Previous Table of Contents Next


3.4.4.2. Smalltalk Supports Rapid Change

One of the most interesting applications in Smalltalk-76 was a multimedia document galley editor written by Diana Merry. It incorporated text and graphics into a general framework for managing document elements. Upon hearing of a visit by an influential Xerox executive, LRG agreed to demonstrate Smalltalk and various applications, including the galley editor. The demonstration was performed on a new PARC workstation, the Dorado, another Smalltalk-76 microcoded byte-code interpreter. We defined a filmstrip as a sequence of images and added a filmstrip as a document element. The task was easy because a small number of messages was sent by the document to any element and therefore to which a filmstrip had to respond. The demo was a page duplicated from a Xerox copier manual that displayed a sequence of text and pictures depicting the flow of paper through the copier. As if on cue, the executive wondered out loud whether the paper could really flow, and we selected the run command.

The demo then turned to the Smalltalk programming environment to explain the cross-referencing capability of the class browser. Again, the executive demonstrated his understanding by asking whether the particular pathway to a retrieved class method could be remembered. Tesler immediately said yes (after all, a computer scientist was asked whether he could do something with software—the answer is always yes). But the then head of PARC, escorting his corporate colleague, admonished that we indeed could not. Tesler left in a silent huff. By the time the demo completed, he was back, loaded some code, and gave the demonstration of bookmarking class method access.

A similar story of the rapid change of system software written in Smalltalk was told the day LRG entertained distinguished visitors from Siemens. Glenn Krasner gave the demo that day, carefully explaining the concepts and mechanisms behind the query interface to the Smalltalk system. But it was clear that the visitors were not pleased. Indeed, they were scowling. “Is there a problem?” we asked. “Indeed, there is,” they answered. “Selected text should not be highlighted that way.” We had been using the well-accepted technique of highlighting text by complementing the color of the text from black on white background to white on black. “And what should it be?” we asked. “You should put a box around the text,” they answered—and so Glenn did. Without engaging in the usual human factors arguments that these gentlemen were prepared to hear, Glenn simply demonstrated how to discover the code that highlighted text and replace it with code that created the box. In fact, what Glenn did was interrupt the process of text selection by typing Ctrl+C to open the Smalltalk debugger as an inspector of the activation stack at the point of interruption. In the debugger, Glenn browsed to the method that highlights text and replaced it with a method to draw a box. He then pressed the proceed command and continued the demo.

3.4.5. Smalltalk-78: Portability Matters

The Dorado was not a Dynabook. A Dynabook is something you can take with you wherever you go. Weighing in at less than 15 pounds, a Dynabook is something you can put on your lap and work with for several hours, and your legs stay awake. A Dynabook is small and weighs less, has a lot of battery power, has a good quality display screen, and runs Smalltalk.

The NoteTaker was a step in the Dynabook direction. It was a computer about the size of a small suitcase. It could fit under your airplane seat if you flew coach. A NoteTaker had three 8086 processors, one for the bitmapped display screen, one to run the Smalltalk interpreter, and one to handle I/O. The battery pack could last several hours. Furthermore, Doug Fairbairn’s planned redesign of the NoteTaker packaging represented a significant decrease in size.

The name NoteTaker came from a proposed educational use for the system. Suppose you were teaching a history or social studies course. The class is studying the history of England and France, noticing the intrigue in the two aristocratic circles. Some students hypothesize that the intrigue was fostered by gossip among the servants during official visits. Suppose you could formulate this hypothesis as a formal model of communications and its effect on government decision making. To test the theory, the students could gather information from the library, “taking notes” about patterns of communication and decision making. Suppose that the model was implemented as a computer system and that the students, while in the library, typed their notes directly into the computer as data to test the model. Would such a modeling tool be feasible? Would such a note-taking tool contribute to the students’ ability to learn about and from history?

Xerox had another interesting problem. The Xerox technicians were not doing an excellent job of servicing customers. The documentation they used was large and filled the trunk of a car. Technicians did not carry that documentation into the customer’s office; technicians did not go down to their cars to read out-of-date documentation. LRG proposed that the corporate documentation team consider using the NoteTaker as a new medium for publishing their materials. Technicians could take the computer into the customer office and search in real time for up-to-date information when any questions arose. The documents could be kept up-to-date by frequent downloads over telephone wires.

The year was 1978. People did not carry computers around with them. Not in a library. Not to customer offices.

And so the then head of PARC canceled the NoteTaker project. “No one wants portability,” he said.


Previous Table of Contents Next