Re: ITS rules for OpenDocument

On Tue, 22 Jul 2008 01:06:24 +0900
Felix Sasaki <fsasaki@w3.org> wrote:

> 
> Hi Yves, all,
> 
> I spent some time on OOXML. See
> 
> http://www.w3.org/International/its/wiki/ITS_Translate_Decorator#Example_usage_for_OOXML_documents
> 
> With the example file I had it was easy to identify all translatable 
> text. However the XLIFF chunks I generated are not very useful without 
> information about text flow, that is "Elements within Text". 
> Nevertheless, comments very welcome.

Hi Felix,

I looked at the files in the link and found these issues:

1) The .docx file linked in the page can't be opened with Office 2007.
It says that a pre-release version of Office was used to generate the
file and it is invalid. I also tried to open it with OpenOffice and
failed (I have Novell's version which supports OOXML). Nevertheless, I
successfully opened it with Oxygen 9.3 and was able to check its
content.

2) The XLIFF file you generated uses has the "xl" namespace in all
elements. That's a kind of  "bad practice". The XLIFF namespace should
be the default one and XLIFF elements should not have a namespace
prefix. 

3) The segmentation applied to the OOXML file for producing the XLIFF
is wrong. You did not consider the internal format of OOXML files for
enclosing formatting information.  I placed an alternative XLIFF file
for your sample OOXML file,  with segmentation that considers inline
markup, at:

ftp://charmed.maxprograms.com/pub/alice-in-wonderland.docx.xlf

Regards,
Rodolfo
-- 
Rodolfo M. Raya <rmraya@maxprograms.com>
http://www.maxprograms.com

Received on Tuesday, 22 July 2008 02:00:25 UTC