(2022-07-22) An Interview With Ted Kaehler

An Interview With Ted Kaehler. The following is an interview conducted via email (September 2021) with Ted Kaehler, who was in the unique position to have worked extensively both on Smalltalk and HyperCard.

Worked on Smalltalk 72, 74, 76, 78, and 80 at Xerox PARC.
Worked on HyperCard and Squeak at Apple.
Worked on Squeak Smalltalk and Etoys at Disney Imagineering.
Continued work on Squeak at ViewPoints Research, Hewlett-Packard Research, SAP, and Y-Combinator’s HARC.

1. Did you and the other creators of Smalltalk have a conscious goal to empower users to inspect and modify their software? If so, what motivated that goal?

Yes, absolutely. Alan Kay wanted to build a “personal computer for all ages.” Just as writing and music are each a kind of “literacy”, so the ability to build simulations and play with them would be a new literacy.

2. Were there any changes to Smalltalk over time that affected how easily users could inspect and modify code?

Larry Tesler created the “Code Browser” to let users see the classes and methods of the entire system. Dan Ingalls implemented the first opaque overlapping windows to let users see more code and other objects on the screen.

We avoided putting code in files so the user would not have to think about or manage files.

Dan and Alan made each class be an object on equal footing with other objects. While LISP had all parts of the system accessible to programmers, having classes as objects made this much clearer and easier. Code written by users is treated the same as code that came with the system.

3. What got you interested in working on HyperCard?

Dan Winkler did a superb job creating the scripting language.

HyperCard came very close to being the World Wide Web six years before the web. Stacks could be on servers and users could browse them. A stack could be a mini-application on a server. We just didn’t push it hard enough. Connections were only over AppleTalk and not the Internet. It wasn’t legal for a company to be on the Internet in 1987-88.

4. When you began working on HyperCard, did you see it as a continuation of some of the themes of Smalltalk’s development experience, or as something very different?

HyperCard evolved a lot during its development and Alan Kay was advising Bill Atkinson.

HyperTalk scripts were command-based. We pushed for letting scripts send messages to objects. The objects were cards, text fields, buttons, backgrounds, and stacks. Bill improved that by making dozens of built-in messages from events, such as mouseDown.

*We knew that Smalltalk was weak as a vessel for holding and showing a lot of text and images. It was not very easy for end users to create cards and fields in Smalltalk. HyperCard did that so well. It was one direction we wanted Smalltalk to go, and HyperCard took it very far.

Scripting in HyperCard is very accessible, and that certainly furthered a goal we had for Smalltalk.*

5. HyperCard has the concept of user levels, whereas Smalltalk does not. What difference did that make in how the two systems empower users?

User levels are good, especially when a teacher wants his students to have a certian experience, and to not look at the answers under the hood. But, in the end, we very strongly want every user to be on the author level.

Later, we built an end user system in Smalltak called Etoys

6. What were some of the things people have created with Smalltalk and HyperCard that impressed you the most at the time?

Apple estimated that HyperCard had six million author-users, many of them teachers

Impressive Smalltalk programs were Alan Borning’s ThingLab, a graphical constraint system. Adele Goldberg’s job shop simulation that we showed to the Xerox executives was impressive. So was Steve Weyer’s Library card catalog and checkout system in ST-74.

7. Do you think users today would benefit from more ability to inspect and customize their software?

Absolutely! It is sad that we have lost this capability in so many apps.