Java nested classes - tips and tricks

I had to prepare this anyway, so I thought I might as well post. 

First of all to get the terminology straight: normal classes and interfaces are top-level. But class or interface can also be nested, if it is defined inside another class or interface. Nested classes originate in Beta programming language, and are available in Java since version 1.1. Non-static nested classes are called inner. Inner classes can be members (declared immediately inside outer class definitions), like this:

class Outer { class Inner { ... } }

Inner classes can also be declared within methods and other blocks, then they are called local, like:

class Outer { void foo() { class Local { ... } ... }

A more popular breed of local inner classes are anonymous classes: 

class Outer { void foo() { ... new Bar(...) { ... } ... }

For visual impression - check out this diagram.

Tip - construction: Creating new instance of nested or inner class within the outer class is simple. From outside it's a bit trickier, suppose we have a class like this:

class Outer {

   static class Nested { ... } 

   class Inner { ... } 

} 

We can reference and instantiate the classes like this:

Outer.Nested nested = new Outer.Nested(...);

Outer out = ...  

out.Inner in = out.new Inner(...); //translated by javac to new Inner(out, ...)

The magic behind inner class constructor is that compiler implicitly adds Outer parameter to all Inner constructors, and passes the enclosing instance when constructor is invoked. From then on, inner class instance (for its entire lifetime) holds a strong reference to the enclosing instance.

Tip - instanceOf: If we have two distinct instances Outer out1, out2 then out1.Inner and out2.Inner denote the same class, but Inner instances will refer to a different enclosing instances. This is different from Scala and Newspeak, where inner class is distinct for every enclosing instance.

Tip - access enclosing instance: To access the instance of outer class from within a contained inner class use Outer.this. Nested/inner class methods/fields hide outer class ones, to access outer class elements prefix them with Outer class name, e.g. Outer.staticMethod(...) or Outer.this.anyMethod(...). 

Tip - nesting and inheritance: Generally, method lookup rules in Java nested classes follow "comb semantics" - first search inheritance hierarchy, then enclosing lexical scopes. This behavior can introduce some wierd puzzlers, like #9 here. In Newspeak the enclosing scope is considered before inheritance, which makes it easier to follow from programmer's perspective.

Tip - generics: Generics type parameters of enclosing class (or method) can be used within inner classes.

Tip - interfaces: Interfaces may have nested classes (necessarily and implicitly static), it may be particurlaly useful for declaring nested enums.

Tip - statics: Inner class cannot have static declarations in it, except compile-time constants. To overcome this, static declarations canbe moved to the top-level class.

 

Trick - loading: Nested class is treated just as any other class by the JVM, e.g. it is not loaded/initialized until used. This fact is used to implement thread-safe lazy singletons using the Holder pattern.

Tip - final: Anonymous inner classes and local classes can access variables in the surrounding scope only if the variables are final:

void invokeAnon(final int number) {

   final String word = “hello”;

   someObject.pass(new Runnable() { 

      public void run() { 

         System.out.println(word.substring(number));

      }

   });

} 

Trick - double braces: Double brace initialization is a trick of putting initialization block inside anonymous inner class declaration, like:

   final Map numbers = new HashMap(){{

      put("one", 1);

      put("two", 2);

      put("three", 3);

      //...

   }};

Trick - tokens: A cool Generics trick that uses local class to capture type parameters is super-type-token, a.k.a. Gafter Gadget. 

Tip - reflection: Starting from Java 5 a bunch of methods have been added to reflection with regards to nested classes. For example, Class#getEnclosingClass() method will let you find out the enclosing class for an inner class, for example:

class Enigma {    

 final static Class MY_CLASS = new Object(){}.getClass().getEnclosingClass();  

}  

Prolog: Last, but not least JLS is the ultimate resource for finding out more.

Oh Null, Null

Introduction

Some say it’s one of the worst inventions in history of programming languages, some say it is number one cause of errors in Java programs. Anyway you look at it - null pointer (in all its incarnations) is a sensitive subject. I wonder why there is no song about it, there are several funny re-works of Beatles songs with geeky lyrics, like Unix Man (Nowhere Man, “he’s a real Unix man sitting in his Unix LAN, making all his UNIX .plans for nobody…”), Write In C (Let It Be), Yesterday (“yesterday, all those backups seemed a waste of pay…”), Something ("Something in the way it fails, defies the algorithm logic...") and a not so funny Eleanor Rigby. I’d like to write a new song, “Null” using the “Girl” tune. Only my inspiration ended after a single verse:

If there’s anybody going to listen to my story, 

All about the Null who brought the fall.

It’s a value that you want so much, it makes you sorry,

Still you don’t regret a single call. 

Oh Null, Null…

I also have a middle verse inspired by Smalltalk, where the equivalent of null is nil, a keyword that returns an object of class UndefinedObject. It has no methods, except isNil and notNil that return true and false respectively, so nil responds to any other method invocation with “DoesNotUnderstand: <method name>”, and in my experience, at least, pretty much equivalent to NPE in Java. Interestingly enough, though, Objective-C turned nil into a “black hole” that returns self upon any invocation.

Nil’s a kind of null, she is not answering your calls,

you feel a fool (fool fool).

When you think the code looks good, she answers “it’s not understood”, 

she’s cruel (cruel cruel).

Ah null, null…

From Beatles to Elvis

One situation where nulls are very annoying, is when we have a “chain” of method invocations, such as account.getOwner().getAddress().getStreet(). If any of the methods return null, we get NullPointerException from the next method invocation. I previously suggested a solution to the problem, but now (due to popular demand) there is a proposal for Java 7 language enhancement, called The Elvis Operators.  One of the stronger arguments for this solution is that it is available in other languages like C# and Groovy, even if not everybody is entirely happy with it. Perhaps this is subjective, but after all, programming language design is about human perception and feelings, as much as about computer capabilities. So, I admit: I feel uneasy with this new way of method invocation.

The Demeter Law Suggestion

One of the more controversial guidelines of object-oriented programming, Demeter Law is related to “Tell Don’t Ask” and usually summarized as “only talk to your close friends”. It prohibits series of method invocations like the one described above; in fact, more than one dot in series of method invocations is disallowed (unless the chain originates in a this object’s member (?)). There’s a good description of the “The Law” in Pragmatic Programmers article; it was originally discovered and named by Karl Lieberherr, who created the Demeter Project  and wrote the book named “Adaptive Object-Oriented Software - The Demeter Method”. Martin Fowler demoted the Law to Suggestion, and in general, Demeter managed to generate a lot of controversy and debates inside both Java and Ruby communities, the debates ranging from intellectual (almost academic) polemics, to flamed battles between consulting firms over preferences in programming styles, music and hairdo. Yes, really. Anyway, if you believe in Demeter, our use-case for “null-safe” method invocations is flawed to begin with. But do we believe in Demeter? It took me a while to form an opinion.

1) I do believe in “Tell, Don’t Ask” and the internal iterator (as presented in Pragmatic Programmers article) to be preferable over external iterator (when applicable, of course). 

“Tell Don’t Ask” begs for closures in the language. For those who say “closures are not object-oriented”, Smalltalk not only supports closures, and with minimal syntactic overhead, but basic things like if and while are designed around closures and would not be possible otherwise. 

But "Tell Don't Ask" goes much deeper and wider. For example let's look through Demeter's eyes on service locator vs. injection: locator is wrong, because you go to a factory to get a factory to get a service. Injection is right, because you pass yourself to injector, which sets your the dependencies directly. How about ORM? How much grief is in loading strategies, because we do getContainer().getItem().getProperty(), vs. calling database.execute(query)? Food for thought.

2) Now the paper boy example – this one I do not buy. First of all, we don’t pay for things in cash these days – we give the seller our credit card. Not very secure? Maybe. But much more handy. 

People should generally stop obsessing around security in software, I think. At one point a highly ranked architect in the firm I worked for said “we can’t deploy any software in the browser because it is not secure”. Yes, it is not secure, it’s a problem and people are working on it. And yet, even as it is, is it secure for the jeweler store to put some jewels in the front window? Why not keep them all in the backroom safe?  Because they won’t have buyers, that’s why. What good is all the security, if you go out of business? 

Back to paper boy, the example also does not hold if you are a large organization – the receiver of the service is not the one who authorizes the payment, and not the one who handles the payment. The supplier may wish for “direct communication” as much as he wants, but the rules of corporate procurement are quite different. Then again, it may be silly to attack the example…? Well, the thing is that I hope to show later that you can’t follow Demeter and scale, just like the example doesn’t.

3) Now there’s “hard to mock” argument. Ok, at risk of starting a flame war here, I mean: I love unit-testing just like the next guy, and I know how useful it is, but let’s not get carried away here. 

Unit-testing, just like static types/compilation, or static analysis, is a means to achieve code quality, not the goal in itself. If it helps you – great, if it gets in your way – ditch it. I think the “100% test coverage” is a fallacy. Dijkstra said “tests only prove presence of bugs, never their absence”, which is even truer for unit-tests. (I heard “we have unit-tests, we don’t need QA” argument once or twice. Yeah, right.) 

Anyway, “hard to test” may be a smell, but unless the suspicion can be substantiated – sorry, circumstantial evidence not accepted.

4) So even though I partially agree that “Demeter violation” is a symptom of a problem, the solution proposed by Demeter is, in my opinion, simply absurd. The grand huge fat mega-façade? Dots replaced with underscores? It cannot scale and it is sweeping design problems under the carpet. I mean, give me hierarchies of nested objects that support subsumption, like Newspeak modules – now we are talking. But mega-façades? No thanks.

So there are good things that result from minding Demeter, but not following the law per se. 

a) For example, “don’t return a java.util.List from your API when you don’t want clients adding elements to the list” – that’s basics of any sound API design – return the minimum interface that the client needs, Josh Bloch has been saying this repeatedly. 

b) We should consider to proxy the returned object (perhaps with our own (inner) class), and then we’ll have control over what the returned object’s methods do.

c) Applied to DI – “do not depend on module A just to give you access to another module B, instead depend on B directly”, is also a reasonable rule of thumb. 

But let’s now go back to our beloved nulls.

Power to the method - let the declaration site decide

Demeter or not, methods are meant to do things. I would argue that typical methods should not be assumed to be wrapping field access, or performing trivial computations. Yeah, I know we all do POJOs, which are really POCS, Plain Old  C Structs in disguise.  (And what makes them structs are frameworks that assume a dumb accessor and mutator, and freak out if they are not.) I believe that most methods should do something, so invoking them on a null target should yield an error, not some ambivalent null… So I am not too happy with changing the programming language to accommodate to behavior that should have been atypical.

Now if some methods are “dumb” and have reasonable defaults (like null for property accessor in null object, or 0 for size/length of a null collection/array, or null for a null string upper case transformation), then the method implementer should decide – not the caller. And again, it’s not problem-free, but worth exploring. Somewhat similar behavior has been proposed by Jacek here.

Suppose at the method declaration we add a default clause: either default <value>, or default <expression>, or default <code-block>, with the return type of value/expression/code-block similar to method return type, and parameters and exceptions (in 2 latter cases) similar to the method’s ones. BTW default is an existing Java keyword, used in switch statements. 

Now examples:

public int size() { return this.size; } default 0; 

public boolean isEmpty() { return this.size == 0; } default true; 

public Person getOwner() { return this.owner; } default null; 

public Iterator iterator() { return … } default Collections.emptyList().iterator(); 

public Something produce(Param1 p1, Param2 p2) throws E1 {

   //normal method code here

} default {

  //default calculation, similar to static method, can use p1 and p2, throw sub-type of E1

  //needs to return Something

} 

The location of the “default clause” can be after method body (and if we stick to coding convention of keeping default it on the same line as the closing curly brace, I think it is preferable), but it can sit also between end of method parameter list and the throws clause, or between the throws and the method body start. I think tail location is preferable, because it reminds of the switch statement. 

