Smalltalk and the Web

A Panel Presentation from
Smalltalk Solutions '96 
New York, NY

Everyone is talking about the World Wide Web and the technology behind it. Some seem to think it is the Silver Bullet for which everyone has been longing--and the latest buzz is that Java is going to be the programming language that takes over the world. This panel consists of Smalltalk experts who also have been doing some real work withthe Web and Web-related technologies. We'll discuss whether the Web is a friend or foe of Smalltalk and what the Smalltalk industry needs to do to find a firm place in The New WebWorld Order amidst Java, Blackbird, and the rest of the hype.

Moderated by:

Ken Auer, Knowledge Systems Corporation


Ward Cunningham, Founder Portland Pattern Repository web site, Cunningham & Cunningham, Inc.
Ralph Johnson, Book Author and Member of Faculty at University of Illinois
Patrick Mueller, Technical lead for the IBM VisualAge Web Connection
Michael Robicheaux, VisualWave expert, ParcPlace-Digitalk

Position Statements:

The panelists were asked to submit strong, even outragous positions.

Smalltalk (and everything else) has lost the applet battle to Java. This is not fatal, because there is more to life than applets, but Smalltalk would have made a good applet language, and it would have turned Smalltalk from a modest success with a hopeful future into a blazing success with a fabulous future. Smalltalk's failing was not technical, but marketing. There will be other opportunities, but will we lose them, too? -- Ralph Johnson
(Ralph elaborates in his web pages)

Everything put into Smalltalk in the last ten years to make development go FASTER has actually made development go SLOWER. If Smalltalk is to survive the development pace of the web, where one human year is said to equal five web years, then we are going to have to rip some things out. -- Ward Cunningham

Web Browsers are dead! At least in terms of using them as the client interface for Web based applications. And Smalltalk is well-positioned to become the client engine for such applications. I'll talk about why web browsers won't work for many classes of applications that people want to build, what a possible solution is, and why Smalltalk fits in nicely. -- Patrick Mueller
(Pat's slides appear here)

Starting the new millennium we will look back and see Java as the pentultimate marketing example of the century. Smalltalk will be a pervasive technology in use by mere mortals as well as the gods. Some people will not even realize they are using it. Smalltalk is the only technology that has proven robust enough to flex with the changing times; mainly because it is the only truely real implementation of objects. Java will be around in its niche. However, it will be remebered as yet another language created by a hardware vendor. The language wars will change from C++ vs. Smalltalk to Java vs. Smalltalk. When will people realize its more than a language? -- Michael Robicheaux


A member of the audience contributed his notes of the actual panel. They appear here with modest editing.

Ralph -- The war for choosing the language for applettes is over--Java is the standard. Java won for marketing reasons, not technical. [Smalltalk community needs to learn to market itself better.] Smalltalk community needs to be more open to what Smalltalk can be used for.

Patrick -- Web browsers are dead. Instead of using Web Browsers with their really boring interfaces, we need to create better interfaces and merely look to the web to provide data.

Ward -- Web years: 5 per year. Legacy years: 1/5 per year. The web moves very fast. What you would normally deliver in a year, you must be able to deliver in 0.2 years on the web. Hypertext: casual, self-explainatory, Objects: modeling, responds to change, Relational/legacy: safe ... All maps well to 3-tier architecture.

Roby -- Everyone asks Roby/ParcPlace: Does VisualWave support Java? Roby's answer: I donno. What do you want to do with Java? Customer's answer: I donno. But I'm sure I need it.

Audience discussion -- You can buy Borland C++ for $89. Students aren't going to buy Smalltalk for $2000. Now Netscape and Java are free and we wonder why it's everywhere.

My own thoughts? -- Why can't Smalltalk produce Java byte-codes? Why not a Java environment that uses Smalltalk syntax, Smalltalk libraries, and produces Java byte-codes? Why not a Java++ VM that extends the Java byte-code set to support blocks, etc.?

Other Press:

David Carr offers this review of the panel for Web Week.

Smalltalk Solutions '96 is presented by SIGS Conferences