| What about spell checkers for the African languages? |
|
Translate.org.za has spell checkers for Afrikaans and South African English. We also have ones in all of the other official languages, but we do not actively promote them as they are not yet at a level where we feel happy for them to be widely used.
Why, you might ask, are the other African languages not as well developed as Afrikaans, for example? Is it some plot to surpress the other languages? No, definitely not! The reason why we don't have very good checkers in all the languages is simply that making a spell checker, with current technology, is difficult for a language in, for example, the Nguni group (Ndebele, Zulu, Xhosa and Swati). The reason is that the language, as written, is highly conjunctive. That is, they join parts of speech together to create 'words'. Current spell checkers are mostly based around word list lookups - if the word is in the list, it is valid; if it isn't, then it's a spelling mistake. But in the Nguni languages, since there are so many permutations, it becomes impossible to create a complete wordlist. To create a good Nguni checker you need to create a basic morphological analyser, something that understands the way the language builds 'words'. Once that is in place a spell checker would have the intelligence to decide whether a 'word' was correcly constructed from the various word stems. There is no free implementation of an Nguni morphological engine. Current work in this area looks as if it will remain proprietary, which unfortunately will hold our languages back. We hope in the future to work with other Africans to create an open morphological engine that could help all Bantu group languages. For the other South African languages, the Sotho group, Venda, Tsonga and also Afrikaans there is less of an issue around morpohology. Afrikaans theoretically should have problems as words are joined together to form new words, yet in practice a word list-based spell checker is quite satisfactory, morphology would just add a little more value, especially in correctly checking aglutenated words and in suggestions. The problem with all languages is the availability and development of a list of valid words. These take time to collect - the Afrikaans lists is based on work conducted in the 80's and 90's. We have only just started collecting for the others and there is a long way to go. You can help by contributing your words and corrections for inclusion in these lists. Although we publish our spell checkers on our site we have begun to rate them as we do not want people to have a false impression of their usefulness. But we do want people to make use of them, so we won't be hiding them. |
Download spell checkers| < Prev | Next > |
|---|
