Currently, multimedia documents can be accessed at anytime and anywhere with a wide variety of mobile devices, such as laptops, smartphones, and tablets. Obviously, platform heterogeneity, users’ preferences, and context variations require document adaptation according to execution constraints. For example, audio contents may not be played while a user is participating in a meeting. Current context modeling languages do not handle such real life user constraints. These languages generally list multiple information values that are interpreted by adaptation processes in order to deduce implicitly such high-level constraints. In this chapter, the authors overcome this limitation by proposing a novel context modeling approach based on services, where context information is linked according to explicit high-level constraints. In order to validate the proposal, the authors have used semantic Web technologies by specifying RDF profiles and experimenting on their usage on several platforms.
TopIntroduction
A huge amount of multimedia documents can be created and accessed by users. These documents may be composed of different types of contents, such as videos, audios, texts, and images (Jedidi, et al., 2005). For instance, good examples of multimedia documents are Web pages or SMIL presentations (Bulterman, et al., 2008). In those documents, multimedia contents are synchronized and organized according to the graphical layout of the presentations (Roisin, 1998). Moreover, users may be able to interact with presentations by selecting particular elements (e.g., a click on a picture plays a video). Besides, many mobile devices (e.g., laptops, smartphones, and tablets) are able to display multimedia documents. This universal access allows users to consult documents anytime and anywhere (W3C Ubiquitous Web Domain). However, such devices have heterogeneous capabilities and characteristics in terms of hardware (e.g., screen size, battery) and software (e.g., players, codecs) (W3C Device Independence Working Group). Moreover, user’s preferences or handicaps may prevent from playing specific multimedia contents (W3C Web Accessibility Initiative). For instance, a user may avoid reading texts written in French and/or avoid playing audio contents, while he is participating at a meeting. All these restrictions introduce constraints that have to be specified with a profile. In a profile, various categories of information have to be managed:
- •
Device characteristics (hardware and software).
- •
Context information related to interactions between the user and the device, such as the preferred languages, the bandwidth, and the surrounding devices.
- •
Document structure like the types of contents that should be played or the preferred presentation organization (e.g., layout, multimedia contents synchronization, hypermedia links).
Consequently, if a multimedia document does not comply with some constraints that are specified inside a target profile, the document may not be correctly executed on the target device. Therefore, in order to display documents on any devices, these ones have to be adapted, i.e., transformed in order to comply with the target profiles.
Since the last decade, a fair amount of research has been conducted on multimedia document adaptation (Adzic, et al., 2011; Laborie, et al., 2011; Lemlouma & Layaďda, 2002; Asadi & Dufourd, 2005; Jannach & Leopold, 2007; Ahmadi & Kong, 2008). Considering some target profiles, these approaches combined multiple operators: transcoding (e.g., AVI to MPEG), transmoding (e.g., text to speech), and transformation (e.g., text summarization). Certainly, each profile expressiveness is exploited by these approaches in order to determine a combination of these operators, such as in Lemlouma and Layaïda (2001), or to optimize their deployments (e.g., for saving battery energy), such as in Laplace et al. (2008) and Girma et al. (2006). However, each proposal exploits specific profile format, which usually contains a list of multiple descriptive information values, such as the screen size, the user languages, and the battery power. Consequently, an adaptation process has to interpret such profile values and to deduce implicitly some constraints. For instance, if a battery power is lower than 10%, an adaptation process may avoid playing hi-quality videos, while another one may provide low-quality videos. Obviously, each adaptation mechanism may deduce different constraints that in many situations might be wrong, thus providing to users incorrect adapted documents. Furthermore, current context modeling languages, e.g., Bolchini et al. (2007), Forough and Reza (2012), do not consider expressing such high-level constraints, while they might be very useful to guide the adaptation process. In this chapter, we overcome this limitation by defining a new profile description model where:
- 1.
Profile information are organized into facets (e.g., device characteristics, context information, and document structure) and composed of services that either provide data or require modifications, and
- 2.
Some profile information are linked by explicit high-level constraints.