Meeting Notes: RE: [xliff] REMINDER: TODAY'S Informal Discussion: XLIFF - DITA interchange best practices

MHonArc v2.5.0b2 -->

xliff message

[Date Prev]
 | [Thread Prev]
 | [Thread Next]
 | [Date Next]

--

[Date Index]
 | [Thread Index]
 | [List Home]

Subject: Meeting Notes:  RE: [xliff] REMINDER: TODAY'S Informal Discussion: XLIFF - DITA interchange best practices

From: "Tony Jewtushenko" <>

To: "'XLIFF TC'" <>

Date: Thu, 15 Jun 2006 16:53:08 +0100

Attending:  Tony, Rudolfo, Doug

 

Topic: Open discussion on XLIFF-DITA
interchange best practices and related issues.

 

Tony> purpose of this meeting is to
brainstorm what the TC can and should do to provide guidance to those who want
to localise / translate DITA content, with potential recommended best practice centered
around XLIFF as intermediate container. 

 

Rudolfo> DITA is just the same as other
XML; Best practices for XML - XLIFF interop would be the same as 

Eg: Heartsome uses skeleton - translatable
content is pulled out to XLIFF, untranslatable remains in skeleton; do not use 

XSLT;  Bryan's XSLT roundtrip tool;
Challenge is Segmentation, one approach to use a second pass across XLIFF file
to create segmentation but it's complicated.

 

Doug: DITA not supported yet, by his
company but folks are starting to ask for it.

 

Rudolfo: Arguments are often made that
DITA without extracting to XLIFF, since can mark content in DITA as
translatable. However, there's no translation related metadata in DITA to
manage localisation process. Anecdotally there's now a growing trend of publishers
to process content into XLIFF and only send out XLIFF.  Vendors are
converging on accepting XLIFF.  

Complications regarding providing explicit
conversion profiles around DITA - XLIFF interchange: DITA files are managed in
a tree structure, intended to be used in a CMS, maybe file system. When
translating have to keep in mind how to create new tree structure. 
Handling the translated material in the CMS is complex as would have to keep
track of references / links and maps in the content and ensure its preserved in
the translated content.  Such complex management is really tools specific
and is not something that can be boiled down to in a best practices document.
So we should not say how to write XLIFF file from DITA - ie., depends on the
context, but should instead provide best practices through something like a
profile document.

 

 

RECOMMENDATION TO XLIFF TC: 

The best practice of interchanging data
between DITA and XLIFF is the same issue as how to interchange data XML and
XLIFF in general... e.g., how to use XLIFF with ITS,  XML:tm, etc. 
We suggest that the TC author profile documents similar to the other profiles
we've developed,  ie,  HTML profile... in them we recommend which
elements to use or avoid, skeleton file approach, transformation approach,
other common problems and recommended solutions.  Should address XLIFF
<-> ITS, XLIFF <-> XML:TM, XLIFF <-> DITA, maybe XLIFF
<-> Docbook.