The Graphics Production Class Library stems from a need for convenient tools for producing quality statistical graphics. Feedback from seminars shows that there is interest in applying new ideas for statistical graphics, but the reality is that people will use what is most convenient to create graphs. The problem worsens when the graphics are included in World Wide Web pages. These rendered bitmaps are often barely legible on low resolution computer screens. To solve both of these problems at once we are developing platform-independent, extensible Java^TM class libraries for producing graphics.

The library itself is composed of a set of reusable components (objects) that encapsulate drawing graphics. These objects can be "glued" together to create applets or applications, and the objects' underlying methods are used to change high level behavior and appearance. Applets are miniature software applications that run within Web browsers, and applications are standalone software products.

Beyond the ability to draw directly on the user's computer screen, one of the most significant advantages of this approach is that we can introduce a whole new level of user interaction with statistical graphics. The user will be able to directly manipulate the graphic to more quickly answer his or her questions about the data. In a sense, the graphic will become the interface, since the interface is the graphic. For instance, moving the mouse over a data point can yield its value in the lower left corner of the browser's display. Other ideas include:

• The ability to drag & drop actual time series from different panels onto one panel to facilitate comparison.
• The ability to drag panels around with the mouse to reorder them.
• Providing windows that pop-up with more information about a data value (such as metadata or hyperlinks to further information) when the user clicks on a particular data point.
• The ability to navigate through the data by dragging or stretching a novel "Pan & Zoom" control.
• The ability to highlight points of interest. For instance, being able to highlight every April in a time series.