Hi Asgeir,
Thanks. I will send you JAR files early next week (one for the GUI version, one for the command line version). Then once I tidy up the Java, I'll send the source (it will give you a nice opportunity to print the code, display it at your office, and let everyone have a good laugh at my crude Java programming skills - not my strongest asset).
Your question and comments are quite useful. The xliffRoundTrip tool is meant to be a starting point. It is mostly driven by the XSLTs that the compiled JARs invoke. There, the user can (if needed) modify the XSLT for more discriminating trans-unit selection. Out of the box, the tool just selects any text() node in any element it finds, and puts it in a source and target. It nests in groups, and puts inline elements in g elements.
The ITS approach is a brilliant approach (it must be brilliant because it's been strongly recommended by some of the more brilliant people I know, like you, Yves, Christian, Felix, and others - not to say those are the only brilliant members on the TC).
Thanks,
Bryan
________________________________________
From: Asgeir Frimannsson [asgeirf@redhat.com]
Sent: Friday, May 15, 2009 5:23 PM
To: Schnabel, Bryan S
Cc: xliff@lists.oasis-open.org
Subject: Re: [xliff] Call for testers: xliffRoundTrip Tool upgrade to XLIFF 1.2
Hi Bryan,
How does the tool know what is translatable content and not?
It would be nice to have a tool that
1) uses ITS for identifying what is translatable/in-line etc..
2) does Rountrips from XML to XLIFF by using a ITS configuration and a supplied XML file
There are tools such as Yves' Rainbow that can already do this in Java, but it could perhaps be beneficial to have a tool that does *nothing but this*, specifically focusing on XLIFF.
I'm happy to help with testing source and binary.
cheers,
asgeir
----- "bryan s schnabel"