Another variation that increases similarity to switch may be adding a colon (default : ) and requiring return keyword, like default : return null; instead of default null. Also, the support for “defaults” can be added gradually, with only values in the first stage and more later if it proves to be successful. Also, alternative to default may be case null : … .

The method would be compiled now additional synthetic static method, e.g. ___size(), with the same signature and throws clause, the synthetic method returning the “default” body provided by the user. 

Unfortunately, though, call sites are affected by such method definition and compiled to something like:

SomeType object;

if (object != null) {

  object.m()

} else {

  try {

     SomeType.___m();

  } catch (NoSuchMethodError e) {

    throw new NullPointerException();

  }

} 

So the “default” would be calculated based on declaration type of the object, just like any static method invocation. Why catching NoSuchMethod? Well, what if SomeType class is recompiled alone, and the default clause, and with it the ___m() method, are gone? We want NPE to be thrown, not NSME. 

If, on the other hand, we add the default behavior without recompiling clients, they’ll be still throwing NPEs, I think we would not want all method invocations to become complex byte-code like the one described above… If I understand Jacek’s proposal correctly, unlike me, he suggests every method invocation to go through a static method. 

It would be interesting to see to what extent JIT could optimize code like that when it learns that object is not null… could it throw away everything but object.m()? This certainly requires more investigation.

Maybe, babe, boom

Another option would be to turn to Functional Programming for an answer. This is one of the things known in FP world for ages: Maybe in Haskell, or Option in Scala and F#. Here’s a nice description by Debasish Gosh, and if you want to learn about Monads, James Iry’s blog or a Haskell book would be a good place to start. But we don’t have Monads in Java, Stephan Schmidt tried to simulate Maybe with Iterable, and the result is cute, very cute even, but lacks the monadic awesomeness of flatMap, as others have pointed out. Hm, but wait, we do have a monad in Java. That’s exceptions. The naïve translation of Maybe in Java would be:

String street;

try {

  street = account.getOwner().getAddress().getStreet();

} catch (NullPointerException e) {

  street = null;

}

Ok, it’s not so clean and pretty as pattern match, but… well, that’s Java. No, wait, smart and experienced programmers like us – we know that we shouldn’t use exceptions that way, because it is expensive! Hm, and why is it? Because filling in the stack trace, actually, otherwise it’s rather cheap. But we don’t need stack trace here, only… the exception is the standard NPE, we cannot subclass it and override fillInStackTrace(). Bummer.

Using exceptions for control flow is debatable. I am not a proponent of this style, just for the record. I am simply exploring options… what if the JVM “magically” knew not to fill in stack trace when it’s not used? Exceptions would become cheaper and whole new programming style in Java would emerge...? 

Some of the Java 7 language enhancement proposals are to do with exceptions, multi-catch for example, I am all for it BTW. Now what if we add another flavor to catch and introduce “no-instance” catch, such as:

try {

  //…

} catch (NullPointerException) {

  //…

}

Note that there is no e, no exception instance, no way to re-throw or print – clearly no need for a stack trace. What if JVM kept a bit-map of exception types in “no-instance” try/catch blocks on the stack, and threw these exceptions without filling in the stack trace? After all, we are guaranteed that they are caught and no “phantom” exception like this escapes to the user… these exceptions could be pooled too. Then we could say

String street;

try {

  street = account.getOwner().getAddress().getStreet();

} catch (NullPointerException) {

  street = null;

}

And if we had closures, we could even make it prettier. We could create a utility method like

public static <T> T nullsafe({ => T} expression) {

  return nullsafe(expression, null);

}

public static <T> T nullsafe({ => T} expression, T default_value) {

  try {

    return expression.invoke()

  } catch (NullPointerException) {

    return default_value;

  }

}

It would be invoked like this:

String street = nullsafe({ => account.getOwner().getAddress().getStreet() });

In all its glorious maybiness and with no, or little, performance overhead?! :-) But hey, I said I don’t like clients deciding on the value that the method returns… true, which is why I am not entirely happy with any of the proposals. But at least here we don’t change the language for the sake of null alone – the “exceptional” invocation is explicitly made with a catch clause or via a special library call. 

Epilog

All this just for the sake of intellectual exercise, so don't take too seriously. :-)

Update: 

So why I don't like any of the solutions, not even the ones I proposed? The "default clause" is essentially another static method. It may even work for interfaces, with a slight twist. The real problem is that static cannot be overridden, unlike the method it is "attached to", so the default clause and the actual method body will not correspond, so we basically gained nothing. 

As for the instance-less exception, call me old-fashined, but I think that exceptions are for exceptional things, and this is stretching the hacks around them just one bit too far. 

So I'll stay with plain old null for now, thank you.

The return of forgotten design patterns

Some design patterns are used all the time and their names are known to all - like facade, factories and proxies. Some design patterns are more popular than they should be. But some, although rarely mentioned by their name, have been recently "rediscovered". I'm talking about Flyweight and Memento.

Flyweight
This one basically lets us share n instances between m concurrent clients when m > n. It separates "intrinistic state", the normal class members, and "extrinistic state", which is maintained via parameter passing and return values. Make your intrinistic state immutable, and you can share the same instance between multiple clients. Cool, ha? Priceless. Look at message passing concurrency instead of shared memory concurrency, REST, transactionless architecture... All these treasures actually follow the same spirit as Flyweight.

Memento
This pattern suggests that if you want to save your object state, you better export it in a new dedicated memento object and store the memento. Then restore your object from the memento.

This is like serialization, only serialization didn't follow the pattern, unfortunately.

Josh Bloch suggests we do it manually with so called "Serialization Proxies" - see item 78 in chapter 11 of 2nd edition of "Effective Java". Here's a slide from JavaOne 2006 preso:

The book lists more advantages of the pattern, like improved security (see item 76 - danger of hackers acquiring references to private fields using de-serialization), ability to de-serialize a different class from the originally serialized instance (RegularEnumSet and JumboEnumSet example) etc.The name "memento" isn't mentioned though.

Now imagine persistence architectures actually using intermediate memento objects... instead of modifying actual objects bytecode, breaking encapsulation with access to their private fields, imposing constraints like public empty constructors and so on... Maybe we would have been better off with mementos...?

