This allows user to see GitLab and translate the externalized strings inline. The translations created will still need to go through review in Crowdin and so on, but I think it will make translating faster, easier and will result in higher quality of translations.
As a member of Ukrainian translation team I must say that this feature is essential for us at this stage.
The thing is that currently all strings that need to be translated are dumped into a single list in Crowdin mostly without any context information. This approach worked well enough a few years back when there were not that many items in that list. However currently there are more than 10k strings there and the number is steadily growing.
Having context is crucial for accurate translation. And currently in order to get that we need to jump through some hoops. For instance in order to correctly translate "issue" it's necessary to understand whether it's used as a verb or as a noun and the only way to do this is to:
look through the source code of Gitlab
find the page where the word is used
figure out the logic of rendering of that page (which is not always trivial because of includes, variables and if-statements).
and then proofreaders have to do the same exercise.
This approach is very slow and requires translators and proofreaders to have coding skills as well as a great deal of patience.
Enabling this feature will:
allow us to produce better translations
widen the group of people that are able to contribute
dramatically lower amount of time spent
Finally, I can only speak for myself but unless this feature (or something similar) is enabled I'll have to leave the translation team because I don't see a point in going through the very inefficient translation process that leads to sloppy results.