Interfaces vs Public Class Members - language-agnostic

I've noticed that some programmers like to make interfaces for just about all their classes. I like interfaces for certain things (such as checking if an object supports a certain behavior and then having an interface for that behavior) but overuse of interfaces can sometimes bloat the code. When I declare methods or properties as public I'd expect people to just use my concrete classes and I don't really understand the need to create interfaces on top of that.
I'd like to hear your take on interfaces. When do you use them and for what purposes?
Thank you.

Applying any kind of design pattern or idea without thinking, just because somebody told you it's good practice, is a bad idea.
That ofcourse includes creating a separate interface for each and every class you create. You should at least be able to give a good reason for every design decision, and "because Joe says it's good practice" is not a good enough reason.
Interfaces are good for decoupling the interface of some unit of code from its implementation. A reason to create an interface is because you foresee that there might be multiple implementations of it in the future. It can also help with unit testing; you can make a mock implementation of the services that the unit you want to test depends on, and plug the mock implementations in instead of "the real thing" for testing.

Interfaces are a powerful tool for abstraction. With them, you can more freely substitute (for example) test classes and thereby decouple your code. They are also a way to narrow the scope of your code; you probably don't need the full feature set of a given class in a particular place - exactly what features do you need? That's a client-focused way of thinking about interfaces.

Unit tests.
With an interface describing all class methods and properties it is within the reach of a click to create a mock-up class to simulate behavior that is not within the scope of said test.

It's all about expecting and preparing for change.
One approach that some use (and I'm not necessarily advocating it)
is to create an IThing and a ThingFactory.
All code will reference IThing (instead of ConcreteThing).
All object creation can be done via the Factory Method.
ThingFactory.CreateThing(some params).
So, today we only have AmericanConcreteThing. And the possibility is that we may never need another. However, if experience has taught me anything, it is that we will ALWAYS need another.
You may not need EuropeanThing, but TexasAmericanThing is a distinct possibility.
So, In order to minimize the impact on my code, I can change the creational line to:
ThingFactory.CreateThing( Account )
and Create my class TexasAmericanThing : IThing.
Other than building the class, the only change is to the ThingFactory, which will require a change from
public static IThing CreateThing(Account a)
{
return new AmericanThing();
}
to
public static IThing CreateThing(Account a)
{
if (a.State == State.TEXAS) return new TexasAmericanThing();
return new AmericanThing();
}

I've seen plenty of mindless Interfaces myself. However, when used intelligently, they can save the day. You should use Interfaces for decoupling two components or two layers of an application. This can enable you to easily plug-in varying implementations of the interface without affecting the client, or simply insulate the client from constant changes to the implementation, as long as you stay true to the contract of the interface. This can make the code more maintainable in the long term and can save the effort of refactoring later.
However, overly aggressive decoupling can make for non-intuitive code. It's overuse can lead to nuisance. You should carefully identify the cohesive parts of your application and the boundaries between them and use interfaces there. Another benefit of using Interfaces between such parts is that they can be developed in parallel and tested independently using mock implementations of the interfaces they use.
OTOH, having client code access public member methods directly is perfectly okay if you really don't foresee any changes to the class that might also necessitate changes in the client. In any case, however, having public member fields I think is not good. This is extremely tight coupling! You are basically exposing the architecture of your class and making the client code dependent on it. Tomorrow if you realize that another data structure for a particular field will perform better, you can't change it without also changing the client code.

I primarily use interfaces for IoC to enable unit testing.

On the one hand, this could be interpreted as premature generalization. On the other hand, using interfaces as a rule helps you write code that is more easily composable and hence testable. I think the latter wins out in many cases.

I like interfaces:
* to define a contract between parts/modules/subsystems or 3rd party systems
* when there are exchangeable states or algorithms (state/strategy)

Related

Application design - When should interfaces be used?

