A Single-Authoring Technique for Building Device-Independent Presentations

One of the major challenges in the prevailing web page development is device heterogeneity. Currently, devices such as cell phones, PDAs, and PCs are already capable of providing access to the Internet. In the near future, we are expecting other devices such as car navigation systems to have the same capabilities. As the number of these devices increases, the traditional approach of developing a separate web page for each device results in too many redundant efforts. It would take too much effort for authors to learn different device-specific markup languages and tools, and then to implement and maintain a large number of device-specific web pages of the same web application. To meet this challenge, we propose a single-authoring technique called ScalableWeb. It allows authors to build a device-independent presentation model at design time. The device-independent presentation model can be used to generate device-specific presentation at runtime, with the possibilities of customization from authors. Introduction Nowadays, we see numerous devices having sufficient computing, networking and display capabilities to run a web browser. However, due to device heterogeneity such as screen sizes and input methods, not all devices can access to the same web page and not all accessible web pages are fully functional on all devices. One of the feasible solutions is to provide a single-authoring technique, with which authors can build a device-independent presentation model that can be used to generate device-specific presentation. Since building a device-independent web page that can be accessible from any kinds of devices is an enormous problem, we would like to approach this problem by taking one step at a time and first considering the most common web-supported devices. They include PCs and Pocket PCs that accept HTML [1], and cell phones that accept cHTML or WML [2]. We also start by initially focus on web pages and web applications which presentations are consisted of purely graphical widgets, and exploiting high-level form controls such as buttons, forms, and text fields. We select graphical modal as it is the most frequently used modality; and we concentrate on the high-level controls as they are commonly exploited in e-commerce web applications. Challenges Within our scope of interest, there are three major set of challenges: • Presentation Accessibility – the ability for a system to produce a presentation data that can be rendered on a target device • Presentation Quality – the ability for a system to produce a highly usable presentation • Function Customization – the ability for an author or a user to customize functions of a web application For presentation accessibility, it is caused by device-specific markup languages (e.g., PCs, and Pocket PCs support HTML, but DoCoMo cell phone only supports cHTML) and input methods (e.g., PCs support keyboards and mouse, Pocket PCs support stylus and virtual keyboards, and DoCoMo cell phones support keypads). For presentation quality, it is affected by different device screen sizes (e.g., PCs’ screen size is above 800 x 600 pixels, Pocket PCs’ screen size is around 320 x 240 pixels, and DoCoMo cell phones’ screen size is around 130 x 120 pixels). For function customization, it is about task preference. Task preference means a task may be suitable for a few devices but not for others. For example, a text inputting task may be suitable for PCs but not for cell phones; as cell phones do not provide a convenient input mechanism. Task preference can also be based on delivery contexts [3] other than device types. Using the example in the W3C Authoring Scenario informal draft [4], if a user is in a hurry, he/she may choose to disable all images presentation and just download the text. With the additional task preference option, the user can further minimize the download time by just selecting the task that he/she is interested.