Re: Architectural problems of the XInclude CR

>On Fri, 27 Dec 2002 19:08:32 -0500
>Elliotte Rusty Harold <elharo@metalab.unc.edu> wrote:
>
>>  I deduce that you want to have the MIME type override the parse attribute?
>
>In interpreting fragment identifiers, I believe that MIME types are 
>authoritative
>and the parse attribute is simply irrelevant.  The XInclude CR is 
>very wrong here.
>

How then do you propose to handle the use-case of including a 
text/xml or application/xml document as an example in a book like 
Processing XML with Java?

The MIME type may be authoritative as to what the document is, but it 
is not authoritative as to how the local process wishes to treat the 
document. If my process (such as XInclude) wants to treat an 
application/xml document as plain text, it should be allowed to do so.

I don't think MIME types and the value of the parse attribute are at 
all the same thing. The MIME type tells you what the server asserts 
the document *is*. The parse attribute specifies what the local 
process wishes to treat the document *as*. These are two very 
different things, and they do not need to be the same.
-- 

+-----------------------+------------------------+-------------------+
| Elliotte Rusty Harold | elharo@metalab.unc.edu | Writer/Programmer |
+-----------------------+------------------------+-------------------+
|           Processing XML with Java (Addison-Wesley, 2002)          |
|              http://www.cafeconleche.org/books/xmljava             |
| http://www.amazon.com/exec/obidos/ISBN%3D0201771861/cafeaulaitA  |
+----------------------------------+---------------------------------+
|  Read Cafe au Lait for Java News:  http://www.cafeaulait.org/      |
|  Read Cafe con Leche for XML News: http://www.cafeconleche.org/    |
+----------------------------------+---------------------------------+

Received on Sunday, 29 December 2002 18:40:00 UTC