So you're basically saying it's better to re-implement the API, with everything it implies:
* examining the docs
* taking care of error handling
* coming up with an easy to use developer interface
* keeping it updated if the API calls were to change
every time you need translation.
You're also implying it's better than having a package that can potentially abstract many different translations API, that you can just add as a submodule of your git repository, and use with one `require()` call away.
I used Google translate API for the last year or so. Implementation took an hour or so. They haven't changed the API, there are no other translation API providers that come close to Google. Wrapping it an abstraction layer is just wasteful pointless idiocy.
If you enjoy wrapping things in extra layers, all power to you. Enjoy the extra complexity and wasteful code.
* examining the docs
* taking care of error handling
* coming up with an easy to use developer interface
* keeping it updated if the API calls were to change
every time you need translation.
You're also implying it's better than having a package that can potentially abstract many different translations API, that you can just add as a submodule of your git repository, and use with one `require()` call away.
I disagree.