[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Dangers of De Facto (WAS RE: Dangers of Subsetting?)

• From: Rick JELLIFFE <ricko@g...>
• To: xml-dev@l...
• Date: Sat, 11 Nov 2000 04:04:28 +0800

"Bullard, Claude L (Len)" wrote:
> 
> I probably could craft language that cites ISO
> SGML then recommends use of the W3C XML specification
> citing the benefits such as lower cost, ease
> of access to trained resources, etc.
> 
> Problem is getting them to accept it.  If this
> were the RFI stage, no problem, but in the RFP,
> such is discouraged.  It will come down to how
> receptive the customer is.  Really, this is all lawyerese
> and contract namespace issues, but experience
> shows the problems of not having a clean
> "namespace with a record of authority" to operate in.
 
ISO 8879 Annex K adds to the SGML declaration an optional parameter
with keyword SEEALSO.  This references (e.g. by URL) an "additional
requirements" document in which any constraints above and
beyond SGML can be noted.  Annex L gives an example of such
additional requirement: XML.   (The markup declarations, the SGML
declarations and the additional requirements allow a complete
specification of the document type definition.  Common XML
core is another example of exactly the kind of thing that can be
appended to an additional requirements document. )

This allows people constrained by policy to talk SGML and do XML.

I proposed SEEALSO for two reasons: first to allow profiles such as XML
to be documented and referenced formally, and second to provide an
escape
from the ISO approach before then, which was to take every 
requirement, try to generalize it, and then add switches to
the SGML declaration.  There is no reason to expect that XML
will not have a successor sooner or later, and even if it not
a subset of XML it will still benefit some by coming under the aegis
of ISO 8879. (Is aegis the word I want?)

There should be no more contractual problem with specifying that
the XML industry profile should be used (as an example of
additional requirements) than specifying that a document type
should use an industry-standard DTD or non-ISO public entity set
or particular SGML declaration.

Cheers
Rick Jelliffe

• References:
  • RE: Dangers of De Facto (WAS RE: Dangers of Subsetting?)
    • From: "Bullard, Claude L (Len)" <clbullar@i...>

• Prev by Date: Re: Why not canonical parsers? [Was: Pull-based XML parsers? ]
• Next by Date: Re: Options in XML 1.0
• Previous by thread: Re: Dangers of De Facto (WAS RE: Dangers of Subsetting?)
• Next by thread: Re: Dangers of De Facto (WAS RE: Dangers of Subsetting?)
• Index(es):
  • Date
  • Thread