It seems it's common practice. We need another way for translators to follow the English files and know up to where they have translated. Maybe a simple HTML comment with the hash and then they can compare after that and see the diff.
Right now, many translations follow this practice which results in ~2x lines. All this does not come for free and it's bloating the codebase IMO.
Also, there's #2450.