7328 Ruby

It's object-oriented. What does that mean? Well, for every ten programmers, there are twelve opinions as to what OOP is. I will leave it your judgment. But for the record, Ruby does offer encapsulation of data and methods within objects and allows inheritance from one class to another; and it allows polymorphism of objects. Unlike some languages (C++, Perl 5, etc.) Ruby was designed from the beginning to be object-oriented.

It's a pure OOP language. Am I being redundant? I don't think so. By this we mean that everything, including primitive data types such as strings and integers, is represented as an object. There is no need for wrapper classes such as Java has. And in addition, even constants are treated as objects, so that a method may be invoked with, for example, a numeric constant as a receiver.

It's a dynamic language. For people only familiar with more static languages such as C++ and Java, this is a significant conceptual leap. It means that methods and variables may be added and redefined at runtime. It obviates the need for such features as C's conditional compilation (#ifdef), and makes possible a sophisticated reflection API. This in turn allows programs to become more "self-aware" -- enabling runtime type information, detection of missing methods, hooks for detecting added methods, and so on. Ruby is related to Lisp and Smalltalk in this respect.

It's an interpreted language. This is a complex issue, and deserves several comments. It can be argued that performance issues make this a negative rather than a positive. To this concern, I reply with these observations: 1. First and foremost: A rapid development cycle is a great benefit, and it is encouraged by the interpreted nature of Ruby. 2. How slow is too slow, anyway? Do some benchmarks before you call it slow. 3. Though some will criticize me, I will say this anyway: Processors are getting faster every year. 4. If you absolutely need the speed, you can write part of your code in C. 5. Finally, in a sense, it is all a moot point, since no language is inherently interpreted. There is no law of the universe that says a Ruby compiler cannot be written.

It understands regular expressions. For years, this was considered the domain of UNIX weenies wielding clumsy tools such as grep and sed, or doing fancy search-and-replace operations in vi. Perl helped change that, and now Ruby is helping, too. More people than ever recognize the incredible power in the super-advanced string and text manipulation techniques. Doubters should go and read Jeffrey Friedl's book Mastering Regular Expressions. So should non-doubters.

It's multi-platform. It runs on Linux and other UNIX variants, the various Windows platforms, BeOS, and even MS-DOS. If my memory serves me, there's an Amiga version.

It's derivative. This is a good thing? Outside of the literary world, yes, it is. Isaac Newton said, "If I have seen farther than others, it is because I stood on the shoulders of giants." Ruby certainly has stood on the shoulders of giants. It borrows features from Smalltalk, CLU, Lisp, C, C++, Perl, Kornshell, and others. The principles I see at work are: 1. Don't reinvent the wheel. 2. Don't fix what isn't broken. 3. Finally, and especially: Leverage people's existing knowledge. You understand files and pipes in UNIX? Fine, you can use that knowledge. You spent two years learning all the printf specifiers? Don't worry, you can still use printf. You know Perl's regex handling? Good, then you've almost learned Ruby's.

It's innovative. Is this in contradiction to #7 above? Well, partly; every coin has two sides. Some of Ruby's features are truly innovative, like the very useful concept of the mix-in. Maybe some of these features will be borrowed in turn by future languages.

It's a Very High-Level Language (VHLL). This is subject to debate, because this term is not in widespread use, and its meaning is even more disputable than that of OOP. When I say this, I mean that Ruby can handle complex data structures and complex operations on them with relatively few instructions, in accordance with what some call the Principle of Least Effort.

It has a smart garbage collector. Routines like malloc and free are only last night's bad dream. You don't even have to call destructors. Enough said.

It's a scripting language. Don't make the mistake of thinking it isn't powerful because of this. It's not a toy. It's a full-fledged language that happens to make it easy to do traditional scripting operations like running external programs, examining system resources, using pipes, capturing output, and so on.

It's versatile. It can do the things that Kornshell does well and the things that C does well. You want to write a quick ten-line hack to do a one-time task, or a wrapper for some legacy programs? Fine. You want to write a web server, a CGI, or a chess program? Again, fine.

It's thread-capable. You can write multi-threaded applications with a simple API. Yes, even on MS-DOS.

It's open-source. You want to look at the source code? Go ahead. Want to suggest a patch? Go ahead. You want to connect with a knowledgeable and helpful user community, including the language creator himself? You can.

It's intuitive. The learning curve is low, and once you get over the first hump, you start to "guess" how things work -- and your guesses are often correct. Ruby endeavors to follow the Principle of Least Astonishment (or Surprise).

It has an exception mechanism. Like Java and C++, Ruby understands exceptions. This means less messing with return codes, fewer nested if statements, less spaghetti logic, and better error handling.

