W3C home > Mailing lists > Public > public-exi@w3.org > November 2008

RE: [EXI] Regarding CH event handling in FragmentContent NonTerminal

From: John Schneider <john.schneider@agiledelta.com>
Date: Tue, 11 Nov 2008 17:15:36 -0800
To: <santhana@huawei.com>, <public-exi@w3.org>
Message-ID: <C1E12ECEC3D94DA7A54367DB11DB9DA9@jcsdell8600>
An EXI fragment is a simple sequence of elements that may potentially
contain intervening comments and PIs. EXI fragments do not support
unparented text nodes, attribute nodes or namespace nodes. This is an
intentional design decision. Without the parent context there is no way to
determine the correct compression channel or type of an unparented node. It
might be possible to create an EXI-specific mechanism for preserving this
parent-derived information in unparented fragment nodes, however, this
information would be lost when round-tripping through text XML and would
create a more generalized and undesirable mismatch between EXI and XML.
    I hope this helps,


From: public-exi-request@w3.org [mailto:public-exi-request@w3.org] On Behalf
Of santhanakrishnan
Sent: Sunday, September 28, 2008 9:26 PM
To: public-exi@w3.org
Subject: [EXI] Regarding CH event handling in FragmentContent NonTerminal

        The EXI Primer it is mentioned that EXI is capable of mentioning the
WhiteSpaces in an efficient manner. It does indeed in certain cases 
1) with the presence of CH events in the Element grammar, 
2) Schema grammar with mixed=true 
3) non strcit Schema grammar etc...
But in the FragmentContent Nonterminal definition where there is a
possibility of characters between two global elements, it cannot be handled
by EXI. 
In the below example fragment, title and para are two global elements of
this book is about exi
<para>Efficient XML Interchange</para>

The character content between these global elements cannot be handled by
EXI. Do we need to insert a CH event in the FragmentContent Nonterminal
definition ? Please suggest on this.

Received on Wednesday, 12 November 2008 01:16:24 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:52:43 UTC