Le lundi 12 mai 2008 à 19:07 +0300, Dotan Cohen a écrit : > 2008/5/12 Yannick Warnier <ywarnier@xxxxxxxxxxxx>: > > Hello, > > > > I've been trying to find something nice to transform an accentuated > > string into a non-accentuated string. Obviously, I'm mostly playing > > inside the European languages, but any method that could transform > > arabic or asian characters to plain non-accentuated characters would be > > perfect. > > > > I have found a number of solutions, ranging from str_replace() for every > > known accentuated character to strtr() to a preg_replace() of a > > conversion of the string to html characters then removing the "&" and > > the "alteration" string (acute, grave, circ, ...). > > > > I must say the last one seems to work better because it's less affected > > by charset changes, but it still seems awfully slow to me and I would > > like to know if there is any function that exists that could do that for > > me? > > > > Yannick > > > > Why are you removing the accents? Why not store/process the data as > UTF-8, which supports all the accents in all the languages, and even > non-latin languages. You mention Arabic, which does not use accented > latin characters (Maybe you are thinking of Turkish, Ubek or Tadjic). > UTF-8 supports Arabic, Russian, Greek, Latin including modified > accented letters, and almost everything else save CJK. > > What is your end goal? Why are you removing the accents? Hi Dotan, I'm trying to give a universally-manageable directory name to an item using a free-text title. I want to avoid every type of accentuated character and everything outside of pure ASCII to make it the most portable possible. Generating a random hash is not acceptable as we want to be the most user-friendly possible. I'm talking about Arabic not to remove accentuated characters, but in case there would be a transliteration function that allows me to turn an Arabic character into something similar in terms of pronunciation but in ASCII. So the goal is to create a directory name that is both intuitive *and* portable for the user and the admin. The title is kept for the user, but there is a generic shortened code that is generated following the given title. We used to ask for a title in a webform, but realised our users liked it much better when we give them the possibility to generate the code themselves, but generating one ourselves by default. I just realised that the developer who did it seemed to make it using html codes directly, so we end up with codes like "EACUTETEACUTE" for an item called "été", while "ETE" would be far better. Yannick -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
