If it's anything like most other translation programs out there, you're gonna get pretty garbled (Engrish to the extreme!) results translating either E to J or J to E. For the simple reason that both languages can depend highly on context. Even professional programs don't give results much better than babelfish, unless the structure of the stuff being translated is highly organized (think of official documentation or textbooks or academic papers, as opposed to conversation or dialogue).
For a translation program to translate well, it would not only have to come up with the literal syntactic and semantic possibilities but it would also have to *understand* the message being conveyed in order to determine the context and use that to fill in omitted words and select the *correct* possibility from the several options it comes up with. And at that point you might as well ask yourself why you're working in the translation field instead of the artificial intelligence field.