RE: Reliable endpoints for XML docs (and doclets)

Boris Moore (Boris.Moore@rivcom.com)
Fri, 6 Mar 1998 22:12:38 +0100

Joe Lapp wrote:
>Tom Bannon wrote:
>> Perhaps I've missed something, but I don't see that the start
>> and end points of XML documents are well defined, ie. something
>> like <xml> and </xml>, akin to HTML's <html> and </html>.
>I wonder whether Tom is remarking on how DTDs and PIs aren't
>themselves enclosed in tags. He seems to be concerned about this
>because he wants to embed XML in HTML. Seems like a good question.

In our work we have a Netscape/IE plug-in which receives downloaded
XML and stylesheet data, and sends a resulting HTML stream to the
browser rendering engine. We have certainly found it useful on occasion
to include sections of HTML data which are skipped by the XML parser
and sent without modification to the browser. Our browser is set, for
the moment, to accept both <HTML> </HTML> sections within XML,
and also <XML> </XML> sections within HTML sections.

It would seem to me that when browsers are able to operate in either
HTML, or XML + XSL modes, then being able to switch modes within
a document (being able, in effect, to embed in either direction) could
indeed become quite relevant and useful.

Boris Moore
RivCom
"Publishing Structured Information"
www.rivcom.com