I kind of understand an interface as being a contract that can be applied to classes that would otherwise have nothing in common (ex: Comparable in Java). However, in what situation(s) would you have the reflex of adding an interface at the design stage?
Whenever you are using a statically typed language, and you want to make it possible for the developer to use your code while providing an alternate implementation - in other words, in such language it is necessary to achieve low(er) coupling.
Languages that use ducktyping as a rule, rather than strict type checking, for example, python, would generally have no need for interfaces.
"I kind of understand an interface as being a contract that can be applied to classes that would otherwise have nothing in common" - that's probably not the way to think about what an Interface is.
An Interface describes behaviour, and implementing an interface means a class enters into a contract to deliver that behavior.
By programming to an interface, rather than an implementation, you enable polymorphism and get more flexible code with lower coupling. For example, this method can take any instance that implements IQuack:
public void DoSomething(IQuack quacker)
{
// ...
}
If you are designing a product and you know the product is going to interact with a type of device, service etc. but not necessarily which, you can use an interface to move forward with the overall architecture, PROVIDED that you know enough about those types of devices to write an interface that can be successfully used by any given device of that type. Of course if you are in the design phase, you better have that knowledge. It's not uncommon to do high level designs using only interface declarations. I'm not saying it's good or bad, but it seems to be a pretty common practice of those who use software (like Rose etc) to generate a skeleton from UML.
Another time would be if you know exactly what device you are going to use but you think there might be a chance that you will need to work with different or multiple types of that device down the road.
A third usage of interfaces is to reduce duplicated code. This is probably the only place people ever get carried away with interface usage and if it wasnt for that, I'd be comfortable saying dont ask "Should this be an interface?" but "Can this be an interface?".

Should we avoid to use Object as the input parameter/ output value of a method?

Take Java syntax as an example, though the question itself is language independent. If the following snippet takes an object MyAbstractEmailTemplate as input argument in the method setTemplate, the class MyGateway will then become tightly-coupled with the object MyAbstractEmailTemplate, which lessens the re-usability of the class MyGateway.
A compromise is to use dependency-injection to ease the instantiation of MyAbstractEmailTemplate. This might solve the coupling problem
to some extent, but the interface is still rigid, hardly providing enough flexibility to
other developers/ applications.
So if we only use primitive data type (or even plain XML in web service) as the input/ output of a method, it seems the coupling problem no longer exists. So what do you think?
public class MyGateway {
protected MyAbstractEmailTemplate template;
public void setTemplate(MyAbstractEmailTemplate template) {
this.template = template;
}
}
It's pretty difficult to understand what you are really asking, but going the route of typing everything to Object does not lead to loose coupling because you can't do anything with the input without downcasting, which would break the Liskov Substituion Principle.
Taken to the extreme it leads you here:
public class MyClass
{
public object Invoke(object obj);
}
This is not loose coupling, it's just obscure and hard-to-maintain code.
The name MyAbstractEmailTemplate makes me believe that you are talking about an abstract class.
You should always program against interfaces, so instead of having MyGateway depend on MyAbstractEmailTemplate, it should depend on an EmailTemplate interface, where MyAbstractEmailTemplate implements EmailTemplate. Then, you can pass your custom implementations around as you want to, without further tight coupling.
Combine this with DI and you've got yourself a pretty decent solution.
Not exactly sure what you mean with "the interface is still rigid", but obviously you should design your interface in such a way that it provides the functionality you need.
MyGateway has to assume something about the inputs. Even if it used XML, it would have to assume something about the structure and content of the XML. Coupling isn't an evil in its own right; expresses the contract between two pieces of code. The oft-repeated advice to avoid tight coupling is really just saying that coupling should express the essence of a contract, not more and not less. Passing a specific type (particularly an interface type) is a very good way to achieve this balance.
The first problem you will run into is that a lot of types are simply not representable by a primitive data type (It's a Java problem that there are primitive types at all.).
The coupling should be reduced by using a proper inheritance hierarchy. What means proper? The method should take exactly that part of the interface as a parameter that is need. Not more not less.
After all you won't be able to avoid dependencies. Methods have to know about what they can do with their input or have to able to make assumptions (see C++ concepts) about the capabilities of the input.
IMHO there is nothing inherently wrong in using objects (wth small cap, not Objects) as method parameters and/or class members. Yes, these create dependencies. You can manage this in (at least) two ways:
acknowledge that by creating this dependency, the two classes become tightly coupled. This is entirely appropriate in many cases, where two (or more) classes in fact form a component, which is a meaningful unit of reuse in itself, and its parts may not make much sense or be interchangeable.
if there are multiple interchangeable candidates for a method parameter, these are obvious candidates to form a class hierarchy. Then you program for the interface and can pass any object of any class implementing that interface as parameter to your method. Note that the phrase "there are multiple interchangeable candidates for a method parameter" is a loose rephrasing of the Liskov Substitution Principle, which is the foundation of polymorphism.
in some languages, e.g. C++, the third way would be using templates. Then you need no common interface, only specific methods/members need to resolvable when the template is instantiated. However, since instantiation happens at compile time, this is entirely static binding.
sThe problem is I would say, that the best java can offer are interfaces and people start to see that they are too rigid. It would be interesting to use something like what is in Go language, that allows flexible checking for all methods of an interface to be present in the type, you do not have to be explicit about implementing some interface. We also need something better than interfaces to specify the constraints - maybe some sort of contracts. Another thing is the interface evolution.

