Scale Diagrams: Understanding Multiscale Interfaces, Proceedings of ACM SIGCHI’95, in press.

We describe Pad++, a zoomable graphical sketchpad that we are exploring as an alternative to traditional window and icon-based interfaces. We discuss the motivation for Pad++, describe the implementation, and present some of the differences between Pad++ and the standard Tk Canvas widget. INTRODUCTION Imagine that the computer screen is made of a sheet of a miraculous new material that is stretchable like rubber, but continues to display a crisp computer image, no matter what the sheet’s size. Imagine that this sheet is very elastic and can stretch orders of magnitude more than rubber. Further, imagine that vast quantities of information are represented on the sheet, organized at different places and sizes. Everything you do on the computer is on this sheet. To access a piece of information you just stretch to the right part, and there it is. Imagine further that special lenses come with this sheet that let you look onto one part of the sheet while you have stretched another part. With these lenses, you can see and interact with many different pieces of data at the same time that would ordinarily be quite far apart. In addition, these lenses can filter the data in any way you would like, showing different visual representations of the same underlying data. The lenses can even filter out some of the data so that only relevant portions of the data appear perhaps those satisfying some search criteria. Imagine also new kinds of stretching that provide alternatives to scaling objects purely geometrically. For example, instead of representing a page of text so small that it is unreadable, it might make more sense to present an abstraction of the text perhaps just a title that is readable. Similarly, when stretching out a spreadsheet, instead of showing huge numbers, it might make more sense to show the computations from which the numbers were derived. The beginnings of an interface like this sheet exists today in a program we call Pad++. We don’t really stretch a huge rubber-like sheet, but we simulate it by zooming into the data. We use what we call portals to simulate lenses, and a notion we call semantic zooming to scale data in non-geometric ways. The user controls where they look on this vast data surface by panning and zooming. Portals are objects on the Pad++ data surface that can see anywhere on the surface, as well as filter data to represent it differently than it normally appears. Panning and zooming is an approach to navigate through a large information space via direct manipulation. By tapping into people’s natural spatial abilities, we hope to increase users’s intuitive access to information. Of course, traditional computer search techniques are also available, but they produce an automatic animation to the area with the desired data bridging traditional and new interface metaphors. Figure 1These lenses shows textual data as scatter plots and bar charts. Advances in the Pad++ Zoomable Graphics Widget Benjamin B. Bederson and James D. Hollan Computer Science Department University of New Mexico Albuquerque, NM 87131 [email protected], [email protected]