Types and components

Just some thoughts following a recent conversation I had. Don't we always want static types? If we can detect errors in our program, why not do it as early as possible? Sure. So when is "as early as possible"? I think the answer depends on what our program is - is it one monolithic piece or a component?

In the past monolithic software was prevalent, but today I would bet that most software is meant to be a component. Just look at open-source as it used to be before Java - you'd usually download the C source-code, change it if you need, build it all on your machine, and create one big executable. (And BTW - congratulations, you are a geek.) Cross platform Java changed the situation. Now open-source projects give you a jar to download, you put it on your classpath and you are ready to go. This lowered the bar for open-source adoption, and contributed (aside less restrictive licenses and other factors) to baby-boom of open-source frameworks and tools. Nowadays a Java developer cannot even imagine having no 3rd party jars in the classpath!

Static type check validates our software component against other components in the compilation environment. Does it match the runtime environment? What about different configurations of the runtime environment - there are tests and real deployments, and various types of deployments, and any given installation environment can change over time - new components being added, other updated or removed? How do we guarantee that compile time checks still hold? The short answer is - we can't. We need dynamic type safety anyway. Now let's examine the added value and the price (yes, there is one!) of deeply static types.

On code organization level, we try to reduce the dependencies to bare minimum - hide classes behind interfaces that we hope will remain more stable. The problem is that number of interfaces and factories in our application grows, while pursuing modularity we sacrifice simplicity... So maybe the problem is in the name? Some go as far as add support for structural types, minimizing dependency to a single field/method signature (not a problem-free solution, but there are interesting refinements). All this may help, but doesn't really solve the problem.

Another aspect we need to deal with is building and packaging the software. Here we enter the world of dependency management, the world of "make", Ant, Maven, repositories, jar versions; if it's a large enough and complex enough software we work on, simply speaking - we enter the world of pain. I still find it strange that we haven't found a better way.

As for application deployment and its problems, we'll get back to it later. But the truth is that no matter how hard we try, we can't guarantee there will be no errors when we deploy our software, so ... JVM doesn't trust us and gives us verification.

Simply put, when class is compiled, some of its requirements from other classes are captured and encoded into the bytecode. Then JVM would check them when the class is loaded, and reject the class if they can't be met. (This is really an over-simplified description of a complex algorithm, which also takes time to execute, despite optimization efforts on JVM side.) So this isn't really a dynamic check, it's something in between - names in our class get linked when it is loaded. In the classic Java SE class-loading scheme, where components are basically a chain, this scheme should work. But if we want real components, ones we can add, override, replace or remove while program is running - sweet turns sour. Our interfaces and factories have names, and classes that represent them need to reside in some "common vocabulary" usually loaded by the parent classloader, because it's not only the class bytecode that matters, but also who loaded what. Since we are talking actual classes, not their names, once we loaded two components, they cannot change their protocol of communication without reloading their parent, they also can't use a different version of a sub-component that the parent component has referenced.

In JEE that sort of things is necessary, that's why classloading in JEE is a terrible mess, not only it does not follow any specification, but it is different in almost each and every app server (wasn't there supposed to be portability?!) If you ever used commons-logging in a JEE app, you probably know what I mean. Maybe it got fixed lately, I don't know, but the Tech Guide for commons-logging is an ode to classloader frustration.

Back to deployment: whenever there are some sort of dynamic components - JEE, Spring or OSGi, there is always reflection. And most of the time there's lots of XML too. It's an escape route from static types. I attended Alef Arendsen's session at JavaEdge that presented OSGi and SpringSource. I carefully watched Alef juggle between XML, source and console like a child who watches a circus magician trying to uncover his tricks. But I didn't quite figure out the magic. And that was a whole session just for HelloWorld. I know Spring folks are doing best they can, and they're smart and all... but comparing to Smalltalk, I wasn't quite impressed. As for other solutions, although I haven't tried this out, there's Guice/OSGi integration without XML and with dynamic proxies and on-the-fly bytecode generation with ASM, but there's some overhead for the user, because it requires intermediate objects for services. So this way or the other, looks like JVM platform is holding us back.

Verification is addition, not replacement of dynamic checks. So what we get is basically a triple check of correctness (javac, verifier, dynamic) but loss of flexibility - we are interfering with components runtime life-cycles. If the invocation target is resolved just in time when the call is made, nothing precludes the target component from being reloaded between calls. But with preemptive validation, we get a static dependency tree at runtime, classes wired with each other "too early" and for good, which makes reloading a component very hard (although people keep trying). The reason for "early linking" is also performance, but late binding doesn't mean that the runtime platform can't do any optimization heuristics... but they'll have to be dynamic optimizations in the style of JIT. Will invokedynamic bring the salvation?

It seems when we are talking about multiple components, "statically typed platform" does not quite do the job. Static type check may mean a lot inside a component, but as for inter-component communication they are not only useless, but harmful. People sometimes dismiss dynamic types, because they think "it's like static types, but without static types". What they may not realize is that you are not just loosing, you are gaining something with dynamic types. You get late binding and meta-programming, and in a multi-component environment, it means a whole lot!

And that's when we are talking "inside the platform" components developed in the same language. Once you work with a system that runs on a different platform or developed in a different language - our type system doesn't normally stretch across the communication boundary. The other system may not even have static types, and since we are only as strong as the weakest link, our static types don't really help us. I think every time we try to encode types into communication between systems we end up with a monster like CORBA or Web Services. But there's another (unfortunately popular) extreme of just sending a string over and hoping for the best - with no checking on our side at all. Then we are relying on the other system to stop us from doing damage, and there's no way to correctly blame the component that made an error - was it a wrong string or an unexpected change on the other side? I think that ideally type or contract checking and conversions can be done dynamically on both sides, and not as part of the protocol. This results in light and flexible data-exchanging protocols (like HTTP or ATOM) which are easier to work with and I think will win in the end. On the more theoretical level I like this model for intercommunication and of course there are Aliens, that model external system as a special object in our system.