Developing to an interface with TDD

I'm a big fan of TDD and use it for the vast majority of my development these days. One situation I run into somewhat frequently, though, and have never found what I thought was a "good" answer for, is something like the following (contrived) example.
Suppose I have an interface, like this (writing in Java, but really, this applies to any OO language):
public interface PathFinder {
GraphNode[] getShortestPath(GraphNode start, GraphNode goal);
int getShortestPathLength(GraphNode start, GraphNode goal);
}
Now, suppose I want to create three implementations of this interface. Let's call them DijkstraPathFinder, DepthFirstPathFinder, and AStarPathFinder.
The question is, how do I develop these three implementations using TDD? Their public interface is going to be the same, and, presumably, I would write the same tests for each, since the results of getShortestPath() and getShortestPathLength() should be consistent among all three implementations.
My choices seem to be:
Write one set of tests against PathFinder as I code the first implementation. Then write the other two implementations "blind" and make sure they pass the PathFinder tests. This doesn't seem right because I'm not using TDD to develop the second two implementation classes.
Develop each implementation class in a test-first manner. This doesn't seem right because I would be writing the same tests for each class.
Combine the two techniques above; now I have a set of tests against the interface and a set of tests against each implementation class, which is nice, but the tests are all the same, which isn't nice.
This seems like a fairly common situation, especially when implementing a Strategy pattern, and of course the differences between implementations might be more than just time complexity. How do others handle this situation? Is there a pattern for test-first development against an interface that I'm not aware of?
You write interface tests to exercise the interface, and you write more detailed tests for the actual implementations. Interface-based design talks a bit about the fact that your unit tests should form a kind of "contract" specification for that interface. Maybe when Spec# comes out, there'll be a langugage supported way to do this.
In this particular case, which is a strict strategy implementation, the interface tests are enough. In other cases, where an interface is a subset of the implementation's functionality, you would have tests for both the interface and the implementation. Think of a class which implements 3 interfaces, for example.
EDIT: This is useful so that when you add another implementation of the interface down the road, you already have tests for verifying that the class implements the contract of the interface correctly. This can work for something as specific as ISortingStrategy to something as wide-ranging as IDisposable.
there is nothing wrong with writing tests against the interface, and reusing them for each implementation, for example -
public class TestPathFinder : TestClass
{
public IPathFinder _pathFinder;
public IGraphNode _startNode;
public IGraphNode _goalNode;
public TestPathFinder() : this(null,null,null) { }
public TestPathFinder(IPathFinder ipf,
IGraphNode start, IGraphNode goal) : base()
{
_pathFinder = ipf;
_startNode = start;
_goalNode = goal;
}
}
TestPathFinder tpfDijkstra = new TestPathFinder(
new DijkstraPathFinder(), n1, nN);
tpfDijkstra.RunTests();
//etc. - factory optional
I would argue that this is the least effort solution, which is very much in line with Agile/TDD principles.
I would have no problem going with option 1, and keep in mind that refactoring is part of TDD and it's usually during a refactoring phase that you move to a design pattern such as strategy, so I wouldn't feel bad about doing that w/o writing new tests.
If you wanted to test the implementation-specific details of each PathFinder impl, you might consider passing mock GraphNodes which are somehow capable of helping to assert the Dijkstra-ness or DepthFirst-ness, etc, of the implementation. (Perhaps these mock GraphNodes could record how they are traversed, or somehow measure performance.) Maybe this is testing overkill, but then again if you know your system needs these three distinct strategies for some reason, it'd probably be good to have tests to demonstrate why - otherwise why not just pick one implementation and throw the others away?
I don't mind reusing test code as a template for new tests that have similar functionality. Depending on the particular class under test, you may have to rework them with different mock objects and expectations. At the least you'll have to refactor them to use the new implementation. I would follow the TDD method, though, of taking one test, reworking it for the new class, then writing just the code to pass that test. This may take even more discipline, though, since you already have one implementation under your belt and will undoubtedly be influenced by code you have already written.
This doesn't seem right because I'm
not using TDD to develop the second
two implementation classes.
Sure you are.
Start by commenting out all the tests but one. As you make a test pass either refactor or uncomment another test.
Jtf

