Have you left your desktop lately? Eviatar Shafrir asked attendees of the World Wide Web Workshop held at Stanford University. If you have, he said, you may have noticed that a new metaphor has been slipped into its place. Instead of the now-familiar program objects such as calculators, word processors, spreadsheets, and file cabinets, this new metaphor consists of "InfoCars" with names such as "Cello" and "Mosaic," which whisk us off into the far reaches of cyberspace. According to Shafrir, who's a member of Hewlett-Packard's user-interaction design group, "the distinction between the application and the information is blurring rapidly."
The two-day workshop, which was part of the Stanford University Computer Forum (a cooperative venture of the university's Computer Systems Laboratory and Computer Science Department) provided a mechanism for developing personal contacts between industrial researchers and their academic counterparts by promoting the exchange of technical ideas between Stanford and industry.
Among the questions workshop chair Terry Winograd challenged speakers to address were:
Among the presenters were Larry Masinter (Xerox PARC), who spoke on "Naming and Resource Location;" Brewster Kahle (WAIS Inc.), who discussed "Search and Query Protocols;" and Dan Dougherty (O'Reilly & Associates), who examined "Browsers." Other sessions included "Page Layout and Portability," by Steve Zilles (Adobe);" On-Line Communities," by Sean White (Interval Research); "Security," by Allan Schiffman (Enterprise Integration Technologies); "Digital Libraries," by Hector Garcia-Molina (Stanford); "Using the Internet for Commerce," by Michael Genesereth (Stanford); and "Web Structure and Meta-Information," by Terry Winograd (Stanford).
In many ways, I was as impressed by the structure of the forum as I was by the information presented. Professor Winograd, in making a natural progression in his career from AI, to cognitive science, to groupware, and now to WWW/Internet research, did an excellent job in tying the many and varied issues into an interesting and coherent whole.
Appropriately enough, as with his previous research, Winograd's new efforts continue to address the issues of background, context, and understanding. This time, he and his students are involved with an NFS-funded research project related to digital libraries. Part of their work will be to help define Internet-related "meta-information"--that is, information about information. Whereas the information of the pre-digital era may have filled up the card catalog in a typical library, making it a reasonable task to find the information desired, this research effort hopes to make it even easier to navigate through the digital sea of information by defining a common protocol for meta-information.
Issues that seemed to come up during conference were a constant acknowledgment of the limitation of HTML (hypertext markup language; see "World Wide Web and HTML," by Douglas McArthur, Dr. Dobb's Journal, December 1994), the propagation of general-purpose viewers, and need (or wish) for more-specific clients addressing the specific needs of user populations (doctors, lawyers, teachers, and the like). As Brewster Kahle sees it, "imagine a program like MacIntax that is network aware. You are using it to fill out your taxes and have immediate access to the most up-to-date government information (possibly through a Z3950 back end totally hidden from the user)."
Pei Wei, author of the popular Viola browser, gave some explicit examples of how to extend the current protocol by using component software.
The basic idea is that, rather than building one single monolithic application that does everything from day one, we should be building a framework or architecture that can be dynamic in its ability to have functionality added or deleted on the fly.
Those component parts can be many different things (such as special navigation control, visualization controls, self-guiding slideshow presentation tools, and so on).
Another approach is the objects with interpreted/compiled scripts model. The scripting language can have performance problems. But a lot of this is a matter of better interpreter or even compiler design. This approach can also be much safer than the executable objects approach, because the interpreter can have a chance to catch dangerous operations.
Yet another approach is to have some kind of interface protocol such that the computation can be happening remotely, and have things rendered locally. This can be quite secure, but could also come with a high bandwidth penalty.
Viola, for example, consists of a very small core engine, plus a toolkit with the primitives coded in C. Part of the toolkit is a scripting language for implementing and gluing together applications. Also, program objects can be embedded into documents, an example of which was a bookmark tool embedded into the Viola WWW's toolbar, which by being linked to the document, would come and go with the document. Pei also demonstrated a mini-chessboard application that is published on the Web server, and instantiated locally by the interpreter. What's particularly appealing about this is that you could have high interactivity without incurring a lot of bandwidth--you can move pieces around, and the board can do simple checks for illegal moves and transmit only the essential movement information instead of using the ISMAP feature and transmitting a new chessboard picture for every move.
In the the commercial-software arena, Adobe came closest to Kahle's inferences with its new suite of publishing tools. Announced first at the Seybold Conference, Adobe has created a number of strategic alliances and improvements to its Acrobat software, placing the company in a strategic position for enterprise-wide and Internet publishing. Highlights from these developments include Spyglass's commercial version of Mosaic browser, which includes links to Adobe Acrobat 2.0, cross-platform, document-exchange software, and Adobe's providing free viewers for the usual platforms (DOS, Windows, Macintosh, and UNIX).
Adobe and its partners will be trying to leverage the efficiency and flexibility of Adobe's Portable Document Format (PDF) in the lucrative Internet server and publishing market. Some of the appealing aspects of PDF for publishing on the Internet will be its ability to work with legacy documentation in existing formats, preserving original representation, indexing text in various parts of a document (such as images), and most of all, its low bandwidth requirements. For instance, a GIF file, now the most common format for images in a Web page containing 2--3 Mbytes of information, would require perhaps 800 Kbytes when rendered in PDF format.
For a complete overview of the two-day workshop, go to http://www-pcd.stanford.edu/workshop/workshop.html.