First, check if you have Episerver Languages Add-on installed. if not, please check my related blog.
Then, you can export a bilingual XLIFF as following: From the add the pages you selected to a translation project and then download the translation package.
I checked these steps and I got an XLIFF file with source/target included content translated or semi-translated. Please check a print screen from this link:
Please let me know if this solves your issue. But, in fact, I would suggest a better approach. Why not export the whole page without a sifting process. Your translators should be using Translation Memories (TM) where all your previous translations are saved. Once the translator Analyzes the newly exported files, they will be able to Identify what segments need to be translated so you pay only for the new untranslated content and not the whole exported files again. I believe this approach will save you time and reduce the manual process.
Thanks for the response.
I have no issue exporting and importing xliff for translation, and I do know the exported file contains some translation. However, those translation are from the file that you imported last time. This mean, if somebody makes a minor change to the page directly on Ektron, the change will not reflect in the xliff file you export. In other word, the translation in the exported xliff is not the lastest and greatest version. And if it's not the latest and greatest version, it's really not very helpful for the localization process.
Also, although I believe our vendor has a TM for our translation projects, they don't handle our reviewing and editing process. Therefore, the TM they are using does not match the TM we manage onsite. (We do send our TM to the vendor periodically.) That's why I'm desperated to find a more efficient way to align text from Ektron.
I am trying to update our translated web pages as our English website has some updates. These updates are minor (i.e. adding a sentence or deleting some words), so obviously, it won't make sense to send out the whole page to the translator again. The goal is to sift out the content that has been changed in the CMS and send that to our translators. What I need for the sifting process is a bilingual xliff that contains "not yet updated English content" which corresponds to the "translated content" in the CMS. This way, I can compare the bilingual xliff to the updated English xliff and get the content that has changed. At the same time, leveraging the old translation we have done before.
The existing features in Ektron do not support my need, and I believe this is a very common issue every localization specialist will run into.
The bilingual xliff will make the translated content more scalable and more convenient for people to manage their translated website.