Monday, February 17, 2014

The place of #language support in #MediaWiki support

At this moment #Wikipedia has 280+ versions. They represent 280+ different languages. Every language is written in a script. This script may be from right to left (eg Latin), from left to right (eg Arabic). Top down is not yet supported.

All these languages use Unicode to represent the characters in their script. As a consequence it is "just" a matter of configuring a language correctly and it will work in a wiki. The characters used in a script have to be defined in Unicode. Without this supporting a language is problematic. An example of this is the SignWriting script and all the sign languages. They are not supported by the Wikimedia Foundation and they are developing their software themselves.

When people want to have their own Wikipedia, they have to localise the MediaWiki software at This is a requirement for the people that request a new project to fulfil. All that is required to support another language is some configuration.  To support all languages other than English, MediaWiki developers have to internationalise their software. This is a well established process, there is nothing special in there for "small languages", it is the same process for any language.

When people use MediaWiki it can be assumed that they use any of these 280+ languages. It cannot be assumed that they know any other language than their own. The Universal Language Selector allows people to select the language they need for their user interface. They can choose a language by typing in the name of their language in their own language.

Once a languages is selected, MediaWiki supports both input methods and web fonts. This is necessary because most devices do not support all the languages MediaWiki supports. At this point a user can read and write in MediaWiki. A special case is the OpenDyslexic font; is enables 7% of a population to read their language with less effort. This font works only for many languages that are written in the Latin script like English.

One other aspect of language support is in enabling search in all the 280+ languages. This is where Wikimedia engineers may have to do special things for a single language. When they do this well, they only have to do this once. Typically however, setting up search for a language is well documented, it is part of international standards.

Language support for MediaWiki is nothing but internationalising the software and for being mindful that MediaWiki is used by people representing over 300 cultures and languages. There is no special consideration for small or big languages.

No comments: