In a world where XML allows multiple
roots, I’m curious whether each root would potentially allow a different
encoding? Would multiple prologs be allowed? Would multiple be necessary?
We recently had basically the situation
described below. However, we were on the receiving end of a “batch”
of multiple XML “documents” stored in a single file. Each document
was separated by prolog. To fit this into our pipeline, we wrote a customer
XMLFilter (SAX) which understood that the file was really multiple XML files
and then generated events...the first of which was a dummy root. It worked
quite well but was an extra step nonetheless.
Matt Johnson
From: COUTHURES Alain [mailto:alain.couthures@agencexml.com]
Sent: Wednesday, February 27, 2008
9:55 AM
To: Richard Salz
Cc: Michael Kay;
xml-dev@lists.xml.org
Subject: Re: [xml-dev] Nested
Documents (was: XML 2.0)
I don't understand what you mean by "more tightly
linked"...
Having multiple roots is almost as having multiple documents, one after
another, in a single message or in a single file... ?!?
I have seen plenty of examples of XML documents which are nothing but a
collection of sub-documents and, in these cases, the root is meaningless, just
there to have a valid XML document. When you convert a CSV file to an XML
document, you have to create a root element, let's say 'Records', without
attributes, and as many 'Record' elements as lines in the CSV file. I think
'Records' is useless...
Alain COUTHURES
<agenceXML>
http://www.agencexml.com
Richard Salz a écrit :
You didn't respond to what I thought was the more
interesting point: multiple roots result in producer, intermediary, and
consumer being more tightly linked. Yes, as someone pointed out, you have
the issue of trailing comments and PI's, but I'll do a hand-wave and say those
are not common. And I admit I'm trying to brush off something I find
inconvenient. :)
/r$
--
STSM, DataPower Chief Programmer
WebSphere DataPower SOA Appliances
http://www.ibm.com/software/integration/datapower/
