[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-docbook
Subject:    summary of po-doc situation
From:       Stephan Kulow <coolo () kde ! org>
Date:       2001-02-25 16:54:58
[Download RAW message or body]

	Thomas Diehl wrote:
> 
> On Sat, 24 Feb 2001 16:10:48 +0100, Francois-Xavier Duranceau wrote:
> 
> >> Well, that's the part we want to overcome with the doc-po thingy.
> >are you talking of using the po mechanism (merge...) for splitting the
> >docbook files? If yes is there any plan for an available tool? IMO this
> >is much more important than using XML, DocBook 3.1, 4.0 or whatever you
> >want. The fact is that we won't have more translaters and translated
> >docs til we have this new processing...
> 
> Hmm -- maybe you got temporarily unsubscribed? There were some
> extensive discussions on the subject recently. (Check out lists.kde.org
> and several threads wit "docs as po files" in their subjects.) I think
> Stephan is already playing around with some files and trying to find
> out whether it will be doable or not.
> 
Actually it might be more important than XML and the Docbook DTD. But SGML
is such a crappy format, that's almost impossible for me to parse it in a
decent speed (if you ever ran jade over all of kde-i18n you might know what
I'm talking about). So we _need_ the XML to do anything useful, but as
XML is only available for Docbook 4.1, we have to switch. 
Eric and I have already been doing quite a lot of work in that direction to
prepare this step and it will happen soon after Eric returned from his well
earned vacation :)

Thanks to the work of Wolfram I created quite some code already to split the
docbook files. But I don't have no good way to take current docbook translations
over as every little change in the markup can break this whole concept. And
most translations aren't in UTF-8 which makes it also impossible. 

So what I would like every translation team with translated documentation to
do is to transfer the docbook translations into utf-8 (you can safely ignore
that our current tools can't handle it gracefully - we're in HEAD) and check
using Frederik's sgmldiff tool if the markup is correct. You should definitly
not add or forget <para>s. 

In kdenonbeta/poxml (it will soon move to kdesdk) I added some tools, you can
use to verify some things:
1. xmlizer - which take a SGML document and creates a XML file out of it removing
   quite some stuff it can't handle (like the preamble - Frederik knows about that
   point and is currently checking to find a way that works with both SGML and XML)
   So don't commit any results of it
2. checkXML - which takes a XML document and checks it for correctness, you can
   call it with --push to pass a SGML document which is run through xmlizer automaticly
3. fixsgml - which corrects common mistakes I found in SGML documents. It can't
   fix all problems, but quite some problems that checkXML finds in SGML documents.
4.! split - which takes two XML files and creates a PO file with msgids from the
   first file and msgstr from the second file. I tried some examples from the german
   translations and even though they claim to be exact translations, there were quite
   some things missing - which only demonstrates the weakness of our current way to
   translate documentation.

So as you see, we could very well start already with translating, but I can't handle
more than german and already convinced the german team to take part in my experiment.

I really want to do some serious work with documentation and translation of it for
2.2 and will put my complete man power behind it. Remind me of that promise :)

Greetings, Stephan

-- 
Teigwaren heissen Teigwaren, weil sie Teig waren * J. v. d. Lippe
MFCH
_______________________________________________
Kde-docbook mailing list
Kde-docbook@master.kde.org
http://master.kde.org/mailman/listinfo/kde-docbook

[prev in list] [next in list] [prev in thread] [next in thread]