So as far as I see - components simply require a dynamic environment, they may be statically checked inside, but act as "dynamic" to the outside world. Sort of hard skeleton and soft shell. Indeed soft parts are much easier to fit together and less breakable, due to flexibility - this is used often in mechanical engineering and in nature, so why not in software?

How not to implement Comparable

I have already blogged about the danger of numeric overflows. I recently came across this example in Java course materials (!!!):

public class Foo implements Comparable<Foo> {
private int number;
public int compareTo(Foo o) {
  if (this == o) return 0;
  return number - o.number;
}
}

How do you think Integer.MAX_VALUE compares to negative numbers? It will appear smaller. This reminds me of even worse case we encountered in a real codebase. Look at this:

public class Foo implements Comparable<Foo> {
private long id;
//...
public int compareTo(Foo o) {
  if (this == o) return 0;
  return (int)(id - o.id);
}
public boolean equals(Object o) {
  return o != null && (o instanceOf Foo) && compareTo((Foo)o) == 0;
}
}

How do you think new Foo(8325671243L) and new Foo(25505540427L) compare? They are equal, but I will do it ala Weiqi Gao and leave you to find out why... :-)

Static initializers - update

After some interesting comments to my previous write-up I decided to make a follow-up post with some additional details.

Here is the code from Effective Java:

public class Person {
 private final Date birthDate;
 //...
 private static final Date BOOM_START;
 private static final Date BOOM_END;

 static {
   Calendar gmtCal = Calendar.getInstance(TimeZone.getTimeZone("GMT"));
   gmtCal.set(1946, Calendar.JANUARY, 1, 0, 0, 0);
   BOOM_START = gmtCal.getTime();
   gmtCal.set(1965, Calendar.JANUARY, 1, 0, 0, 0);
   BOOM_END = gmtCal.getTime();
 }

 public boolean isBabyBoomer() {
   return birthDate.compareTo(BOOM_START) >= 0 &&
     birthDate.compareTo(BOOM_END) < 0;
 }
}

and here's how I would have changed it:

public class BabyBoom {
 private static final BabyBoom boom = new BabyBoom();
 private Date start = null;
 private Date end = null;

 private BabyBoom() {}

 public static BabyBoom getInstance() {
   if (boom.start == null || boom.end == null) {
     Calendar gmtCal = Calendar.getInstance(TimeZone.getTimeZone("GMT"));
     gmtCal.set(1946, Calendar.JANUARY, 1, 0, 0, 0);
     boom.start = gmtCal.getTime();
     gmtCal.set(1965, Calendar.JANUARY, 1, 0, 0, 0);
     boom.end = gmtCal.getTime();
   }
   return boom;
 }

 public boolean contains(Date birthDate) {
   return birthDate.compareTo(start) >= 0 && birthDate.compareTo(end) < 0;
 }
}

public class Person {
 private final Date birthDate;
 //...
 public boolean isBabyBoomer() {
   BabyBoom boom = BabyBoom.getInstance();
   return boom.contains(birthDate);
 }
}

I didn't synchronize getInstance, because initializing the dates twice does no harm, so it's not worth the price of synchronization. However I did check that both fields are initialized before returning the object in getInstance. 

A case against static initializers

"Effective Java" is an excellent book. I recently bought the 2nd edition, and it is absolutely fabulous, priceless. However after quite some time in the industry, I've learnt not to take any advice blindly. First edition was also excellent, but several of the items were revisited since then. So here's an item that I have mixed feelings about - Avoid creating unnecessary objects. The advice is to use static initializers for the expensive computation. 

Static initializers are double-edged sword. It's like with the stock exchange in times of crisis - for a particular individual it may be a good idea to sell the stock, but the trouble is that everybody's doing it, and in the end everybody's losing big-time. Same applies to static initializers. One or two may seem harmless, but they add up and together create a big problem. The fact that static initializers are (edit) may be invoked at program start-up affects everybody and since they potentially interfere with classloading, it's very hard (edit) harder to debug them if anything goes wrong.

Here is how it usually gets out of hand: people start with initializing static members in static blocks. Map of values, sort of configuration details. That alone sounds harmless. But soon comes the time when the values in a map need to be read from a properties file, so here we got IO within static block. Uh oh, better catch these exceptions. Before you notice the whole thing turns into a puzzler. Don't believe me? Here's a real problem I had.

Server in Java, client is either applet or JNLP. On certain machines, only one of them can run. You run server first - client never comes up, no error whatsoever. You reboot and connect as client to another machine - no problem. But if you try to start up the server locally - silent death. A team in India spends months on it. In vain. The whole release is detained, escalation to senior management. The thing ends up on my desk after a ruthless blame game between teams. Long story short: it's a DirectX problem. Why the **** does the server need DirectX? Ah. What's next after reading defaults from a properties file? Reading them from the database. Oh, but it's a different process, and the database needs to be up. So we not just connect to database from initialization block, we wait for the database process to be up. Great idea. How? We follow "best coding practices" and reuse: find a poller utility somewhere in the JDK. Apparently there is a java.awt.Timer. Why not? Great idea. Apparently, a touch of one AWT class causes a bunch of other AWT classes to load, which in turn loads DirectX and OpenGL dlls. And guess what - Windows on some machines has a nasty bug, that only allows to load them once per machine, regardless of the user. And when another user tries to do it - the loading gets stuck. And our server is of course a system process, while the client belongs to the logged in user. 

Since it was a last minute fix, we solved it with some JVM flags that disabled DirectX and OpenGL. The problem was not the fix, but the diagnosis. If it was part of the regular code, it would have been easy to connect with a debugger, see what call gets stuck, investigate it from there. But as it was part of start-up, people didn't know where to look. Not to mention the man-months accumulatively spent by developers who waited for the server to restart when testing. 

So... what's the lesson here? Life is better without static, avoid it as much as you can. 

Java... tea?

No clone for you!

Java Practices call Cloneable/Object#clone API pathalogical and suggest to avoid it. I have to agree. Here's my little story. 

I have a method that receives a certain object, which is Cloneable. Now I want to actually clone that object, so I am calling the clone method. Makes sense, no? Apparently not. My slightly outdated version of IDEA was confused, just as I was, and did not complain. But Javac was cruel and uncompromizing. No clone for you. 

