During the expert tests and the first verification phase 262 bugs were discovered, 152 of them were 'major' bugs, which may cause serious troubles using FASTY. Most bugs were caused by the user interface; some bugs address the installer and the predicted words. These bugs were to be taken care of during the redesign phase in order to provide a new, better prototype for the second verification phase. The important list of the technical problems was however the subject of a longer revision than foreseen so the redesign phase had to be prolonged until December 2003.
During the first verification phase, we also noticed that the general design was inconsistent. For the PT2, a redesign according general design guidelines was done.
In the PT2, we have planned the following: all keys are caught, passed through the FASTY system and generated newly. Some problems concerning delays which has to be ensured for a proper key generation has been noticed too. In the PT2, a new functionality for key generation avoids the delays and the maybe problematic usage of special key codes for the differentiation of own generated keystrokes and external keystrokes. Additionally the shift related problems are solved with that. Most of the display problems are caused due to the special way of creating the prediction list (using rich edit controls). This bunch of bugs is removed in the PT2 by a new way of creating the prediction list.
During the first verification, we noticed that there was an insufficient initialisation or missing drivers, which led to unpredictable behaviour. During the implementation of PT2, an error handling avoided problems with not correct working FASTY systems (e.g. not selected main drivers).
The Adjustment Tool has to be open for all kinds of drivers. Therefore it is impossible to integrate all possible settings into the Adjustment Tool. Thus, the Adjustment Tool selects the used drivers and the configuration of driver-specific options (such as: port-number, data rate, aso.) is only induced and the driver itself displays a setting dialog. The storage of the settings had to be done by the driver as well. This is not the case for PT1 but has been done in the PT2.
During the first verification phase, keys that are reserved as hot keys (NUMPAD) or double functions (F12) cannot be used for other purposes. In the PT2, a new way of catching the keys is used to avoid dead keys. Additionally the keys used for FASTY internally (e.g. prediction selection) are selectable. In this way every user can select the keys he or she can give up for normal writing. Additionally, keyboards with extra keys (e.g. Internet keyboards) can be used too (but most likely the drivers of the keyboard must not be installed). The prediction window was taking the input focus in PT1. In the PT2, the prediction window rejects the focus.
The contents of the zip archive containing the FASTY prototype 2 system is listed below.