Yes, but it makes it incompatible with most XML tools, which suppose that 
formating is either automatic ( <?xml-stylesheet ),  in the tags (like the 
"id", "class" and "style" attributes of XHTML with CSS)  or inline (like 
for example <b> element in XHTML.)

Yes, KWord's maindoc.xml file is in XML, but not in a way that permits the 
use of any extension of XML, like XSL for example.

The main problem are the "pos" and "len" attributes of <FORMAT>.
If a citation program for example, have to change a two letter sequence by 
a six letter one, you have to change all <FORMAT> of that paragraph that 
points at the first character of the two letter sequence or later. In 
another word, you have to know exactly KWord's format, as this relationship 
is not in KWord's DTD. (I am not meaning in the comments!)

Compare to the same in (X)HTML:

<p>One Two Three <b>CC<b> <i>Four<i> Five</p>

into

<p>One Two Three <b>DDDDDD<b> <i>Four<i> Five</p>

Here I have just to change the text, not any formating! I could even have 
made it automaticly with a program so simple as sed.

And that is making citation programs that are not plugged in KWord so much 
difficult to develop: you have to understand a big part of KWord's file 
format to work correctly, not just the <TEXT> element.

References: http://www.w3c.org (XML, XSL, XHTML, HTML, CSS)

Have a nice day/evening/night!
-----Original Message-----
From:	Thomas Zander [SMTP:zander@planescape.com]
Sent:	Wednesday, April 25, 2001 3:00 PM
To:	koffice@max.tat.physik.uni-tuebingen.de
Subject:	Re: Hopeful suggestion for kword

>
> First KWord's file are tarred and gzipped but the most important problem 
is
> that the formating of the text is defined using character positions. So 
for
> every change of text you make, you have to adapt the formating 
information.
>
> And completely changing KWord's file format will not be done for long (if 
> it will be done!), as the problems of KWord 0.8 have made lose many 
months
> of development. So it is not the time to break any important think like a 
> file format now!

The way the markup is stored is correctly from an xml POV, the thing is 
that
in the file you want to seperate context (the text) from markup (things 
like
<b> etc).

This makes it possible to search better, to change formatting without 
touching
the content etc.

The file format is correct, I see no need to change it.

--
Thomas Zander 
                                           zander@earthling.net
The only thing worse than failure is the fear of trying something new