Alternative to [i2auto] Tag

Was it really such a performance hog that it needed to be removed?

Yes, there was a HUGE performance cost. When the spreadsheet only have a couple thousand terms, is not noticeable, but as it grows bigger, it was taking up to several minutes importing/exporting. (if I recall correctly, the change made it 70 times faster)

The problem was that the WebService can't do string manipulation as fast a normal desktop application, so it was taking a lot of time just to parse the terms and then to assemble them!

Do you you have any ideas of a good alternative method?

You can keep two spreadsheets, one for development and one for releasing.
Normally, you link the development one, and work on it.

Translators should be the only ones authorized to modify the release spreadsheet. They will copy from the development spreadsheet into the release and approve the changes.

The final game links with the release spreadsheet. So that its easy to verify if a translation is missing (there is even an option to display a warning on all labels with a missing translation).

Then, after release, you can keep using the development spreadsheet, and those changes will not be visible in the released game. Unless the translators move the changes manually, at which point the live-auto-sync will automatically download the new changes into your game.

That solves two issues: translations needed to be approved by the translators and that you could continue developing your translations after the game is released without fear that temporal translation get automatically downloaded to your released game.