Retrospective course rationale, and future study

Why (did you) study languages? (A non-exhaustive list)

• New languages come along all the time, often accompanied by massive hype. You must learn to develop some taste and see through the hype to a language's technical core. The ability to evaluate languages will make you a better programmer, and a better technical leader (whether of a product group, an IT/IS department, or an open source project).
• When you get to the real world, you will be shocked at the diversity of programming languages in actual use. Windows GUI applications written in C++ are only a tiny portion of the software world. Using one language and one environment limits your career choices.
• Languages teach you to think in diverse ways. Each (good) language learned is another weapon in your intellectual arsenal. Good designers use the appropriate paradigm for the problem, whether it be functional, object-oriented, declarative/constraint-based, etc.; and they can apply these paradigms in various implementation languages.
• Some languages make you measurably more productive than others. High-level languages provide you with powerful tools for abstraction and verification that lower-level languages lack, so that you can write code that does more in less time.
• There is no One True Language. No one language is ideal for everything, even within a single project! For example, dynamic languages like Smalltalk are ideal for quickly writing prototypes, to understand your problem domain. Once the prototype is written, you can write the "real thing" in a more efficient, statically typed language. Similarly, if you wanted to write a symbolic manipulation program, Java is probably not the best language---ML, Scheme, or some other Lisp dialect would make your life easier. But you might want to link your Scheme interpreter to C libraries for efficiency in critical sections of code.

Further study: topics for which we had no time

• Exceptions: are not just an ML feature. They are the most common way of expressing error conditions in Java and other langauges. Sun has a lot of online documentation, including a tutorial with a chapter on exceptions; read it at java.sun.com.
• Scripting languages: are lightweight, dynamic languages for rapid, low-overhead development. Often come with lots of convenient libraries. Examples: Python, Perl (blech), Visual Basic (blech), etc. These should be easy for you to pick up.
• Logic and constraint languages: What if you could write code that just specified the logic of a program, but not the control? For example:

  abs(X,A) :- X >= 0, X = A.  /* a CLP(R) "relation */
  abs(X,A) :- X < 0, -X = A.
  
  ?- abs(1, X).               /* CLP(R) query */
  X = 1
  *** Yes 
  
  ?- abs(Y, 2).               /* Another query; notice that it goes */
  Y = 2                       /* in the opposite "direction".       */
  Y = -2
  *** Yes 
  

• Domain-specific languages: not all languages are for general-purpose programming. For example, SQL is a language only for querying databases, and it looks and acts very differently than a general-purpose language.