This extended abstract is a contribution to the User Modeling for Accessibility Online Symposium of 15 July 2013. The contents of this paper was not developed by the W3C Web Accessibility Initiative (WAI) and does not necessarily represent the consensus view of its membership.

Developing a Semantic User Preferences and Device Modeling Framework that supports Adaptability of Web Applications for People with Special Needs

Introduction and background

The introduction of user and device models to customize applications has been the subject of research for decades. This position paper presents a modeling framework that supports dynamic adaptations of the user interface of web applications. This work builds upon previous efforts of the authors (Velasco et. al. 2004) leveraged with the use of the semantic framework Composite Capability/Preference Profiles (CC/PP, (CC/PP, 2007), which allows the matching of device capabilities and user preferences. The combination of these models with those of the corresponding web applications, enables an adaptive transformation process that facilitates access to users with special needs, derived of their functional restrictions or because of context-related handicapping situations. Our approach differs from traditional user models based upon cognitive or physiological characteristics of the users.

Approach and framework components

Our models are based upon the Composite Capability/Preference Profiles (CC/PP) 2.0 (CC/PP, 2007) RDF framework. CC/PP enables a description of device capabilities and user preferences that can be used to guide the adaptation of content. CC/PP allows to construct user and device profiles as a two level hierarchy: a profile with a number of components, which have a number of attributes. The attributes of a component may be included directly in a profile document, or may be specified by reference. This simple framework allows an efficient processing of profiles without the parsing overhead of more complex semantic models based on OWL, for instance. The framework has three major components, which are described in the following:

One of the novelties of our approach is the definition of a decoupled web technology model, which could be used by any other component of our architecture. This approach offers several advantages: (i) it allows the expression of device capabilities for either hardware or software components; (ii) it allows the expression of user preferences; and (iii) it simplifies the expression of matching rules for adaptation with different granularities, ranging from top components like, i.e., images or movies to HTML elements and attributes, or CSS properties, for instance. The model is characterized by a hierarchy of classes, in which subclasses like MarkupTechnology, XMLTechnology or StyleTechnology include containers for their constituting parts. The model is available at: http://i2web.eu/ns/2012/technology#.

Our approach for the device model is similar to that of UAProf (User Agent Profile, 2006). Due to its limitations, we decided not to extend UAProf but to design our own vocabulary, denoting a semantic equivalence or difference where appropriate (the model is available at: http://i2web.eu/ns/2012/devicemodel#). The main components are:

After evaluating different approaches, we have focused on modeling the user preferences and not her physiological characteristics. On that way, we can focus on transforming different UI components. Our work is based upon the CAMELEON Unified Reference Framework (Calvary et. al. 2002) and (Calvary et. al., 2003), which “… is a framework that serves as a reference for classifying user interfaces supporting multiple targets, or multiple contexts of use in the field of context-aware computing.” Furthermore, following the work of the SERENOA project, we have newly introduced a semantic container that holds categories of pre-defined transformations, which can be mapped to algorithms similar to those of the library of algorithms for advanced adaptation logic defined by SERENOA (SERENOA Project, 2012). Due to the abstraction level introduced by the Web Technology model, we can avoid redundancies in their classification due to different content-types. The latest version of this model can be found online at: http://i2web.eu/ns/2012/usermodel#.

The transport of the models is solved via two additional extended HTTP headers: x-i2web-uprof for user preferences profiles and x-i2web-dprof for device profiles. These headers are general header fields which MUST contain the following:

The whole modelling infrastructure is being made available through a Model Management System (MMS). The MMS is an assistive system that supports the user in configuring her preferences and in detecting the configuration of her system (operating system, browser and assistive technologies), while it enables model-aware web applications to react dynamically to these specific devices and user preferences. The MMS system completes a set of open interfaces that allow the integration of the infrastructure into third-party products and systems. With the help of the MMS, users communicate their preferences and the characteristics of their systems to applications in a transparent form to them, and enable the communication of clients with the model repositories. The system also provides a multiplatform configuration user interface to support end users to define their preferences.

User and expert evaluation

The i2web project uses a highly user-centered approach, working closely with the various user groups, and has conducted both expert and user evaluations and several field studies throughout the project. We have validated the modelling framework through these user and expert tests, which implemented our framework in three different web applications. The results of these evaluations are reported in (Power et. al. 2013). Additionally, the user interface of the MMS has been subject to 3 evaluation rounds by accessibility experts and end-users, which has led to 3 designs that improve the accessibility and usability of the system.

Conclusions and future work

The whole modeling framework has been used to adapt industrial applications in three domains (e-goverment, online banking and video on-demand services) under the scope of the Inclusive Future-Internet Web Services (i2web) project. These implementations have demonstrated the feasibility of our approach through extensive user testing. Future work is targeted to the extension of the models to include further standard preferences of users and state-of-the-art devices, assistive technologies and user agents.

Acknowledgements

Parts of this work have been undertaken under the framework of the Inclusive Future-Internet Web Services (i2web) project, co-financed by the 7th Framework Programme of the European Commission under its ICT Programme (Grant Agreement number 257623).

References