Using desktop applications is still a nightmare, where opening any document from the internet or navigating to a foreign page is hit or miss.
Also, the main frustration is not that it happens often; it's that when you encounter it, fixing it may involve studying the whole software stack to find out where the setting is switched or at what point in the toolchain the format was mishandled.
It also often happens with pages in Chinese or Japanese. But then you bring another problem: when a document is shown as a list full of ▯▯▯▯▯▯,▯▯-.▯▯▯, there's no way to tell if the problem is with the encoding or the fonts, and no obvious way to know which font should be installed and what part of the software stack is the one responsible.
The whole field of displaying documents to end users could benefit from some user-centered design to build tools that helped users to fix these situations, based on some common heuristics to solve the most frequent problems.
I don't know, maybe it's me who's doing something wrong and turn to using only a small subset of software tools, properly configured?
But I do need to process documents from many origins using a variety of different tools; I've never found a good programable multimedia editor that satisfies all my needs (something like emacs but with WYSIWIG capabilities).
It got much better in the last years as UTF-8 gained wider adoption. The problems are mostly with legacy systems in government. (Cannot comment on that since my name fits in ASCII.)