Projects, Committees, Etc. |
||
Research ProjectsmSpaceA problem with accessing material from a database is that generally, one needs to know something about the data domain to form a query before it can be accessed. Often, however, we want to access data where we have no or little domain expertise. The mSpace project is about helping users access data without needing domain expertise - but where the interface also supports users learning about the domain as they explore it - if they want to learn more about it - they don't have to for the interaction to still be valuable mSpace is a project to develop both data and interaction models for this kind of "domain naive" information exploration architecture. An example of such high dimensional space is music. As our first prototype, the mSpace project is focussing on representing the classical music domain to users so that both neophytes and experts can discover tunes as well as knowledge about tunes on multiple dimensions. The project is bringing together research from front end to back end: - Interaction Design (Graduate Students: Gonzalo Ramos, PhD candidate and Paulo Pacheo, Masters Candidate) - Hypermedia Systems: in particular, link bases and agents - Databases, in particular XML structures for both multimedia metadata and connections with the linkbase (Theophanis Tsandilas, PhD student) - Related to this is work in Clustering for what may be called just-in-time user modeling to best anticipate user selections within the music space - Multimedia Metadata Standards - such as mpeg7 - and how the project may help advance this standardization effort - Networks and pre-caching to evolve making multiple streams of multimedia content available, near simultaneously (to reduce client-server delay between tune selections) (in collaboration with Michael Murphy of Ryerson University, Rogers Center for Communication) - I/O design for multiple platform access to the same mSpace interaction. (in collaboration with Steve Mann, ECE, U of Toronto) Interview for CITO Trends in Research about the mSpace music prototype Hunter GathererThe Hunter Gatherer project looks at (1) how to support the collection of within-Web-page components and (2) how making the collection of within-Web-page components affects the perception of web-based information. To address the second question, we needed to develop a tool to support this collection making process. That's Hunter Gatherer (see the Hunter Gatherer Overview if you'd like to give HG a try). In the tool design process, we worked with users through surveys, lab evaluations and field work to understand how best to develop and support transparent capture of within-web-page components. Our goal was to minimize any forced divided attention which a user experiences switching between information capture (gathering page components) and information management (openning a separate file in a separate application, pasting the information, copying the URL, moving back to the file and pasting it into the file; adding a header to the component to contextualize the material, etc.). On the architecture side, there have been certain intersting research questions related to the development of this tool with respect to the limits of an XML architecture and proxy service approach. We are preparing to present these in a paper for WWW2002. On the HCI side, we are now both refining the tool, based on user feedback, and doing a longitudinal study of tool use to address whether Hunter Gatherer use affects style of interaction with web information resources. Intensional Hypermedia SystemsR&D of demand-driven, intensional systems such as Intensional Hypertext (see Papers for further Information). Overall, Intensional Hypertext is a way to support the authoring of user-determined version control in a hypertext. That means that users can do two main things: fine-tune a digital document to reflect their needs with the information ( less of this more of that; french version of this; turkish version of that with more detail). The second thing a user can do is set which of these changes they wish to have range over an entire site, and which they wish to maintain over just a page, or just a component in a page. It's part of a philosophy of mine that says: the right data should come to the user; not the user go to it. That's the interaction side of the equation. Beneath this is the architecture to support both the authoring of such documents and the interaction structures within them for users. Please see the papers section for architecture and version refinement semantics. Collaborators include W.W. Wadge from the University of Victoria, Canada; John Plaice, University of New South Wales, Australia, and myself. We are using the Intensional approach to page and page element verioning within the IVIS Project, Biomedical Communications, Faculty of Medicine, U of T . The architecture is letting us evaluate two things in particular (1) if there are patterns between user-type or content type and kind of illustration prefered by users (photo, illustration, 3/4 profile, etc) and (2) how we can design a model for determining how to determine content which ought to be versioned. Multidimensional Video Help SystemsMultidimensional Video Help Systems is a collaborative research project among Ron Baecker, Expressto Software and myself to investigate the efficacy of versionable, web-based video help systems for self-instruction. What this means in a nut shell is that we want to see what happens if you can let users change levels (to create new versions) of a streaming media presentation on the fly. This is based on previous work in Intensional Hypermedia (described below) to let users manipulate versions of data in real time across a variety of dimensions, such as language, expertise, degree of detail and so on. The wrinkle in the video version of versioning is of course that video flow is linear, so what is the best way to support change of a dimension? For instance, when a user says i want more of "this" - what is the this? or more particularly when does the this refer to? should the version shift move back to the start of a particular clip, or just from that selected moment? SupervisionCurrent:
Research Assistants
Recent Graduates:
CommitteesLocalGraduate Committee, Computer Science Department, U of Toronto ConferencesCo-chair, Tutorials, Hypertext 2001, Aarhus, Denmark. Program Committee, Third International Workshop on Structural Computing, in conjunction with Hypertext 2001 Program Committee, Seventh International Workshop on Open Hypermedia, in conjunction with Hypertext 2001 Steering Committee, The Knowledge Media Design Institute (KMDI). Web Media Chair, Grace Hopper Celebration of Women in Computing Conference, 2000, Cape Cod, Sept. 2000. Research Groups / LabsBell University Labs Collaborative Effectiveness Lab University of TorontoThe Collaborative Effectiveness Lab (previously the Personalization Technologies Lab) considers a variety of aspects, from the social to the technological, of the development and implementation of customizable virtual environments and data delivery. I've been involved with several projects in the Lab: Customized Text Generation, with Graeme Hirst and Virtual Organizations/Communities Infrastructure Implementation with Mark Chignell, Gale Moore, Barry Wellman and Ian Spence, Dept of Psychology, U of T. Other more recent projects involve developing methodologies to help support tool development for collaborative work which leverages what tools users already know how to use. The Dynamic Graphics ProjectThe Knowledge Media Design Institute |
||