It's certainly not a bad idea. But there is the complication of determining what language Ace should be using, though using the same language as the sample answer field would cover most cases.
The problem from my perspective is that even relatively simple changes take a lot more time than you'd expect. If you wanted the checkbox to be persistent (which it really should be - it would be annoying to have to re-click it on each save), you'd need to add a field to the database, which means defining a DB update-protocol. Then there's all the testing - new functionality needs new tests to be added to the test suite. And the functionality needs to be documented.
To be blunt, I can't myself be bothered as I don't really feel a need for that functionality. If you're keen enough to implement all that yourself, and if it doesn't clutter the UI too much, we would certainly consider merging a pull request.
The problem from my perspective is that even relatively simple changes take a lot more time than you'd expect. If you wanted the checkbox to be persistent (which it really should be - it would be annoying to have to re-click it on each save), you'd need to add a field to the database, which means defining a DB update-protocol. Then there's all the testing - new functionality needs new tests to be added to the test suite. And the functionality needs to be documented.
To be blunt, I can't myself be bothered as I don't really feel a need for that functionality. If you're keen enough to implement all that yourself, and if it doesn't clutter the UI too much, we would certainly consider merging a pull request.