The full unicode title is present in the json and can be reproduced in the link. A server could index pages by full title and offer to retrieve them with something like a url friendly version of base64 passed as a query parameter on page fetch.
The example of distinguishing C, C++ and C# would work for that server and others like it but collapse back into the restricted slug space for everyone else. I would recommend this approach for a chinese federation.
A good start would be to provide the query parameter now to set the precedent.
Jon has brought this up several times.