Blog

Update on the Matnsaz Beta

I would like to offer my deepest gratitude for everyone that has expressed support for the project – by signing up for the beta, getting in touch to help and collaborate, spreading the word. It has all meant a lot and indicated that this is something worth doing.

Given that the public beta has still not launched I wanted to share an update. An initial build after months of testing on one phone is now being run by a number of people on their devices. This has identified a number of issues that are now being addressed before the build is shared more widely.

Of the issues identified are some minor changes such as interface problems, missing diacritics, but also slightly more interesting changes such as disambiguating between a number of different Unicode characters and choosing the correct ones intelligently as context dictates. This includes for example:

  • و and ؤ
  • ئ and ء

These of course highlight a number of interesting cases where Unicode (the underlying text encoding standard now used by all modern software), separately encodes certain letters with a hamza below or above, instead of encoding them as a series of two characters. 

The second reason for the delay in launching this beta is the design and review of an accompanying research study. To my knowledge, few if any formal, quantified research studies exist that measures keyboard performance for Urdu (or the Arabic script in general). This is hence a great opportunity not just to design a keyboard but to contribute to literature in Human Computer Interaction, from which the hope is that this work and its learnings will provide pathways for other researchers, designers and developers. 

The design of this research study is complex, given the distributed, non-standardized nature in which the keyboard will be operated by its initial testers. Since it is also being supported by the University it is also following due process through the Institutional Review Board that ensures no harm is done to the participants of the study. Designing the study correctly to ensure valid and usable results, and to ensure its accordance with necessary review takes some time which unfortunately delays the release of a public beta.

In this time I'll continue posting updates when possible. And please don't hesitate to get in touch

Zeerak Ahmed