Lloyd Rutledge - Research Plan

Main Theme

I am currently researching the extention of existing techiques for the automatic creation and adaptation of hypermedia presentations. I have been involved in developing constructs for adaptivity in SMIL, such as test attributes, the switch element, and the excl element. I plan to continue contributing to making SMIL more adaptive. I have published papers on extending existing multimedia presentation constraint techniques to be more powerful and expressive. I plan to continue exploring how these constraint extensions facilitate writing documents in terms of more abstract concepts such as rhetorics. Finally, I have been, and plan to continue, investigating how presentations can be generated from archived media and meta-data. These three areas have been my foci for exploring the "synthesis" of hypermedia presentation -- that is, the creation of increasingly varied hypermedia presentations either fully automatically of through more extensive computer involvement in the work of the human author.

This research area is part of the large goal of making hypermedia information on the Web more adaptive and widely applicable. Web authors should be able to publish the information they create once: at one time, in one place and in one form. Web users should be able to perceive this information in a wide variety of presentation circumstances. Idally, this means that users can ask for exactly what they want and how they want it, and get exactly that, from whatever sources exist. It also means that authors organize their information as they see fit, independent of its presentation, but still be assured that it will always be presented in a way that respects their intentions for it. This frees hypermedia information posted on the Web for much more varied use.

My approach is to take the Web as it currently exists and augment it bit by bit towards this desired goal. The Web, as it currently stands, is authored in a presentation-oriented manner. Documents are typically describes a specific presentation, not the information the presentation conveys. With this as one end of an extreme, the other is whe documents are encoded as the information itself, independent of any presentation of it. This information can then be shown in any presentation to which it applies, and shown in the manner that is most appropriate for the circumstances. This presentation-independent other extreme is a guiding goal of my research. I'm exploring how the hypermedia information we put on the Web, and later perceive on the Wev, can be made increasingly presentation-independent.

Some Lists

There are three perspectives from which to view such a widely-adaptive hypermedia Web. These are:

Some specific topics within this reasearch goal that I have explored and will explore this research are:

Progression of Publications

The formal basis for the most conceptual consideration of hypermedia presentation generation process has been, and will remain, the Standard Reference Model for Intelligent Multimedia Presentations Systems (SRM-IMMPSs). The SRM-IMMPSs is a theoretical model structuring the presentation generation process. For conceptual-level research, we can fill in and refine the components of this model and the processing and communication that occurs between them. For more pragmatic research, an application of existing standard and tools can be designed in terms of this model to test both the model and the effectiveness of the standards and tools for collective meeting the goals of the model. All of the SRM-IMMPSs-based research approach describe above has been performed already, and can continue to be performed. What can also be done is the application of the results of exploring the use of standards and industrial tools to improving the further development of these standards and tools.

A key aspect of this research approach is exploring different types of meta-data from which presentations can be generated. These different types will exist at different levels of presentation-independence. Since the generation of presentations from meta-data that is more presentation-oriented is better understood than that from more presentation-independent meta-data, this provides a ladder of increasing presentation-independence that can be climbed. A current research topic of mine is to relate traditional multimedia constraints with the introduction of cross-dimensional constraints, which are more presentation-independent. Another step on top of this in my current research is to relate rhetorical structure to these cross-dimensional constraints. Rhetorical structure and its processing into presentation is one step further from cross-dimensional constraints to presentation-independence. My research, with Brian Bailey, introducing cross-dimensional constraints and the relationship with rhetoric structure, will be published in Hypertext 2000.

Continuing this direction of research will involve further development of these meta-data types and the introduction of more types into the generation process. The topic of cross-dimensional constraints will be further developed into a submission to Multimedia Modeling 2000. I will work with Jim on exploring the role of rhetorical structure in generation hypermedia presentations and submitting this work to ACM Multimedia 2000. A potential new type of meta-data to use in presentation generation is narrative structure. I'd like to incorporate the narration-related research of Craig and Frank N into this semi-automated generation framework. Narrative structure could be shown to be more presentation-independent than rhetorical structure. The MPEG-7 research of Frank N could provide more types of meta-data to incorporate into the framework.

Involvement with Groups

My involvement with several outside groups will help guide, support and apply this research. Participation in conferences, of course, will be a primary community involvement for exercising my research and relating it to that of other groups. My participation in the W3C will be the primary venue for applying my research to standards. This participation includes involvement in several W3C working groups, primarily those of which I am an active member: SYMM, which is developing SMIL, and XML Linking, which is developing XLink and XPointer. Involvement with the company Oratrix and their development of the GRiNS editor for SMIL multimedia documents will be one resource of industrial application of research.

Primary among my external group involvements are the two current sources of my external funding: the MIA consortium, and TALMUD project. Both groups involve research on multimedia database storage and retrieval across the Web. The TALMUD project is a consortium of several European telecommunications companies whose purpose is to make a prototype for multimedia organization and retrieval on the Web. KPN is one member, and they fund my involvement in TALMUD.

The MIA consortium is a collection of groups located in and around Amsterdam that research multimedia database storage and retrieval across the Web. The purpose of the consortium is to encourage collaboration in developing multimedia storage and retrieval systems and subsystems among its members. MIA members include several industrial partners, providing another venue for industrial application of research.

Involvement in these groups will help to more quickly test these research concepts. The results of this testing will enable significant improvements to the research models. Furthermore, involvement with these groups will quicken the development of these research concepts into industrial applications and general use.