Because Cloneable does not require a clone() method, and Object#clone() can be protected. WTF? My objects subclass different not-necessarily-cloneable classes, so they can't have a common cloneable superclass. So I'll define my own ReallyCloneable interface that has a public clone method, sounds simple enough...? Not!

public interface ReallyCloneable extends Cloneable {
    public Object clone() _
}

Now I realize there's CloneNotSupportedException to throw. A checked exception. I have 2 options: be a good girl, adhere to best practices and clone the signature of Object#clone, or ... not. Let's explore both options. 

public interface ReallyCloneable extends Cloneable {
    public Object clone() throws CloneNotSupportedException;
}

Now all my classes (let's assume they are simple enough to contain only primitives, Strings and arrays of the above) implement ReallyCloneable and add those 3 lines:

public Object clone() throws CloneNotSupportedException {
    return super.clone();
}

But whoever clones my objects also needs to add:

try {
    //call clone
} catch (CloneNotSupportedException e)  {
    //handle it... how???
}

Because implementing Cloneable and all the rest means that JVM will not necessarity throw CloneNotSupportedException, but as far as Javac is concerned, it still may. So no matter what, the caller needs to prepare himself for the worst. Then why bother with interfaces and all that, I ask? People say Java is verbose. I don't mind verbose, but bloated with boiler-plate to this extent?! 

The other option is not to declare the exception in ReallyCloneable. This is possible because overriding rules do not require to throw the exceptions of the overridden method.

public Object clone() {
  try {
    return super.clone();
  } catch (CloneNotSupportedException e) {
    //hm...? 
  }
}

Even though JLS allows it, my IDE and static analysis tools will complain. So I have to either reconfigure them or add some "escape" annotations. More typing... By now I feel like a typomaniac. And wait, I should handle the exception - I don't want to swallow it. Besides, what if someone really objects to being cloned? There is a need for a runtime exception. 

public class CloneFailedException extends RuntimeException {
public CloneFailedException() {
super();
}
  public CloneFailedException(CloneNotSupportedException e) {
    super(e);
  }
}

and

public Object clone() {
  try {
    return super.clone();
  } catch (CloneNotSupportedException e) {
    throw new CloneFailedException(e); 
  }
}

Now at least the callers don't have to catch. But they will still have to cast, because there is no way to express a self type with Java's static types system. Sigh.

And all this why? and for what? Couldn't we just have Cloneable with a public clone and a RuntimeException to throw if we didn't want to be cloned? Simple as that? 

This is yet another example of static type system's failed attempt to protect us from ourselves. 

DSL - fuel for life

Do you feel that your programming language is too bloated? I do, and I know I am not alone. 

Let's take a look at Java. You may ask - what do you mean, when you say Java? Ah, good question. There is Java, the language, as described in the spec. Java the Standard. Don't you wonder where's the new edition of the book, BTW? Anyway, then there's mini-edition, enterprise edition, real-time Java... there's a huge stack of official/certified technologies (e.g. all the JEE stuff - is JSP or JSF Java? is JPQL?), for which there are often multiple vendors. That's not all, there are all the popular open-source frameworks that don't bother getting Sun's approval, and yet they possess lion-share of the market (e.g. Eclipse, Spring). There's no chance to even keep track of all this, forget mastering. And yet, I don't feel that I have all I need. I have all that, and yet I can't write a descent program the way I'd like to, because I am missing some core features. What? Let's see - how about proper modularity, closures, tuples, local type inference, properties... 

Java made the grade in expanding layer by layer - to the extent where it reminds me the state of the Earth in Wall-E. If you haven't seen the movie, I'll just say that the Earth drowned in human-produced garbage, and the garbage made it inhabitable for anything organic. The garbage in Java makes it impossible for the core, organic features of the language to grow. In the movie, people are leaving, and robots stay back to clean up - now you make the analogy. 

As a Java programmer I really identify with Wall-E. Moving tons of garbage around, day after day, in a desperate attempt to clean up the world - something obviously impossible. And man, I'd love to be that flying-iPod-looking Eve from outer space. She's so strong, so modern, so clean and shiny...  And she has a mission!

If Wall-E is my Java, then my Eve is no doubt Newspeak. 

So what does it mean in a wider sense of programming languages? I think good language should have a small core, as little redundancy as possible. And it should grow from then on. Here's Guy Steel at OOPSLA '98:

I think that the best way of growing is via internal domain specific languages, DSLs, - you don't change the core language, yet you cover more and more domains. Anders Hejlsberg also talked about it at the recent PDC conference - adding features as a DSL; and then, if the DSL is very successful and popular, add syntactic sugar to the core language, as they did with LINQ. I don't think that you need that latter part if your language is well suited for DSLs to begin with. So the language should have a small core and be well suited for DSLs. Java isn't DSL-friendly. Scala is much better, e.g. the Actors library. Haskell, Ruby and of course Smalltalk are really good at it. 

The last part is getting rid of garbage as you grow, and again, Gilad has an idea how to do that. 

I could go on forever and ever about DSLs, showing examples like parser combinators, unit-test and SQL libraries and so on. Martin Fowler is writing a book. So, instead of boring you with repeating what's been said many times, and enumerating lots of references, I will just finish with this cutest quote:

DSLs are for making languages bear-able.

For I am a bear of very little brain and long words confuse me. [Milne 1926]

The premise of this subject is that computers should adapt to the ways of people, and not the other way around.  

Types and other virtues

Here's a thought. Why do Java Generics attract so much bad vibes? If we're talking about types again, gotta visit our old friends - ML, Haskell. They are doing fine, they've got it all figured out. Hindley-Milner etc., they've proved themselves to be right.

But them and Java, it's apples and oranges. So to bring them to the same arena with object oriented languages, let's see how are they doing on the front of extensibility? Well, there are some recent advancements, polymorphic variants in OCaml, extensible records in Haskell, but it's not like the other stuff they've figured out ages ago, there's a slight hint of hesitation here.

