FAQ
Up
The Video
FAQ
The People
Motivation
Overall Design
Navigation
The Technology
Windows
Visual/Audio Cues
User Studies
Press Clippings
References
Links

Current Status

The Task Gallery project was a research test bed from 1999. From it we learned many valuable lessons about metaphors, interaction techniques, and underlying technological infrastructure for application redirection. Development on the Task Gallery has not continued directly, but many of its lessons and techniques are being used in ongoing projects. Information about these projects can be found at the web page for the Microsoft Research VIBE group page. In answer to the most popular question, we are sorry to say that there is no download beta version of the software available.

Thanks for your interest in the TaskGallery! The amount of email related to this work has been a little overwhelming and thus replies to individual queries may be quite delayed! Also, because of the number of requests, we are not able to send out VHS copies of the TaskGallery video. For those of you who would rather get the mpg video from an FTP site, I have placed it at:

There have been many questions about this project – I hope this email answers most of them:

We started this project back in January of 1999 as a way of learning about the role of spatial memory in finding large numbers of documents. We finished a first prototype in time to be shown at WinHEC ’99 in April of 1999. Subsequently we ran several user studies that focused on how easily people could find tasks that they themselves had created. The results were encouraging and are informing our current work. The prototype has been shown at several conferences including WinHEC 99, CHI 2000, and ACM 2001.

A copy of the paper, as presented at CHI 2000 is available at::

There has also been some press coverage. You can see a sample of that at:

 

The TaskGallery is a research prototype meant to demonstrate the feasibility of a 3D user interface for managing large numbers of documents and applications. The project had a limited scope and was never meant to produce production code. As implemented, the TaskGallery was a stand-alone application built on top of a modified version of the Windows 2000 operating system. It took advantage of “Redirection” technology that allowed any existing Windows application to “live” inside of a 3D virtual environment. We cannot release the patches to the operating system that are required to run the TaskGallery.  Because of this we are also not able to release working versions of the TaskGallery for public-consumption. The TaskGallery is not a future version of the Windows operating system or user-experience. Instead, it is a research infrastructure that we are using to try out a myriad of different designs.

3D will be a component of future user interfaces. The question is for what tasks does it actually make sense. Our observation is that most of the time users just want to get their work done. They want to write a letter to someone, balance a budget, or order something off the web. For most of these common tasks, the UI will probably remain full-screen single-application and 2D. The magic starts to happen when users want to bring in information from multiple places, see relationships between multiple items and quickly get overviews of changing data-streams like stock quotes to weather to traffic to email). 3D will provide a natural way to do this.

3D allows users to switch between multiple tasks in a fluid, discoverable, and repeatable way.  We can show multi-dimensional data in easy comprehensible ways (channel guides for future televisions) and let users manage large collections of information in a personal and comfortable manner.

A great deal of our research takes advantage of natural human abilities. Our intent is to shift much of the user’s efforts from the cognitive (“thinking”) to the perceptual (“sensing”). By doing this, we hope to address many of the problems with current interfaces.

The particular metaphor used in the TaskGallery was one where the user is essentially “inside the data.” This can also be referred to as a first-person-perspective as is seen in many video games. The main difference is that we focused on providing higher-level navigation controls. Rather than a user driving down a hallway in a free-form fashion, we gave the user several controls including “next task,” “previous task,” “show me more,” “show me all,” and “show just this task.” This high-level interface does not require careful hand-eye coordination and is well-suited to users who are not well-versed in the realm of first-person video games. Additionally, since transitions are accomplished with smooth-animations between understandable vantage-points, disorientation on the part of the user is avoided.

We worked on several iterations of the TaskGallery, and with each, incorporated feedback from user studies. Some of the areas that we refined were:

  • The items held in the user’s virtual right and left hands were changed so that the user does not have to look away from the center-of-interest. The right and left hand slide up into view during interaction and quickly slide back afterwards.
  • Icon designs became simpler.
  • We added an “X-ray” specs button to each document toolbar. When toggled on, the front-most document becomes semi-transparent, thus letting the user select windows that underneath the front-most document.

We have also identified several areas that are ripe for future research:

  • Currently when a user demotes a document (places it in the background) the entire representation of the document shrinks. It would probably be more appropriate to use some kind of semantic-zooming, whereby different aspects of the document differentially scale. A web-page title and identifying logo might stay screen-size invariant while body text shrinks linearly.
  • We still need better texture filtering mechanisms. As text-heavy documents scale down, their contents become quickly unreadable. We have identified several trade-offs between “smoothness” and “detail” but are still not satisfied with current hardware based filtering techniques.
  • In general, we would like to avoid situations where users have to individually inspect large numbers of documents in order to find one particular document. More work needs to be done on generating auto-summarizations for documents and the creation of effective visual landmarks for way-finding among large numbers of documents.

We are now exploring new metaphors in addition to the first-person-perspective of the TaskGallery such as dynamic virtual objects that can co-exist with traditional desktop-applications and operating systems. We are also looking into the use of different types of display and input device configurations including very wide-aspect-ratio displays, touch-sensitive mice, and peripheral notification schemes. More information on this work is available at:

I have put together a presentation on some of the more fundamental aspects of 3D User Interface Design and how they relate to the TaskGallery. You can access a web-ready version of this presentation at:

http://research.microsoft.com/~dcr/talks/3dui_00_02_holder.htm (note: there are several videos referenced in this presentation which do not appear in the web-version). Under special circumstances, I am willing to give out the source PowerPoint version of this talk. Just let me know what your intended use is.

There are also some other excellent resources out there for learning about metaphors for representing large information spaces. One of the best is:

For people interested in getting into this field, I recommend a program of study in Human-Computer-Interaction, Interaction Design, and/or Industrial Design. One of the best general resources about the HCI field is the SIGGHI web page:

Thanks, and I apologize for any delayed responses,

- Dan

 

 
Back
Next

©2005 Microsoft Corporation. All rights reserved. Terms of Use |Trademarks |Privacy Statement

For problems or questions regarding this web contact dcr@microsoft.com.
Last updated: July 28, 2005.