I'm not going to address the question of whether your files would be
TEI-conformant calling a DTD elsewhere, except to say that if they aren't,
then there's an awful lot of non-conformant TEI produced on a regular basis
by readers and contributors to this list!
As for your earlier questions, about XML rendering in browsers: The basic
reason for the discrepancies you are seeing have to do with various
different decisions made by the different vendors (or rather, their
developers) over how to conform to various parts of the XML Rec, which
specify how a DTD is accessed, when, and for what. The short version of the
story is that an XML parser doesn't even have to process a DTD at all, if
it aspires only to be a "non-validating parser" -- which leads to various
problems and complications having to do with the fact that sometimes
parsing and application of the DTD to an instance (irrespective of whether
the instance turns out to be valid) makes a difference in the data that is
seen in that instance (for default attribute values, for example) -- and
that whether this complication applies to a given case can't always be
known up front (it being impossible to know what a DTD declares without
looking at it).
The architects of XML processing systems have to decide on whether and what
processing overhead they are willing to accept given the particular design
tradeoffs they are facing for the class of applications they are
targetting. What makes this worse is that the boundaries of this problem
for them are not evident to us -- for example, IE's behavior seems to be
programmed in light of the application of its XML engine (MSXML) outside IE
as well as within it, whereas Mozilla's seems to be considered to be more
of an integrated module that won't be reused outside the browser for tasks
that browsers don't ordinarily have to do, like validation. So while we can
see that the two vendors have dealt with the problem differently, we can
only guess at why.
The bottom line for all this is that you simply can't expect XML parsers,
and even parsers in browsers, all to behave the same way.
To cut this Gordian knot, when many of us want to make XML available to
clients over the Internet, we present only normalized versions of our XML,
which are scrubbed altogether of DTD-dependencies. This means all entity
references are resolved and attribute values expressed -- and the document
is delivered with no DOCTYPE declaration whatsoever, thus requiring nothing
of even the most enthusiastic parser in the way of DTD processing. Such a
document loads much more nicely in the widest possible range of XML
parsers, which can all deal with at least this much.
I'll leave it to you to decide whether such a file can be "conformant TEI".
Or to the priesthood reading this list to rule on it. (I'm more like a
renegade monk, I'm afraid.)
Wendell Piez mailto:[log in to unmask]
Mulberry Technologies, Inc. http://www.mulberrytech.com
17 West Jefferson Street Direct Phone: 301/315-9635
Suite 207 Phone: 301/315-9631
Rockville, MD 20850 Fax: 301/315-8285
Mulberry Technologies: A Consultancy Specializing in SGML and XML