2009-10-02, 07:16 AM
Well, I have been around for some time also, but I never had to dig into EBCDIC luckily. I started programming in C on Sun workstations in 1989 (UNIX), avoided windows programming altoghether and took the plunge to Java in 1998. Been there since.
Yes, proper i18n handling is a messy business, even if you just want to support one character set, e.g. swedish chars. There are many implementations that may conflict. And a few stupid reasons too, like the windows codepage (850 is it?) that is similar, but not equal to is 8859-1. So most of the time it works but there are spcial cases where it fails. M$! :mad:
When we come to web there is two proper ways to encode swedish, iso 8859-1 or utf-8. But you need to keep your selected encoding style consistent through the code and the database or you end up with those errors again. I've been there last week correcting such a bug
Well, there is only one proper way to handle it all and it is going all utf-8. That is the path for our coding here now, i.e to convert all old iso-8859-1 encodings to utf-8.
Yes, proper i18n handling is a messy business, even if you just want to support one character set, e.g. swedish chars. There are many implementations that may conflict. And a few stupid reasons too, like the windows codepage (850 is it?) that is similar, but not equal to is 8859-1. So most of the time it works but there are spcial cases where it fails. M$! :mad:
When we come to web there is two proper ways to encode swedish, iso 8859-1 or utf-8. But you need to keep your selected encoding style consistent through the code and the database or you end up with those errors again. I've been there last week correcting such a bug

Well, there is only one proper way to handle it all and it is going all utf-8. That is the path for our coding here now, i.e to convert all old iso-8859-1 encodings to utf-8.