Language codes (e.g. en for English or cy for Welsh). Using these may affect whether search engines, browsers and other machines are able to determine the language of your pages and offer responses based on this information. For example, pages like http://www.swansea.ac.uk/cy/ will bring up a Google translate option in many browsers.
Western European languages may appear to use characters from a standard alphabet but this may not be the case. For example, the Welsh alphabet has 28 characters, does not include all the characters in the English alphabet and includes eight characters (ch, dd, ff, ng, ll, ph, rh, th) that are recognised as single characters (graphemes) in Welsh but as two characters in English. This means that someone called Dafydd or Llinos may define their name as being five characters long or six characters long, depending on which language they are using. This also means that it's not possible to have an "a-z" listing in Welsh: there's no "z" in Welsh, so it would be "a-y".
Equally, some languages will have accents, umlauts and other special characters. These often get affected by the environment in which you're running your code, so be particularly careful to test in your live environment. You should also make sure that these render on the devices (whether smartphones in country x, PCs or tablets) that will be used to access the service.
Welsh has additional accents not available in standard Western language Unicode support. Particular care needs to be taken with font selection to ensure full Latin Extended-A support as ŵ and ŷ may not be included in the font set.