The “encoding problem” for Asian languages is real, but it goes back earlier than that, as it emerged with typewriters. See Thomas Mullaney’s “A history of the Chinese typewriter”, and Nanette Gottlieb’s “Word Processing Technology in Japan” for more. By the late 90s/00s it wasn’t a significant primary factor (that is if it was a factor, it was due to cultural inertia, and not lack of access to technology itself).

I think the problem here is looked at backwards. It’s not that Japan sucks at software - it’s that the US/North America got extremely good at software in the last 20 years (by a combination of factors that feed into each other, as always: the best research labs are in the US, the tools are built in the US, the companies that make use of them for competitive advantage are in the US, the economic context most favorable to founding such companies is in the US, etc). The rest of the world is catching up, but the gap is still visible.