On Translia, for all file types that the source text are already segmented, say TTX or BIF (Trados files), their segmentation will be followed as is. For .txt file, the source text is divided into segments by paragraph. Here a "paragraph" means a text string that is separated with return carrier or line break.

At present, we don't add the feature for translator to customize the segment length. It may reduce the match rate when we have TM available in near future. While, we know well sometimes we translators think the segement is too short, which makes our translation is not fluent enough; or it's too long which gives us less "sense of achivement". With pors and cons considered, we use paragraph as segmentation unit for txt, but willing to adjust it based on the feedback of all of you - when we feel better, we can do better translation :-)

Regards,
Yolanda