One of the biggest challenges working with contextualized apps in Ethiopia is the lack of existing supportive language tools. For instance, creating a Tigrigna or Amharic based system usually lacks seamless Geez (ግዕዝ) writing tool that contains integrated grammar and spelling checker. The same is true with the other latin based languages including Oromiffa and Qafar.

Layered language architecture

Layered Language APIs Architecture


In order to create robust Ethiopic apps, the languages to be integrated need to have basic support for seamless writing and reading. Otherwise, it is difficult for users to adapt to such systems that have limited functionality in  their local languages. Even if app developers are aware of such drawbacks, it is always difficult for them to implement it without the involvement of other stakeholders in academy and linguistic institutes.

Working with languages requires that all the characters, words, and grammatical structures should be first stored in a suitable computer system. This work is directly related with linguistic professionals or people working on related fields. Collecting every word of a language is not also a simple task especially for those languages which have varied accents (ላዛ). In Tigrigna for instance, a single word can be pronounced in various tones and accents: ኸልቢ፣ ከልቢ፣ ኻልቢ refers to the same object. Such challenges are actually addresses by the app framework developers. The availability of such archived language data gives an underlying framework that will be used by any researcher or app developer.

Aelaf, in collaboration with MIT, is starting an open source project for the development of a framework or API so that other app developers in the language will directly use it for their systems. This project will start from registering Geez and Tigrigna characters, words and grammatical structures. Based on this data, an associated API and libraries will be prepared so that external systems will make use of them.