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

RE: CDATA by any other name... (was The raw and the cooked)

• From: "Rick Jelliffe" <ricko@a...>
• To: "XML Dev" <xml-dev@i...>
• Date: Wed, 4 Nov 1998 04:41:00 +1100

> From: John Cowan

> True.  However, since the standard encodings of XML include all the
> characters there are (and if they don't include yours, just you
> wait, 'Enry 'Iggins), that isn't as much of an issue.

(An optimistic view of ISO10646: there are dozens of new Han ideographs
created every day, apart from other scripts.)

The situation I am thinking of is, for example, where I am creating an XML
document which will be used, after processing by a non-XML Macintosh
application that only understands MacRoman. The CDATA marked section is the
only constraining/signalling mechanism in XML which could be applied, and it
goes without saying that it is a pretty poor one, but I don't want to say it
is useless.

If the consensus of developers is that they dont want to allow marked
sections to be used in this way, I hope that the schema people will look at
a solution for constraining strings to use certain repertoires of
characters. I believe the Balise parser and SGML/XML processing system has a
"sanity checking" option for names in markup for this kind of
repertoire-limitation purpose.

> The current XPointer draft allows PIs to be referred to on equal
> terms with elements (except for not having a GI or attributes or
> sub-elements).

> The DOM has a ProcessingInstruction node, though pseudo-attribute
> parsing is not performed.

Which is my point. By not even providing some minimal kind of token-locating
within a processing instruction for people to use if they need it, PIs are
barely useful as far as I can see.  People will always try to use what is
provided, rather than extend an API, so it is almost the kiss of death to
PIs except for the more sophisticated applications. When schemas come along
with better lexing of attribute values and PCDATA, I wonder if they will
also bother to allow scanning of the PI into tokens too.

Rick Jelliffe


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@i...
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo@i... the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@i... the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@i...)

• Follow-Ups:
  • Re: CDATA by any other name... (was The raw and the cooked)
    • From: John Cowan <cowan@l...>

• References:
  • Re: CDATA by any other name... (was The raw and the cooked)
    • From: John Cowan <cowan@l...>

• Prev by Date: RE: Z references (was Re: DTD,s)
• Next by Date: Re: XML APIs
• Previous by thread: Re: CDATA by any other name... (was The raw and the cooked)
• Next by thread: Re: CDATA by any other name... (was The raw and the cooked)
• Index(es):
  • Date
  • Thread