In mathematics, one exciting application of web technology is CECM's Organic Math project, mentioned above. The project seeks to provide a system for true collaborative mathematics and web-based dynamic publication. In particular, it seeks to provide for the browsing of mathematical documents, the verification of those documents by programs like Maple, and the annotation of those documents with comments by researchers. These documents can contain information stored in all the various forms the Web allows, thus producing dynamic, graphical, documents that contain far more information than can be represented on a simple black and white piece of printed paper. It is this type of technology that will allow scientists to communicate accurately.
A system such as that proposed in the Organic Math project, however, must provide an interface to the user that
The VIN system contains an interesting intelligent agent that performs a very important, and basic, service. That service is storage and retrieval of annotated images. VIN itself handles the information displayed on the electronic blackboard as an image, simply described by its network URL, with chalk marks and typing written on top of it. This information can be transferred to an agent, and stored at a central site. The amount of information is tiny, because the original document, the image, is not stored, but only the annotations to the document. The original author of the document retains possession of it. This is realized by the Vault agent. The Vault agent provides easy access and storage of what has been written on a blackboard. Access to the service is controlled through a password mechanism (these things are easy to program into a client using Java), and can be allocated to a user, or to the channel on which a discussion was taking place. The service uses the power of Java to make its use as easy as possible: the user is hardly even aware that the information is being transferred across the network, and even less aware that, automatically, other agents are called in to help translate images into the formats needed by the web browser. The user simply clicks on a list of available files, and the intelligent distributed network that VIN represents takes over: the stored blackboard simply pops up on the screen. Vault, again, represents complete ease-of-use. One only needs a modern web browser, like the new NetScape, and then one can use all the services that it offers. Obviously the extended document vault that will be needed by the Organic Math project, and any large distributed, networked, institute, can gain a lot by such functionality. Of course, we must not rush to do everything using one concept like executable content. We want to provide plain Web services for users of older browsers. We must, however, try to move with the technological capabilities that are now emerging.
The second, rather powerful, way in which Java, and executable content can be used, is by its direct inclusion in the actual publications we are interested in. A demonstration of this is included in an online overview of my experimental differential geometry research, which contains executable content that provides animations from images generated by me at CECM. This is a very simple application that makes a big difference. One can now see the results of the research. Animations, however, are just a start: one can easily set up Java programs that allow the online Reader to look directly at the Author's data, and to manipulate that data in a fully interactive way. This moves us significantly beyond the simple inclusion of images, and so on, in web documents, which are often no more than colour representations of what we can find in the standard paper journals. With executable content, we can suddenly immerse our readers into our research itself. A simple demonstration is a short paper (available online) that I wrote that describes the Mandelbrot set, and then provides a window that allows the Reader to actually wander through the set, zooming in on any region they like:
The interaction is as simple as moving the mouse over the region desired. The browser itself then calculates the next image. The applications of this kind of interactive networking to real-world mathematical science are obvious. We can expect to see intelligent documents which provide many tools that the reader can use to gain the most from them. These tools could be anything from indexing to a mechanism for commenting on the paper.