Private vs. Public members in practice (how important is encapsulation?) [closed]

Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 5 years ago.
Improve this question
One of the biggest advantages of object-oriented programming is encapsulation, and one of the "truths" we've (or, at least, I've) been taught is that members should always be made private and made available via accessor and mutator methods, thus ensuring the ability to verify and validate the changes.
I'm curious, though, how important this really is in practice. In particular, if you've got a more complicated member (such as a collection), it can be very tempting to just make it public rather than make a bunch of methods to get the collection's keys, add/remove items from the collection, etc.
Do you follow the rule in general? Does your answer change depending on whether it's code written for yourself vs. to be used by others? Are there more subtle reasons I'm missing for this obfuscation?
It depends. This is one of those issues that must be decided pragmatically.
Suppose I had a class for representing a point. I could have getters and setters for the X and Y coordinates, or I could just make them both public and allow free read/write access to the data. In my opinion, this is OK because the class is acting like a glorified struct - a data collection with maybe some useful functions attached.
However, there are plenty of circumstances where you do not want to provide full access to your internal data and rely on the methods provided by the class to interact with the object. An example would be an HTTP request and response. In this case it's a bad idea to allow anybody to send anything over the wire - it must be processed and formatted by the class methods. In this case, the class is conceived of as an actual object and not a simple data store.
It really comes down to whether or not verbs (methods) drive the structure or if the data does.
As someone having to maintain several-year-old code worked on by many people in the past, it's very clear to me that if a member attribute is made public, it is eventually abused. I've even heard people disagreeing with the idea of accessors and mutators, as that's still not really living up to the purpose of encapsulation, which is "hiding the inner workings of a class". It's obviously a controversial topic, but my opinion would be "make every member variable private, think primarily about what the class has got to do (methods) rather than how you're going to let people change internal variables".
Yes, encapsulation matters. Exposing the underlying implementation does (at least) two things wrong:
Mixes up responsibilities. Callers shouldn't need or want to understand the underlying implementation. They should just want the class to do its job. By exposing the underlying implementation, you're class isn't doing its job. Instead, it's just pushing the responsibility onto the caller.
Ties you to the underlying implementation. Once you expose the underlying implementation, you're tied to it. If you tell callers, e.g., there's a collection underneath, you cannot easily swap the collection for a new implementation.
These (and other) problems apply regardless of whether you give direct access to the underlying implementation or just duplicate all the underlying methods. You should be exposing the necessary implementation, and nothing more. Keeping the implementation private makes the overall system more maintainable.
I prefer to keep members private as long as possible and only access em via getters, even from within the very same class. I also try to avoid setters as a first draft to promote value style objects as long as it is possible. Working with dependency injection a lot you often have setters but no getters, as clients should be able to configure the object but (others) not get to know what's acutally configured as this is an implementation detail.
Regards,
Ollie
I tend to follow the rule pretty strictly, even when it's just my own code. I really like Properties in C# for that reason. It makes it really easy to control what values it's given, but you can still use them as variables. Or make the set private and the get public, etc.
Basically, information hiding is about code clarity. It's designed to make it easier for someone else to extend your code, and prevent them from accidentally creating bugs when they work with the internal data of your classes. It's based on the principle that nobody ever reads comments, especially ones with instructions in them.
Example: I'm writing code that updates a variable, and I need to make absolutely sure that the Gui changes to reflect the change, the easiest way is to add an accessor method (aka a "Setter"), which is called instead of updating data is updated.
