Open issues in model transformations for multimodal applications

Lengyel, László and Charaf, Hassan (2015) Open issues in model transformations for multimodal applications. JOURNAL ON MULTIMODAL USER INTERFACES, 9 (4). pp. 1-9. ISSN 1783-7677

Text 2015 JMUI - Open issues in model transformations for multimodal applications.pdf Restricted to Registered users only Download (1MB)

Abstract

Multimodal human–computer interaction refers to the interaction with the virtual and physical environment through natural modes of communication. Multimodal input user interfaces have significant role in different domains, for example industrial plants or hospitals, also they have impli- cations for accessibility. To develop multimodal applications in formally rigorous settings, software developer teams may use tools or a software development kit to increase the effi- ciency and the quality of the resulted software artifacts. Such a technique is performing the design with software modeling and applying model transformations to generate well-defined components of the software. Furthermore, representation- bridging communication is a discipline of cognitive info- communications, where the sensory information transferred to the receiver entity is filtered and/or converted. When- ever such approaches are used, the challenges associated with the modeling of information requirements, user capa- bilities and cross-model interactions are compounded and further increase the need for formal design and verification tools. Applying model transformations is a way to sup- port this activity. Communication-intensive solutions often require complex methods, i.e. significant model transforma- tion efforts between the different representations. Important semantic information should be preserved and not misin- terpreted in a complex model transformations. Therefore, methods are required to verify that the semantics used during the application generation and analysis are indeed preserved across the transformation. As a case in point, such a model transformation could yield embedded code for a given type of electronic driver assistant system based on a high-level characterizations of the information to be transferred and the driver’s cognitive capabilities. Later, a multimodal inter- actions expert could easily modify those characterizations on demand, and regenerate a modified version of the soft- ware without having to know about the low-level details of the embedded platform. This paper provides a strong motivation regarding the necessity of methods to support ver- ification and validation of model transformations supporting multimodal application development and cognitive infocom- munications. As the main result of the paper, we compile a list of open issues in the field of verification/validation of model transformations, and link those issues to the devel- opment of multimodal interfaces. Through its discussions, the paper makes the point that the design practices behind multimodal interfaces could strongly benefit from the use of formal modeling techniques in general, and model transfor- mation approaches in particular.

Actions (login required)

Edit Item