CS 536 (F02) Homework 2: Evaluating Languages

Assignment Goals

• To understand how you would evaluate a language without before having specific training in programming languages. Given this goal, I won't tell you now how to evaluate a language. Focus on the aspects that you think are important, and list any aspects that you are consciously ignoring because they seem unimportant.

• To make you think about languages that lie outside of your current repetoire. Languages shape thought, so the broader your linguistic range, the bigger the toolkit you can apply to programming problems.

The Assignment

1. Choose a language to evaluate. See the next section for my expectations for the language you choose. If you choose something other than one of the suggested languages, you must clear your choice with me by email before doing this assignment.

2. Write a one-page essay excluding bibliography, pictures, and large code fragments that evaluates your chosen language. Good, clear writing counts! Pay attention to the quality of English writing you do on this assignment (I'm aware of — and sensitive to — the fact that some of you are newly in the USA and haven't had much practice with written English. This essay is a chance to practice. The writing center can help you with your written English skills.)

   Do this assignment on your own, without collaborating with other students. Usual standards of academic honesty apply; if you use material from the Web, cite it! (And be sure it's trustworthy ...) Don't quote large segments of text or code from a Web site or other source as a surrogate for doing your own writing! See the collaboration policy for more details.

Which Language Can You Use?

You may choose to evaluate one of the specific languages listed below, or you may choose your own; in either case, choose a language that you do not currently know. Some possibilities include:

• Erlang
• NESL
• Ruby

If you choose you own you must clear your choice with me by email before doing this assignment. Your proposed language should satisfy the following criteria:

• No large mainstream languages like Java, C, C++, Fortran, Pascal, Perl, Python — get outside the box!
• Pick something in contemporary use!
• Pick something reasonably high-level, in which you can implement significant systems.
• No toy languages. Lots of compiler and programming language courses, for instance, create little toy languages as the object of study in the course. Skip these.
• No phantom languages. The Web has no shortage of proposed programming language projects that never get off the ground (or at least not more than a few inches off of it). Skip these as well. Make sure your chosen language has an implementation and at least a small, real user base.
• Avoid overly specialized languages. There are many languages designed to cater to very specialized niches. A language whose design was motivated by particular domain is okay, as long as you could reasonably implement some programs for other domains in that language.
• Avoid the temptation of choosing one of the billion new ``scripting languages''. This past decade has seen an explosion in their number, but all they have to distinguish themselves is sometimes nothing more than a logo. In particular, every software company that had its own hacked-up in-house language seems to have decided the smart thing to do is to get other people to do some of their work for them — so they release it as an Open Source effort, and hope lots of people will pick it up and hack on it. The end result is, of course, lots of languages each getting very little attention. Do everyone a favor and pay them even less attention! (Aside: The proliferation of these in-house languages lends credence to the premise of this course, which is that you may well find yourself designing languages sooner than you think.)
• No proprietary languages for which you (or I!) can't find any useful documentation!

Overall, be sensible.