If I make that data public, and something changes the variable without going through the Setter method (and this happens every swear-word time), then someone will need to spend an hour debugging to find out why the updates aren't being displayed. The same applies, to a lesser extent, to "Getting" data. I could put a comment in the header file, but odds are that no-one will read it till something goes terribly, terribly wrong. Enforcing it with private means that the mistake can't be made, because it'll show up as an easily located compile-time bug, rather than a run-time bug.
From experience, the only times you'd want to make a member variable public, and leave out Getter and Setter methods, is if you want to make it absolutely clear that changing it will have no side effects; especially if the data structure is simple, like a class that simply holds two variables as a pair.
This should be a fairly rare occurence, as normally you'd want side effects, and if the data structure you're creating is so simple that you don't (e.g a pairing), there will already be a more efficiently written one available in a Standard Library.
With that said, for most small programs that are one-use no-extension, like the ones you get at university, it's more "good practice" than anything, because you'll remember over the course of writing them, and then you'll hand them in and never touch the code again. Also, if you're writing a data structure as a way of finding out about how they store data rather than as release code, then there's a good argument that Getters and Setters will not help, and will get in the way of the learning experience.
It's only when you get to the workplace or a large project, where the probability is that your code will be called to by objects and structures written by different people, that it becomes vital to make these "reminders" strong. Whether or not it's a single man project is surprisingly irrelevant, for the simple reason that "you six weeks from now" is as different person as a co-worker. And "me six weeks ago" often turns out to be lazy.
A final point is that some people are pretty zealous about information hiding, and will get annoyed if your data is unnecessarily public. It's best to humour them.
C# Properties 'simulate' public fields. Looks pretty cool and the syntax really speeds up creating those get/set methods
Keep in mind the semantics of invoking methods on an object. A method invocation is a very high level abstraction that can be implemented my the compiler or the run time system in a variety of different ways.
If the object who's method you are invoking exists in the same process/ memory map then a method could well be optimized by a compiler or VM to directly access the data member. On the other hand if the object lives on another node in a distributed system then there is no way that you can directly access it's internal data members, but you can still invoke its methods my sending it a message.
By coding to interfaces you can write code that doesn't care where the target object exists or how it's methods are invoked or even if it's written in the same language.
In your example of an object that implements all the methods of a collection, then surely that object actually is a collection. so maybe this would be a case where inheritance would be better than encapsulation.
It's all about controlling what people can do with what you give them. The more controlling you are the more assumptions you can make.
Also, theorectically you can change the underlying implementation or something, but since for the most part it's:
private Foo foo;
public Foo getFoo() {}
public void setFoo(Foo foo) {}
It's a little hard to justify.
Encapsulation is important when at least one of these holds:
Anyone but you is going to use your class (or they'll break your invariants because they don't read the documentation).
Anyone who doesn't read the documentation is going to use your class (or they'll break your carefully documented invariants). Note that this category includes you-two-years-from-now.
At some point in the future someone is going to inherit from your class (because maybe an extra action needs to be taken when the value of a field changes, so there has to be a setter).
If it is just for me, and used in few places, and I'm not going to inherit from it, and changing fields will not invalidate any invariants that the class assumes, only then I will occasionally make a field public.
My tendency is to try to make everything private if possible. This keeps object boundaries as clearly defined as possible and keeps the objects as decoupled as possible. I like this because when I have to rewrite an object that I botched the first (second, fifth?) time, it keeps the damage contained to a smaller number of objects.
