > so maybe these are worth
keeping.  However, seem y comments below.

I would like to suggest a convention for using a
minimum subset of XML, so that all "SXML" documents would also be
well-formed XML documents, and, with the addition of
a DTD or Schema, could be valid.

One way to do this would be to define your own element and name space:
SXML, say.

Suppose that an SXML document had a head and a body, but that the body
could contain whatever XML you wanted.

The head would contain, optionally,
* definitions of repeated fragments
* metadata about the document
* an optional reference to a Schema, so as not to have
  to use processing instructions or comments
* an optional erference to one or more style sheets, again
  avoidig these silly processing instructions

All the references would use XLink, unless XLink  ends up not supporting anything as
simple as an href attribute, in which case you might as well
use an href attribute.

Getting rid of DOCTYPE means you have no way to define yuor own entities,
and in that case might as well get rid of the entity mechanism
altogether.
For inclusion, use id/idref topoint to an object defined
in the header:
<SXML:HEAD>
 <SXML:OBJECT NAME="Gribble">My name is Gribble</SXML:OBJECT>
</SXML:HEAD>
<SXML:BODY>
 <P>Ths is Gribble. <SXML:USE WHAT="Gribble" /> </P>
</SXML:BODY>

An SXML parser wold not need to understand namespaces, only to allow
colon (:":) in identifiers.

Lee

--
Liam Quin, Barefoot Computing, Toronto;  The barefoot agitator
liam at holoweb.net [note changed address]
Ankh on irc.sorcery.net, http://www.valinor.sorcery.net/~liam/


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo@ic.ac.uk the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)