Well, saying about Latvian pages it's of course a different situation, I know it perfectly. Yes, UTF-8 is a saver if you run forums in multiple languages at once. I've meant one language forums (I've mentioned it).
By double less space let's not also forget the traffic... of course the space on disk doesn't matter a lot, but the traffic double grows. Sometimes it's critical... specially if you have let's say 300 visitors per second ;-)
Anyway, it doesn't really matter if you choose UTF-8 or not. I just want to say that for about 80% of cases UTF-8 is not the most suitable solution. BTW, PHP6 is planned to work with UTF-8 texts, I don't know when it will happen actually :-) but it would solve many problems.
Regarding migration problems - if you know how to do it properly, it appears very easy. The main problem comes from mysqldump utility here, which (from some time) transfers all databases in UTF-8 by default. In the final result, the original encoding may be lost.
When setting up the site, you must just know precisely how to set up the database in the proper encoding. So if you run the site in windows-1251, the database and all related client/server connections should be set in this mode on server (cp1251). The same appeals to any other specific non-latin encoding. When you transfer the database, you must put the proper encoding flag to mysqldump or whatever exporting tool. Also, if you run the site in UTF-8, then the database itself and all client/server connections must be set up in this mode. But as I know, setting up the whole server in UTF-8 is a painful task. Not every admin is even supposed to do so.
The nightmare appears mySQL itself here. Not the thing for newbies I must say... |