If you couple the objects tightly enough, it may be more straightforward just to combine them into one object. If you relax the coupling constraints enough you're back to structured programming.
It may be that if you find that a bunch of your objects are just accessor functions, you should rethink your object divisions. If you're not doing any actions on that data it may belong as a part of another object.
Of course, if you're writing a something like a library you want as clear and sharp of an interface as possible so others can program against it.
Fit the tool to the job... recently I saw some code like this in my current codebase:
private static class SomeSmallDataStructure {
public int someField;
public String someOtherField;
}
And then this class was used internally for easily passing around multiple data values. It doesn't always make sense, but if you have just DATA, with no methods, and you aren't exposing it to clients, I find it a quite useful pattern.
The most recent use I had of this was a JSP page where I had a table of data being displayed, defined at the top declaratively. So, initially it was in multiple arrays, one array per data field... this ended in the code being rather difficult to wade through with fields not being next to eachother in definition that would be displayed together... so I created a simple class like above which would pull it together... the result was REALLY readable code, a lot more so than before.
Moral... sometimes you should consider "accepted bad" alternatives if they may make the code simpler and easier to read, as long as you think it through and consider the consequences... don't blindly accept EVERYTHING you hear.
That said... public getters and setters is pretty much equivalent to public fields... at least essentially (there is a tad more flexibility, but it is still a bad pattern to apply to EVERY field you have).
Even the java standard libraries has some cases of public fields.
When I make objects meaningful they are easier to use and easier to maintain.
For example: Person.Hand.Grab(howquick, howmuch);
The trick is not to think of members as simple values but objects in themselves.
I would argue that this question does mix-up the concept of encapsulation with 'information hiding'
(this is not a critic, since it does seem to match a common interpretation of the notion of 'encapsulation')
However for me, 'encapsulation' is either:
the process of regrouping several items into a container
the container itself regrouping the items
Suppose you are designing a tax payer system. For each tax payer, you could encapsulate the notion of child into
a list of children representing the children
a map of to takes into account children from different parents
an object Children (not Child) which would provide the needed information (like total number of children)
Here you have three different kinds of encapsulations, 2 represented by low-level container (list or map), one represented by an object.
By making those decisions, you do not
make that encapsulation public or protected or private: that choice of 'information hiding' is still to be made
make a complete abstraction (you need to refine the attributes of object Children and you may decide to create an object Child, which would keep only the relevant informations from the point of view of a tax payer system)
Abstraction is the process of choosing which attributes of the object are relevant to your system, and which must be completely ignored.
So my point is:
That question may been titled:
Private vs. Public members in practice (how important is information hiding?)
Just my 2 cents, though. I perfectly respect that one may consider encapsulation as a process including 'information hiding' decision.
However, I always try to differentiate 'abstraction' - 'encapsulation' - 'information hiding or visibility'.
#VonC
You might find the International Organisation for Standardization's, "Reference Model of Open Distributed Processing," an interesting read. It defines: "Encapsulation: the property that the information contained in an object is accessible only through interactions at the interfaces supported by the object."
I tried to make a case for information hiding's being a critical part of this definition here:
http://www.edmundkirwan.com/encap/s2.html
Regards,
Ed.
I find lots of getters and setters to be a code smell that the structure of the program is not designed well. You should look at the code that uses those getters and setters, and look for functionality that really should be part of the class. In most cases, the fields of a class should be private implementation details and only the methods of that class may manipulate them.
