Bay Area Roundtable (BART)

Scaling for the Internet series, meeting 5

Languages for the Internet

Automating the Web with WebL
Hannes Marais, Digital Systems Research Center, marais@pa.dec.com

The Hypertext Markup Language (HTML), eXtensible Markup Language (XML), and off-the-shelf tools like web servers and web browsers are enabling technologies for cheap and efficient inter-business E-commerce. A side-effect of this development is that the web is turning into a large and semi- structured database. However, in contrast to traditional relational databases, which can be accessed via SQL, the web has no query and manipulation language for performing tasks in an automated manner. We are specifically interested in automating tasks that require reliably filling in forms, retrieving and processing web pages, manipulating web pages, etc.

In this talk I will present a programming language called WebL, which was engineered for information processing on the Web. In contrast to other languages, WebL incorporates two novel features: service combinators and a markup algebra. Service combinators are language constructs that provide reliable access to web services by mimicking a web surfer's behavior when a failure occurs while retrieving a page. The markup algebra extracts structured and unstructured values from pages for computation, and is based on algebraic operations on sets of markup elements. WebL can be used to quickly build and experiment with custom web crawlers, meta-search engines, page transducers, shopping robots, supply chain integration systems, data aggregators, web mining, etc.

Biography: Hannes Marais is a member of the research staff at DIGITAL's Systems Research Center (SRC) in Palo Alto, California. He holds a Ph.D from the Swiss Federal Institute of Technology (ETH) in Zurich. His interests include user interfaces, programming languages, and information retrieval applications on the Web.

Does Java Have Alternatives?
Michael Franz, UC Irvine / IRUS, franz@ics.uci.edu

At first sight, Java's position as the defacto standard for portable software distributed across the Internet seems virtually unassailable. Interestingly enough, however, it is surprisingly simple to provide alternatives to the Java platform, using the plug-in mechanism supported by the major commercial World Wide Web browsers.We are currently developing a comprehensive infrastructure for mobile software components. This is a long-term research activity and not directly related to Java and the World Wide Web. However, purely as a technology demonstration, we have started a small spin-off project called "Juice" with the intent of extending our experimental mobile-code platform into the realm of the commercial Internet.

Juice is implemented in the form of a browser plug-in that generates native code on-the-fly. Although our software distribution format and run-time a rchitecture are fundamentally different from Java's, once the appropriate Juice plug-in has been installed on a Windows PC or a Macintosh computer, end-users can no longer distinguish between applets based on Java and those based on Juice. The two kinds of applets can even coexist on the same web-page.

This, however, means that Java can in principle be complemented by alternative technologies (or even gradually be displaced by something better) with far fewer complications than most people seem to assume. As dynamic code generation technology matures further, it will become less important which code-distribution format has the largest "market share"; many such formats can be supported concurrently. Future executable-content developers may well be able to choose from a wide range of platforms, probably including several dialects of Java itself.

Biography: Michael Franz is an Assistant Professor in the Department of Information & Computer Science at UC Irvine. He holds an Engineer's degree and a Ph.D. from the Swiss Federal Institute of Technology (ETH) in Zurich. His interests include programming languages, extensible software systems, mobile code, and run-time code generation.

Coordinator: David Redmiles, UC Irvine / IRUS, redmiles@ics.uci.edu

Directions to meeting are available

Upcoming Meeting:

Friday, July 10, 1998
Topic: Scaling for the Internet series, meeting 6:
            Component Technologies
Speakers: Rick Hayes-Roth, Chairman/CEO, Teknowledge Corp.
                  Reginal Adkins, JavaSoft/Sun Microsystems, Inc.
Coordinator: David Rosenblum, UC Irvine/IRUS, dsr@ics.uci.edu

The Irvine Research Unit in Software wishes to thank its corporate sponsors:

Sustaining:

The Boeing Company * Boeing North American, Inc. * Northrop Grumman Corporation *
Raytheon Company * Sun Microsystems Laboratories * TRW

Supporting:

FileNet Corporation * Beckman Instruments * Microsoft Corporation

Continuus Software Corporation * Hewlett Packard * Printronix, Inc.

For further information on BART or IRUS, contact
        Debra Brodbeck at (949) 824-2260; brodbeck@ics.uci.edu