It has an advanced Array class. Arrays are dynamic; you don't have to declare their size at compile-time as in, say, Pascal. You don't have to allocate memory for them as in C, C++, or Java. They're objects, so you don't have to keep up with their length; it's virtually impossible to "walk off the end" of an array as you might in C. Want to process them by index? By element? Process them backwards? Print them? There are methods for all these. Want to use an array as a set, a stack, or a queue? There are methods for these operations, too. Want to use an array as a lookup table? That's a trick question; you don't have to, since we have hashes for that.

It's extensible. You can write external libraries in Ruby or in C. In addition, you can modify the existing classes and objects at will, on the fly.

It encourages literate programming. You can embed comments in your code which the Ruby documentation tool can extract and manipulate. (Real fans of literate programming may think this is pretty rudimentary.)

It uses punctuation and capitalization creatively. A method returning a Boolean result (though Ruby doesn't call it that) is typically ended with a question mark, and the more destructive, data-modifying methods are named with an exclamation point. Simple, informative, and intuitive. All constants, including class names, start with capital letters. All object attributes start with an @ sign. This has the pragmatism of the old "Hungarian notation" without the eye-jarring ugliness.

Reserved words aren't. It's perfectly allowable to use an identifier that is a so-called "reserved word" as long as the parser doesn't perceive an amibiguity. This is a breath of fresh air.

It allows iterators. Among other things, this makes it possible to pass blocks of code to your objects in such a way that the block is called for each item in the array, list, tree, or whatever. This is a powerful technique that is worth exploring at great length.

It has safety and security features. Ruby borrows Perl's concept of tainting and allows different levels of control (levels of paranoia?) by means of the $SAFE variable. This is especially good for CGI programs that people will try to subvert in order to crack the web server.

It has no pointers. Like Java, and with a grudging nod to C++, Ruby does not have the concept of a pointer; there is no indirection, no pointer arithmetic, and none of the headaches that go with the syntax and the debugging of pointers. Of course, this means that real nuts-and-bolts system programming is more difficult, such as accessing a control-status register for a device; but that can always be done in a C library. (Just as C programmers drop into assembly when necessary, Ruby programmers drop into C when they have to!)

It pays attention to detail. Synonyms and aliases abound. You can't remember whether to say size or length for a string or an array? Either one works. For ranges, is it begin and end, or first and last? Take your pick. You spell it indices, and your evil twin spells it indexes? They both work.

It has a flexible syntax. Parentheses in method calls can usually be omitted, as can commas between parameters. Perl-style quotes allow arrays of strings without all the quotation marks and commas. The return keyword can be omitted.

It has a rich set of libraries. There is support for threads, sockets, limited object persistence, CGI programs, server-side executables, DB files, and more. There is some support for Tk, with more on the way.

It has a debugger. In a perfect world, we wouldn't need debuggers. This is not a perfect world.

It can be used interactively. Conceivably it could be used as a sort of "Kornshell squared."

It is concise. There are no superfluous keywords such as Pascal's begin, then after if, do after while. Variables need not be declared, as they do not have types. Return types need not be specified for methods. The return keyword is not needed; a method will return the last evaluated expression. On the other hand... it is not so cryptic as C or Perl.

It is expression-oriented. You can easily say things like x = if a<0 then b else c.

It is laced with syntax sugar. (To paraphrase Mary Poppins: A spoonful of syntax sugar helps the semantic medicine go down.) If you want to iterate over an array x by saying for a in x, you can. If you want to say a += b instead of a = a + b, you can. Most operators are really just methods with short, intuitive names and a more convenient syntax.

It has operator overloading. If I am not mistaken, this originated long ago in SNOBOL, but was popularized more recently by C++. It can be overdone or misused, but it can be nice to have. Additionally, Ruby defines the assignment version of an operator automagically; if you define +, you get += as a bonus.

It has infinite-precision integer arithmetic. Who cares about short, int, long? Just use a Bignum. Admit it, you always wanted to find the factorial of 365. Now you can.

It has an exponentiation operator. In the old days, we used this in BASIC and FORTRAN. But then we learned Pascal and C, and learned how evil this operator was. (We were told we didn't even know how the evaluation was done -- did it use logarithms? Iteration? How efficient was it?) But then, do we really care? If so, we can rewrite it ourselves. If not, Ruby has the good old ** operator you loved as a child. Enjoy it.

It has powerful string handling. If you want to search, substitute, justify, format, trim, delimit, interpose, or tokenize, you can probably use one of the built-in methods. If not, you can build on them to produce what you need.

It has few exceptions to its rules. The syntax and semantics of Ruby are more self-consistent than most languages. Every language has oddities, and every rule has exceptions; but Ruby has fewer than you might expect.