So having types and being object oriented at the same time - maybe it's just, hm, non-trivial. If Martin Odersky says it's hard for the compiler (and he's a genius!) - then it really must be. Now look at it as a reverse Turing test - if the machine can't figure it out, how the **** are we supposed to? Which means we need an escape route. If declaring the right type is hard, we need an easy way out.

Pre-generics Java didn't bother itself too much with complex static types, the trick was to fall back on the dynamic types whenever in doubt: arrays - ArrayStoreException, collections - ClassCastException. Some say it's broken. It's a hole in a fence, but it's a way out . "Oops"

Then Generics came and fixed the hole in the fence. But, without realizing, they violated the status-quo, the eco-system, they stepped on a butterfly. Suddenly, the brain hurts, and the only way out is dumping all the angle brackets and jumping over the fence. It's a "No Exit"!

What does Scala do? Ah! First of all Scala has a much more advanced type system to start with. But the real trick is implicits. If it's too hard to declare the right type, just imperatively describe the conversion. Done deal, here's your way out. It's pretty cool, but implicits have their share of bad vibes too. Can't live with them, can't live without them, I guess.

So how about implicits for Java 7? No, actually how about Scala for Java 7? It's been done before, you know...

The Great Divide

What hasn't been said about static vs. dynamic types in programming languages? Read on at your own risk, because here I go again...

When you think of static types what comes to mind? Haskell? OCaml? Scala? Dear friend, you are better than most of us, but you have clicked the wrong URL. Peace. See you in another post...

Did you say Java? Still with me? Good. Listen, now when the others have gone, just between you and me, the guys from the previous paragraph - they're on to some good stuff. Check it out, you won't regret it. But don't quit your day job, not just yet. It's a bit complicated, but did you ever witness extreme programming methodology implemented in a big corporation? No? Then picture this: Elbonians take over Dilbert's firm and make everybody do XP. They even send pointy-haired boss to a Certified Scrum Master course. Get the outcome? It can only end like the implementation of Carl Marx's ideas in Russian countryside. I am trying to say - there are ideals, and there is reality. In reality, Haskell programs have bugs too.

So Java, you say. How do you feel about dynamic types? Cool? Get out of here. No really, it's no fun preaching to the converted. See you!

Oh no, heaven forbid, you won't touch them with a stick. You're my guy then. So let's rewind to Java 1.4 days, after all many Java developers still use 1.4 and many others look back at it with nostalgia. Are you one of them? Ok. So what about pre-generics collections, do you think they are statically typed? Hmmm... And what percentage of your code involves collections? So this code was not entirely statically checked. Now add all the reflection stuff...

But then of course came Generics. And suddenly Java is much more complex. How do I make my code compile, gee, wildcards, captures... ?! I am trying to get something done, hello!... It's easy of course to blame Generics implementation, but if we learn something from the folks whom I kindly asked to leave in the beginning, they'll tell you that finding correct static type for every element in your program is hard. They of course think that hard is good, they are noble men with ideals, they like overcoming challenges. But you and I, we're just trying to make a living. So we curse Sun and back off to an untyped collection. Hm, maybe we're just doing the right thing? Maybe sometimes we just know that our program is correct, but the compiler demands more and more typing and wastes our time?

Java 5 was all about improving type-checking. If pre-defined types were not enough, annotations came handy. Define your own and test it at compile-time or at run-time. Did it ever happen to you that there were so many annotations, that you couldn't see the code?

See, more types is not always a good thing. Unless you're very keen on intellectual challenges. James Gosling said this about Scala - functional programs will make your brain hurt, they are for calculus lovers. He's right. It's the kind of pain you feel in your muscles when you start working out, you know, that indicates they are still alive... So working out is good, but we can't afford doing it all day, ha?

Maybe the appeal of plain old Java was that it's a combination of static and dynamic checks? So it's not that all dynamic is evil, maybe it's a matter of how much and where?

Give dynamic types a break. Who knows, you may find eventually that they're good for some things.

Peace.

P.S. I've done some role playing here, just for the record. I do love Generics, even though they were hard to master. Annotations are overall very useful. Right now I don't do as much Java as I used to, and I do other fascinating languages (static and dynamic) as I, for long time, wanted to.

Beware of the subs

No, I don't mean these cute yellow things beneath the waves. I am going to talk about List#subList and String#substring methods in Java.

Apparently many people are unaware of what exactly these methods do. Unfortunately their ignorance may lead to unpleasant consequences. So getting straight to the point: both methods do not copy a portion of the original data, instead they create a view, or, in other words, a proxy to it. The important thing is that the new wrapper object holds a strong reference to the original object.

The Javadoc of subList at least admits that it's a view, as for substring, the only way to find out is by looking at the source - the method redirects to this constructor:

//Package private constructor which shares value array for speed.
String(int offset, int count, char value[]) {
  this.value = value;
  this.offset = offset;
  this.count = count;
}

So what's the problem? Let's look at the following snippet from a real code-base:

List leaky = ...; //long list of big & hairy objects
leaky = leaky.subList(from, to);

Assuming that leaky wasn't referenced anywhere else in the code, there is no way for the programmer to access the elements that lie beyond the (from,to) range. But these bytes aint going to rehab, no, no, no - as far as JVM is concerned they are still strongly referenced. So if you really mean to extract a portion of a list (or string), and throw the rest away - copy it manually to a new list (or string). For example:

List sneaky = ...;
sneaky = new ArrayList(sneaky.subList(from, to));

Is there any better way a "sub" could be implemented? Well, maybe the reference to the original data could be kept weak, and only when (if) the original object is enqueued for garbage collection then the data could be copied into the "view". This would require backwards references from "original" object to "views", which would also need to be weak, so... overall this doesn't seem worth the effort, and hence avoiding leaky lists and strings shall remain the responsibility of the programmer.

Speaking of subList, another nasty thing about it is that sub-list is not Serializable, nor Cloneable or anything like that, even if the original list was. And speaking of leaky things that are caused by undercover strong references - never forget non-static inner classes that refer to their enclosing instance.

