Some text editors don't display those characters correctly, so when copy/pasting from the edit window into an text editor and back again will lose the characters. The same can happen after exporting pages and editing them and importing them again.
It's an extreme edge case of course, but one I've run into. (I just checked and å works fine)
The RTE normally converts those HTML entities back into the extended characters anyway, so my manual replacements on some pages didn't last long.
I guess we'll see after they fix the current editor problem exactly what the intended behaviour is.
edit: Oh, and it's a habit I picked up years ago - I don't know if it's a good practice, but I routinely just encode HTML entities when storing strings to a database to avoid possible problems.
Kinda offtopic, but when I was working on a file hosting site, I discovered that PHP doesn't allow writing some extended characters in windows filenames, but can handle the strings in the database just fine - so I needed to store the files on disc with a different filename. I still haven't worked out a solution to this problem, but I haven't really thought about it in a while...
Thanks. I was just wondering as I've setup my AWB in an effort to 'fix' the rte issues we see of late, and decided (on a whim) to change pretty much every common symbol used as well as letters with various accents/umlauts in case rte decides to pull another stunt like this again. It had never occurred to me that there could be issues down the line, or why they might be useful.
Yeah, AWB. Wikipedia has a somewhat lacking guide on it's features - lists them and generally assumes you know what each does - but it's simple enough to get the hang of for most parts. I tend to use it for template replacement/category (addition/removal/renaming) but it has all sorts of stuff built in. The general downside is that it's engineered for Wikipedia and lacks stuff we might use that they don't.