The "current situation" is worse than before the change made about a year ago. The web site is all in UTF-8, but the database remains in some ancient codepage. The script performs a mapping on the (Unicode) input, but, alas, not to HTML escapes (which would be finde), but to characters of the said codepage that the programmer deems "close enough". The "Preview" no longer makes good for its name - what a shame! All this has been discussed endlessly, and in the end the worst possible solution was chosen. Some Western languages turn out OK, most others miss out, and many posters are deceived by the "Preview".
|