Charles de Miramon wrote:
Abdelrazak Younes wrote:
Xml is slow to parse,
That's a generalisation that is not really true, it really depends on
your XML syntax. We are free to do what we want, so I am pretty sure the
parsing will be as fast as nowadays. Really, do you see an increased
difficulty when parsing<document>...</document> instead of
\begin_document...\end_document ?
The slowness of XML parsing has been a complaint for KWord, OpenOffice,
MsWord2007.
Those are different more complicated beast that have to care for finger
painting, page positionning, reference linking etc. We don't have this
kind of information in the format and that's why our XML format will be
simple, very simple. As a first step it would look a lot like the
present format (just replace \begin...\end with <>...</>). In the
future, it could use a minimal subset of odf. But there is no way in
hell that our format is going to be as complicated as odf, simply
because the WYSYWIM concept freed us from that problematic.
If you cannot use a parser from the shelf (QtXml, LibXml2) but
have to tweak something, what is the point.
We'll see. If using on the shelf library is too slow, we'll just keep on
parsing the file ourself and it will be fast.
Because XML is easy to transform. That is really the *main* reason. For
example the transformation to TEI will be very, very simple, believe me.
Not only that, but also transformations to HTML, ODF, etc.
If it was true, there would be today rocksolid ODF<-> Docbook, ODF<-> TEI,
ODF<->XML-FO converters. The idea that is very easy to transform one xml
jargon to another xml jargon, is a myth.
I understand that but see above. The reason why it will be simple in our
case is precisely because our XML format will stay simple.
Heck, even a
transformation to LateX would be possible if one day we decide that it
is not worth maintaining the LateX code in C++.
At some point in Koffice development, Robert Jacolin wrote a XSLT converter
between KWord first XML format and LaTeX. He left the project but nobody
continued its work because everybody hates XSLT.
Or just that nobody found an interest in learning it, or that no
developer was interested in LateX. XSLT is quite easy in my experience.
XML is nice for archival, interchange, standardized formats but is not an
universal swissknife.
Oh, I agree. I tried to understand once the odf format but I got a head
ache :-)
Abdel.