Take care, and keep your head above the water :-)

JunkedIn

Indeed LinkedIn runs on Java, no doubt about it:

In whining there is truth?

What do Java developers want? Hard to say. But here is what Java bloggers don't want Sun to do:
No evolving Java syntax - no properties, no closures, etc. - fix what is already there first.
No extending Java capabilities via annotations.
No investing in a new JVM language (JavaFX, or JRuby in the past)
Ok, maybe add Groovy, but don't change JVM spec.
Hey, wait, why are key people leaving Sun? We didn't want that either!

Synthetic bridge method annotations

Update regarding "walls and bridges": Sun has kindly opened an enhancement request on my behalf for the issue. Please vote if you are affected by it too.

Java properties links

Since I am trying to keep up to date with what's going on in "Java Properties" land, I started with Alex Miller's collection and have been adding things that I find either relevant or influential. Another idea I nicked from Alex was to put them on Tumblog.

A disclaimer: this is a personal collection, so if something looks irrelevant, it means that I think it is influential :-)

I'll be more than glad to hear suggestions for more links or feeds.

Static methods as function objects

In one of the earlier posts I described a policy object created by "functional style programming" in Java. The basic idea is to make methods into Function and Predicate objects and then composite the logic from these building blocks. The problem though was that the policy implementation looked alien to Java. To re-cap: I wanted to invoke a different function based on the class of the object - classical multi-dispatch with a single argument.

So inspired by extension methods idea I added a new mechanism to define properties (and functions in general) - via static methods. You define a bunch of static methods with the same name (or appropriately annotated) like this:

static int size(Collection c) {
 return c.size();
}
static int size(Map map) {
 return map.size();
}
static int size(Object obj) {
 return (obj.getClass().isArray() ?
   Array.getLength(obj) : 1);
}

Now assuming I have a property declaration

Property<Object,Integer> size =
 new Property<Object,Integer>(0) {};

The underlying implementation of the size.of(object) would be a "policy" (composite function) dispatching to one of the static methods according to run-time type of the object. So that size.of("foo") is 1, size.of(Arrays.asList("a","b","c")) is 3 and size.of(null) is 0.

There are couple of issues though.

Issue #1: automatically ordering the rules. Now when rules are not added manually, I should be careful to test for more specific class first. The solution is to calculate a distance between two classes and always look for a best match - basically same algorithm javac uses when choosing methods at compile-time. This isn't fully unambiguous though: if class C implements 2 interfaces A and B, having different implementations of size(A) and size(B) would put me into a dilemma which one to choose for size(C). Javac in such a case reports an error - thanks to Java being statically typed. But once I make it a run-time decision the only logical solution is to report an error at run-time. To make the problem less acute, I could add an annotation that controls ordering of the methods.

Issue #2: this is less generic than a general predicate-based policy - it is limited to "instanceof" predicate. One of the policies I use internally in the properties project is reflectively invoke getFoo() if the class has a "getFoo" method. This isn't possible with static methods.

Issue #3: reflection. It's not really a problem, but rather an enhancement I want to make: instead of using reflection I could generate the functions at compile-time with APT, like Bruce Chapman does here. This is one of my next adventures, which I will hopefully describe in a future post.

The taste of Java Generics

As the closure debate flames are rising, people are attacking Generics in order to attack closures. Some use it as a negative example against particular closure proposal, some try to discourage Java language changes in general and some to promote other languages. When I read articles that start with "Generics are hard" and lead into "Generics are bad", I can't help thinking of the fox and the grapes. Yeah, they are hard. So is concurrency, NIO and many other things. So? That's why people create frameworks on top of Java.

Don't get me wrong - Java Generics have faults, and they are long known. Here's a hit-parade of my biggest complaints, but I will also mention some solutions I have used:
1. Erasure
The way to work around it is by using type token or super-type token. As any workaround, it is limited, but it can solve part of the problem - Class#cast and Class#newInstance don't issue the annoying warnings.
2. Generic parameters not covariant
It is usually a good idea to use bound wildcards extensively.
3. No lower bounds support (except wildcards)
We can use static methods instead (ala extensions) with <T extends Super> rather than <S super Sub>
4. Verbosity, instead of more aggressive type inference
Not much to do, hm, learn to type faster? :-)
5. The angle brackets
Yeah, I hate them most when typing on blogger, luckily I am old enough to have been coding HTML in plain text editor a long time ago, so... back to GT and LT escapes.

BTW Java didn't die, it has simply grown beyond its hipness age. And I am glad that it spawned even better things, such as Scala. But there is a difference between "good" and "popular". Java is popular and not that bad, other languages may be better, but not as popular. Will Scala be willing to sacrifice some of its expressive power and shortcuts (gee, stuff like that reminds me why I don't code in Perl) to become simpler to learn/read and potentially more popular? Should it? (Ruby consciously didn't and ... didn't.) We shall see.

Wildcards observation

"A bug in poker is a limited form of wild card."

- Wikipedia

While looking for academic references for my properties project, I bumped into the expression problem - the write-up by Phil Wadler and follow-ups by Mads Torgersen and Matthias Zenger/Martin Odersky are really fun and educative readings.

This stuff caused me to re-read the the wildcards paper and I suddenly understood why I had this strange gut feeling about wildcard usage. Almost all learning material about Generics, including the wildcard paper itself, emphasizes the distinction between read and write methods - having a wildcard in a collection parameter prevents you from adding elements to the collection, but allows reading. This is true, but somewhat misleading. Actually, wildcard in the parameter allows us to call a method that returns the thing behind the wildcard and treat it as if the returned object is of the type of wildcard bound (or an Object, if the wildcard is unbound); the wildcard prevents us, however, from calling any method that takes "the thing behind the wildcard" as a parameter. Now obviously, to add something to a collection you need to call a method that takes that something as parameter, and it is indeed prevented, but other modifications are not disallowed, e.g. clear() .

The consequence of wildcard usage applies to any method, not just "write method", for example - containment test. So looking at Kevin's example, it now seems to me that doSomeReading method shouldn't have used a wildcard if it wished to perform a safe containment test - that's the catch (or should I say, hm, capture...)

Happy New Year!