Re: [docbook-apps] Docbook to WordML

["Christian Roth" <roth at infinity-loop dot de>]

Phil Weston [Mo, 26. Jan 2004]:

>A direct XSL transformation from Docbook to WordML must (in theory) be a
>more reliable process that relying on XML-->FO-->RTF-->.doc given that
>(in my experience) the RTF conversion doesn't support all features of RTF
>and RTF in itself doesn't support all the features you need to build a
>decent, structured Word Doc.

RTF includes essentially all structuring possibilities you also have in a
.doc file, though honestly, I don't know which additional structuring
layers Word 2003 resp. WordML introduces. What RTF is missing is some
meta-info, password protection and macros.

That said, XML-->FO-->RTF-->.doc is a problematic workflow since semantic
information (essentially style names in Word) is lost in the XML->FO
step. This is why we have created for our products a DTD that's similar
to WordML meaning that it expresses the structural capabilities of the
Word application resp. RTF, but in a very concise, non-verbose way (the
upCast DTD):

<http://www.infinity-loop.de/DTD/upcast/4.0/upcast.dtd>

It does not include layout properties as elements, as these are attached
to the basic, structuring elements using plain CSS if desired.

Our product downCast is an upCast-DTD+CSS to RTF converter. The advantage
in an

  arbitrary XML-->upCast-DTD (or similar)-->RTF 

workflow (as opposed to FO as intermediate format) is that:

(1) style names (and therefore, semantic information) is preserved in the
resulting Word document, which makes converting it back to some
reasonable XML - once edited - a lot easier,

(2) it still suffices to write the RTF generator only for one DTD

(3) you don't need to change the XML-->upCast DTD conversion step (i.e.
in general the XSLT processing sheet) to change basic layout, since that
comes from an external CSS stylesheet most users are familiar with and
confident in editing.

Another advantage is that with RTF, in contrast to WordML, you are
targeting not only Word 2003, but also legacy applications like Word 97,
Word 2000, and many custom components using the reasonably documented RTF
format as their storage/interchange format (like Mac OS X's TextEdit and
the underlying engine).

Best regards
Christian Roth

-- 
infinity-loop GmbH
Christian Roth
CTO
http://www.infinity-loop.de
roth@infinity-loop.de



To unsubscribe from this list, send a post to docbook-apps-unsubscribe@lists.oasis-open.org, or visit http://www.oasis-open.org/mlmanage/.