I'm pretty sure I'm missing something here, but using HyperCard= to present practice problems is a lot different from writing a= program to solve the problems. Or did the stacks solve the problems themselves?

Insofar as the stack had to figure out the right answer, yes. I'm practically positive the right answers were algorithmically-derived, not handed to the program up front -- in at least one case I remember, at least two correct solutions were possible, and the program accepted two (the two I tried, anyway) as correct. Essentially, the stack would show you a problem, let you solve it, and tell you whether you were right or wrong. If you were wrong, it would show you the problem ("parallel fifth here" "that isn't really a ii chord") so that you could revise your solution.

I understand that natural language is a wonderful thing, and a lot of the interesting stuff in CS is closely related to language. I just think that it would be a hard way to teach programming and still

teach language concepts at the same time. Can writing a program give greater appreciation for a work of literature? Maybe it can, and I would love to see some examples. With math, however, programming can uncover things that would otherwise go unnoticed, and the programming is relatively simple.

Same for language (which is not the same thing as literature; let's not get our fields mixed here). It's relatively simple to write a program to produce regular verb paradigms from natural-language-of-choice. In the process, though, students absolutely will uncover a lot of knowledge about language they didn't know they had. Toss irregulars into the mix (easier for Romance languages, harder but still possible for English), and you have a winner -- tough-but-interesting programming problem that touches on a significant pedagogical problem. As for literature, I'd turn kids loose on writing a concordancer in a hot minute. It's absolutely astonishing what you can learn about an author's use of language from a concordance. For English poetry, I think stress analysis would be fun (especially combined with a pre-written speech synthesis module). If you stress a poem precisely according to The Rules, does it sound right? Why or (more often!) why not? What departures from The Rules do poets make? Does a poem that precisely fits The Rules necessarily sound better than one that breaks them every now and then? (For Spanish poetry, syllable analysis is the analogous-but-not-identical problem.)

Other subjects would require a pretty extensive re-evaluation of current curricula and teaching methods

This is probably true, but once again, if our goal is genuinely Programming for Everyone, we can't balk at that. In a way, too, part of what I'm putting forward is Natural Language for Programmers, who are reputed (how fairly or unfairly I can't say, not having taught a large enough sample) to have difficulty learning foreign languages. I think the two-way enrichment is part of what CP4E is after. Dorothea -- Dorothea Salo Impressions Book and Journal Services, Inc. phone: (608) 244-6218 fax: (608) 244-7050 http://www.impressions.com