The problem you mention is one of the common reasons. PHP doesn't handle UTF-8 well. I run a phpBB forum, and if you try UTF-8 the search function doesn't work with the multi-byte characters, nor does the text highlighting code. So I stick with ISO-8859-1 (in my case).

However, I think the biggest reasons for not moving to UTF-8 are the tools (read Windows text editors), a general lack of knowledge about character encoding, apathy and inertia.

Japanese can be quite problematic with UTF-8. Some sites are using it successfully but you've really got to know what you're doing in terms of the server setup. If you are on a shared server / virtual host it's doubtful they will have all the mbstring modules complied into PHP that you'll need.

Shift-JIS is a good encoding to use for a Japanese site.

Thanks a lot!

But really, theses answers come as a big surprise.
During all the past years, I've always used ISO-8859-1 for latin text and Shift-JIS for the Japanese ... and am still using them.

But recently, I've found many web articles pushing webmasters toward the UTF-8 trend because it's supposed to be "safer" and "more efficient" for multilingual sites. And you can display multiple languages on the same page.

I started to think that I was doing a very bad thing in sticking with the "good old encodings".

Where is the truth then? After some research, I came to the conclusion that UTF-8 is good for western sites. As for the Japanese sites, EUC-JP seems to be prefered to Shift-JIS by a huge majority ... and I still don't know why.

We're having pretty much the same discussion in your other thread in the Asia/Pacific Forum: EUC-JP better than SHIFT-JIS? [webmasterworld.com]

UTF-8 works pretty well for Western character sets. It's the double byte languages like Japanese and Chinese that still have issues.