Beni Cherniavsky wrote: > The main point of the size argument is not HD size however - it's > cache usage and processor/memory bandwidth which translater to speed > of text processing. Not that I find that a convincing argument, I > find it hard to imagine a text-intensive application whose performance > is critical enough, except for parsers - but most syntaxes out there > are still ASCII ;-).
Maybe a word processor when you are writing a book? And when you have a small-memory computer. :-) But I am finding my arguments a little weak. In maybe 5 years' time (or less), the double space occupation of UTF-32 may become a non-issue, as are Western users accustomed to the UTF-16 based Microsoft Office. So maybe the strongest argument for UTF-16 is compatibility with existing software, especially when you have binary interfaces between them. Though, it is really a pain that wchar_t have different meanings bwteen Windows and Linux. It really makes cross-platform life harder. Best regards, Wu Yongwei -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
