But also in this case, the sheer charm of some of the less-than-wholly-satisfactory translations was such that I did the post. I hope it’s clear to the original author (if he sees this) that I’m not denigrating what appears to be an excellent rebuttal.

I wonder whether we’ll see truly satisfactory machine translation within my lifetime. (That’s “wonder,” not “don’t believe.”) There are two fundamental problems–one of which may be a bit less severe for scholarly resources:

1. Different languages have such different constructions that effective translation can be difficult. (In the example cited, at least one marvelous result is, I believe, a case where typical German sentence order results in a different effect in English.)

2. (“Why a roundtrip translation for Cites & Insights or Walt at Random material might be particularly gruesome.”) Idiomatic language is, I suspect, much harder to translate effectively, as is colloquial language. I love good idiomatic American, and try to use it, but it can’t be easy to translate, and I’m sure the same is true for other languages.

The one tendency of Google’s translate that I found bemusing, rather than amusing, was its tendency to turn “Open Access” into “open ACCESS.”

Serious progress in machine translation will be as revolutionary as serious progress in open access, and I look forward to both for similar reasons. They remove different access barriers and together could create effective no-barrier access to whatever we may want to read.