Putting the Science back in Computer Science

That's interesting. I hadn't run into this before in college or grad school, and it seems like a good way of bridging the analytical gap between theory and practice that Sedgewick talked about.

It's unlikely that Sedgewick hadn't heard of this before. I suspect that he didn't mention it simply because it isn't often taught. According to your citation, it was introduced in 2001 which is quite recent in academic terms. How did you come across it?

Yes, it is quite interesting, and although the main idea is straightforward, the analysis of concrete algorithms gets quite technical. Also, it's still actively researched and maybe not entirely accepted into the mainstream yet. AFAIK it has been successfully applied to many well-known algorithms which have bad worst-case, but surprisingly good behavior in practice (e.g. quicksort). I heard about it from my advisor.

A list of related papers: http://www.cs.yale.edu/homes/spielman/SmoothedAnalysis/index...

I'm not entirely sure that's a cogent argument. Two major objections:

1. If anything, I think the problem with Java is that it encourages too much abstraction. This isn't inherent in the language design, but is a product of Java culture—the massive UML diagrams and fragile class hierarchies which are glorified in Java tend to encourage adding extra abstractions to things which don't really need to be abstracted away, and you spend time doing extra work to accommodate them. There's an attitude among computer scientists, especially of the academic variety, that there's no such thing as "too much abstraction," and while abstraction is generally a decidedly good thing—I have, in my time, written Strategy-Factories, after all—I think it can be possible to overemphasize and overuse it. (John Carmack: "It is not that uncommon for the cost of an abstraction to outweigh the benefit it delivers. Kill one today!")

2. I'm not entirely sure that I agree with the statement that Python has poor data abstraction, especially in comparison to Java. Python in many ways lets you abstract away more details than Java does; e.g. duck typing allows you to treat several pieces of data as identical black boxes without having to worry about the implementation details and without using subtyping or interfaces, which is exactly the definition of abstraction. On the other hand, Java fails to abstract away some of the implementation details which are unimportant to the kind of programming that beginners do, e.g. the differences between int and Integer.

My own two cents is also that I also prefer Python's OOP-but-not-necessarily-always philosophy to Java's OOP-all-the-time-everything-in-a-class-somewhere philosophy. I am involved with teaching a Java course right now, and it has made me really wish that students could start out not worrying about classes and just writing scripts, as in Python, before they have to learn about classes and inheritance and data abstraction. (I have previously taught everything from C to Scheme, and it really is wonderful to just start people out writing code in a high-level Python-like language without having to handwave about static and classes and so forth.)

I really agree with this, but think that we'll see even more of a split than just math/application. I got my PhD in Information Systems and Technology, which fused a social science approach with a topic of information systems being used in different domains. I found it really useful, and think that a business-oriented/programming-oriented/math-oriented division will probably split up the field.

This is an old topic. Auguste Comte and Emile Durkheim, both foundational to the social sciences, each sought to bring empiricism and the scientific method to the study of social institutions. I would argue that they were largely successful and it is a shift in our modern understanding of science that questions whether Social Science is "real" science.

Specifically, I think the power of mathematics as a tool for objective description of natural phenomena has given us the impression that any field without underpinning equations is soft, subjective, and inherently unscientific. Biology is Chemistry is Physics is Math, but the study of social institutions resists reduction to "hard" sciences, to the objectivity of equations. I have even seen people say that Chemistry is not science!

In this view, Sociology seems inherently less legitimate than Biology. And I think that is dangerous. If this is how we are to understand science, then we are doomed to belittle and ignore the possibilities for better understanding society and intelligent behavior.