HtDP is the textbook for the curriculum approach described in the paper. It's supported by the student-oriented DrRacket IDE (which is included with a normal Racket install): https://www.racket-lang.org/
Besides HtDP as an intro course for undergraduate CS majors, they also have an interest in teaching program design as a part of basic literacy, including to younger students.
Sadly, you are right. SICP has vanished from most introductory courses. CTM a fantastic sequel, more oriented towards formal semantics, has never really made it into mainstream.
A central topic in CS studies should be formal system semantics. These days, students only get some computability and complexity notions. Usually embedded in an algorithms course. At that's it.
Which is sad. You can get a lot of insights into CS, and Software Engineering as true engineering, by going through some basic semantics notions. Take a theorem prover, and teach students how to model different computation systems and verify some properties on them.
If time allows, expand into more efficient but less general approaches like abstract interpretation, static analysis, model checking and type/effect systems.
You don't need a ton of background. Some beautiful texts cover these with just FOPL [1,2].
Unfortunately, the accompanying Mozart 2.x interpreter/programming environment doesn't yet support some important features of the Oz language, and the old 1.4.x implementation is so outdated the easiest platform to install it for is... Windows. The macOS version seems oblivious to the relatively recent tightening of app permissions, and Linux is something I have yet to figure out.
Quote 1 "Clearly, a computer science curriculum must not, and doesn’t have to, become a vocational training ground for the latest industrial programming language and programming tools. Superficial aspects of industrial practice change as fast as fashion trends. No academic department can switch its course content fast enough and maintain a curriculum that passes on tested wisdom."
Quote 2 "Applied to the first-year courses, these suggestions say that the year should start with a heavy emphasis on principles and should add some industrially relevant components during the second semester. Even more precisely, the first semester should emphasize programming principles and habits; the second part should illustrate the use of these principles in currently fashionable programming languages. Of course, the “principled” semester may integrate fashionable parts where they aren’t an obstacle, and, more importantly, the “fashionable” part of the first year must continue to practice good design habits."
First, organizations that employ software developers don't choose programming languages and toolchains because they are "fashionable". They choose them because of proven utility and effectiveness, prior experience, employee knowledge base, and inertia. Introducing the "fashionable" language to your organization typically requires overcoming considerable resistance from both management and peers. As a result the turnover of languages used in any particular sector of software development is really slow. The changes in relative frequency of use of programming languages more often reflects new areas of application, such as mobile, bioinformatics, machine learning, etc where early practitioners set precedents for the use of a new language.
The age of the top languages in Tiobe Index are C, 47; C++, 34; Python, 29; Java, 24; JavaScript, 23; C#, 19; and VisualBasic.net,17.
Yes there was. Java’s adoption itself was largely a result of corporate fashion and was brought into university and high school curricula to meet their needs. Fashion has always been part of the industry in some way.
Java was a very plausible language for teaching software development at the time. It had memory safety, good multithreading support, static typing, exceptions with some static checking, and supported OO. The sandboxing and VM was also appealing, for tools support and research.
OO was very popular, and not just for fashion: (1) academia was concerned with things like interacting objects as models of computation; and (2) OO methodologies of the time were very useful tool for practical reasoning about systems designs, and Java arguably supported mapping design models to code better than any non-research language of the time (though there's still a gap).
I was involved with a university CS department effort to learn about and evaluate many OO languages, for rethinking intro CS education, and I recall Java did very well. This was before most companies other than Sun had ever heard of Java, so I'm not aware of any employer push at that time.
Turning exciting Java (which initially seemed targeted at the kind of programming work then using C++, not corporate MIS programmers using 4GL frameworks and such) into a bureaucratic enterprise language, came later.
Regarding pre-corporate university interest in Java, I can mainly tell you what I recall, but you can still see this 1996 Emacs mode for writing in an assembly language for the JVM (like a student might go to the trouble to spend time on, if Java had research target appeal): https://www.neilvandyke.org/jasmin-emacs/
In 2004, Java had been out for nearly a decade and was Java 5 at that point with the collections framework maturing the standard library.
At that point, Java had just as much maturity as Go (2009 release) or Powershell (2006 release) does now.
Java is only three years younger than Python. It would be surprising to see people claim that programming python in 2001 was done because it was fashionable.
Surprising quote: "Several instructors reported that female students like the HtDP curriculum exceptionally well. In a controlled experiment, an htdp-trained instructor taught a conventional AP curriculum and the Scheme curriculum to the same three classes of students. Together the three classes consisted of over 70 students. While all students preferred our approach to programming, the preference among females was a stunning factor of four. An independent evaluator is now investigating this aspect of the project in more depth."
That said, what I really liked about my SICP course at Boston Univeristy in 1987 was the fact I could do all the exercises on our office PC (an 80286), with Emacs as the IDE. Quite heady stuff back in day.
