Seems to me that there's room for improvement in this article. For example, the author does not try to compress any written language samples. If you wanted to find a "most efficient" language, you'd compress samples, maybe normalize them with the uncompressed length because it's hard to get the same thing written in different languages. My guess would be that natural languages would compress quite a bit, with English close to the least compressible. Artificial languages like Ithkuil and FORTRAN would be much more compressible, putting lie to Dembski's assertion that "natural" information is more likely to be incompressible.
EDIT: I think English would be least compressible. But I suspect that all written languages would compress quite a bit.
EDIT: I think English would be least compressible. But I suspect that all written languages would compress quite a bit.