It's true that shutting down such service which was based essentially on work made by a large worldwide community is a bad sign. At least Twitter should release all their current efforts in an open-sourced repository with an open licence without commercial restrictions, possibly with attribution (BSD-like, CC-BY) or with some basic share-alike or copyleft requirements (ODbL, CC-BY-SA, GPL, LGPL), or left completely free (CC-0, Public domain, WTFPL).
Shutting down the service compeltely and reasing the data would be an insult to those that participated and from which Twitter benefited a lot.
We can understand that Twitter no logner wants to support it directly itself and wants to save some costs, but there are certainly interesting data to keep which could be reused in a possibly larger project as a translation memory and corpus useful for many other free (of commercial) translation engines. It's up to Twitter to choose their opensource repository provider (such as GitHub), make the decessary licencing requirements, and possibly open a competition to transfer the administration of the project to some organized open groups (if the project does not have to be left in readonly with no further development and can be resurrected by independant groups creating their own branches or integrating the data in their own corpus). For that, Twitter should ask their community about the pros and cons of several candidate licences.
And may be Twitter will continue to be able to profit from the resurrected or derivated projects (but it will jsut no longer be alone to use it).