On Sat, Nov 13, 2010 at 18:01, Nicolas Mailhot <nicolas.mail...@laposte.net> wrote: > I despair of making *nix input people understand that LANGAGE ≠ INPUT > Please stop trying to derive one from the other, they are *distinct* and > one can (and often does) use a non-english layout to type English. It's > about as smart as trying to find German people in Europe by searching > for Volkswagen cars. Sometimes it will be right, most often it will be > terribly wrong.
Yes, and it has nothing to do with system-wide or even session-wide settings IMHO. I'm a French guy living in GB. I type on French AZERTY or UK QWERTY hardware layouts, occasionally German QWERTZ. My software layout layout is always QWERTY US. I mostly use the en_US.UTF-8 locale, but some systems use en_GB.UTF-8. My timezone is Europe/London on my desktop, UTC in my servers and most virtual machines. And the one time I could really significantly benefit from a spell checking mechanism is when I try to improve my Spanish on #fedora-es. Only the application can often have lucky guesses or can be efficiently taught, unless one comes up with über-heuristics (for #fedora-es, the IRC client based on the channel I'm in). If you have good reasons to put language information in the input layer of the UI infrastructure, I'd love to hear which :) And I'd be really pleased if software kept letting me get rid of the Magic most might want. That's one big criteria when I pick my alternatives. Cheers, -- Pierre Carrier -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
