I'm not completely sure what the actual question is, but I'll try and explain why I think your browser is seeing garbage when using Perl: if you are using Apache/mod_perl (or even Apache without mod_perl, I suppose) to talk to Oracle then you'll need to set NLS_LANG in your httpd.conf.
rdfield
Comment on Re: UTF8, Perl 5.6 and Oracle Revisited....
Hmm, that is something to think about. We are using Oracle running on a Sun Solaris Server. The Perl implementation on it is 5.6 but I do not think it is Apache Mod_Perl. However, any hint that might help me solve this issue is welcome. I will look into it. I am also thinking that the ISO-2022-JP code could be damaged as happens from browser so we are applying the text cleaning techniques from Ken Lunde's excellent book.
The question is not clear. I apologize. I am trying to to find out how to make sure I get good output file from the database that if I bring up in Japanese Word, or some Japanese text editor, will come up in Japanese. Only the browser is consistent with this. I did apply all the help from my previous question and that allowed me to get the raw UTF-8 data from the database, but it is not viewable as Kanji characters unless I put into web browser. The data is not the HTTP unicode #& format though so I know this is not issue.
I hope that clears up some. And thank you very much for your response. :-)
I've noticed problems viewing UTF-8 data in Word97, but the same data looked OK in Word2000. Generally speaking, all the language/UTF-8 work I've been involved with was via XML and the browser was the only was they only way we could reliably view the data. Word2000 uses some of the same libraries (another Monk may have more information/knowledge in this area...?) as IE and is much better than previous versions at displaying XML/UTF data.