The Pan African L10N had a conference in Morocco in February. I learned about it through a blog I read today. I am really happy with the progress that is reported on of more languages being supported. I am however coming more and more to the conclusion that there is a need for a stage before actual localisation that will provide a service to the bilingual people of a language.
At this stage, the support of a language is very much an all or nothing affair. There is a localisation or there is nothing. This is not how it needs to be. When a language is known to exist, the lowest level of support for that language is the acknowledgement that this language exists. This is currently not done, and I think it is a missed opportunity.
The first thing to consider is, what languages and linguistic entities exist and, how do you support this. This is a surprisingly complex question. Languages are recognised in the ISO 639 standard. There are several versions of the standard and not all languages have a script that is supported in Unicode. Even when a script is supported in Unicode, it does not mean the an associated font is available for a language. The consequence of these two points is that a subset is needed on computer. On the other hand the currently recognised versions of the ISO 639 do not recognise orthographies or dialects or other entities that make a difference to how documents are to be supported.
This is not an issue the organisations that develop and localise software want to tackle. For them this a distraction. Deciding what linguistic entities can be supported is something that is best addressed by one organisation that exists to deal with issues like these. The World Language Documentation Centre (WLDC) is that organisation. Through its association with Geolang and because its board of experts in many of the relevant fields, it is already in a prime position to the research that goes into the development of the ISO 639-6.
With the WLDC and Geolang able to provide researched and verified information about linguistic entities that can be safely supported, it is then up to the applications to at least acknowledge the existence and allow a user to create content in that language. As more information becomes available, spell checkers can be added specific to that linguistic entity. In this way slowly but surely the functionality grows without the need to first localise the application.
In a way this is a solution for a "chicken and egg" problem. This problem is solved when you think of it in an evolutionary way. First there was the egg, the support of the language, and then the chicken evolved, the localisation of the application.