Technologically, I think one reason was that the data you needed for dealing with the Japanese writing system was much too large to do in software in the 1980s and 1990s. There are thousand of characters with different reading. en.m.wikipedia.org/wiki/Japanese_…