Having both getters and setters is equal to the field being public (when the getters and setters are trivial/generated automatically). Sometimes it might be better to just declare the fields public, so that the code will be more simple, unless you need polymorphism or a framework requires get/set methods (and you can't change the framework).
But there are also cases where having getters and setters is a good pattern. One example:
When I create the GUI of an application, I try to keep the behaviour of the GUI in one class (FooModel) so that it can be unit tested easily, and have the visualization of the GUI in another class (FooView) which can be tested only manually. The view and model are joined with simple glue code; when the user changes the value of field x, the view calls setX(String) on the model, which in turn may raise an event that some other part of the model has changed, and the view will get the updated values from the model with getters.
In one project, there is a GUI model which has 15 getters and setters, of which only 3 get methods are trivial (such that the IDE could generate them). All the others contain some functionality or non-trivial expressions, such as the following:
public boolean isEmployeeStatusEnabled() {
return pinCodeValidation.equals(PinCodeValidation.VALID);
}
public EmployeeStatus getEmployeeStatus() {
Employee employee;
if (isEmployeeStatusEnabled()
&& (employee = getSelectedEmployee()) != null) {
return employee.getStatus();
}
return null;
}
public void setEmployeeStatus(EmployeeStatus status) {
getSelectedEmployee().changeStatusTo(status, getPinCode());
fireComponentStateChanged();
}
In practice I always follow only one rule, the "no size fits all" rule.
Encapsulation and its importance is a product of your project. What object will be accessing your interface, how will they be using it, will it matter if they have unneeded access rights to members? those questions and the likes of them you need to ask yourself when working on each project implementation.
I base my decision on the Code's depth within a module.
If I'm writting code that is internal to a module, and does not interface with the outside world I don't encapsulate things with private as much because it affects my programmer performance (how fast I can write and rewrite my code).
But for the objects that server as the module's interface with user code, then I adhere to strict privacy patterns.
Certainly it makes a difference whether your writing internal code or code to be used by someone else (or even by yourself, but as a contained unit.) Any code that is going to be used externally should have a well defined/documented interface that you'll want to change as little as possible.
For internal code, depending on the difficulty, you may find it's less work to do things the simple way now, and pay a little penalty later. Of course Murphy's law will ensure that the short term gain will be erased many times over in having to make wide-ranging changes later on where you needed to change a class' internals that you failed to encapsulate.
Specifically to your example of using a collection that you would return, it seems possible that the implementation of such a collection might change (unlike simpler member variables) making the utility of encapsulation higher.
That being said, I kinda like Python's way of dealing with it. Member variables are public by default. If you want to hide them or add validation there are techniques provided, but those are considered the special cases.
I follow the rules on this almost all the time. There are four scenarios for me - basically, the rule itself and several exceptions (all Java-influenced):
Usable by anything outside of the current class, accessed via getters/setters
Internal-to-class usage typically preceded by 'this' to make it clear that it's not a method parameter
Something meant to stay extremely small, like a transport object - basically a straight shot of attributes; all public
Needed to be non-private for extension of some sort
There's a practical concern here that isn't being addressed by most of the existing answers. Encapsulation and the exposure of clean, safe interfaces to outside code is always great, but it's much more important when the code you're writing is intended to be consumed by a spatially- and/or temporally-large "user" base. What I mean is that if you plan on somebody (even you) maintaining the code well into the future, or if you're writing a module that will interface with code from more than a handful of other developers, you need to think much more carefully than if you're writing code that's either one-off or wholly written by you.
Honestly, I know what wretched software engineering practice this is, but I'll oftentimes make everything public at first, which makes things marginally faster to remember and type, then add encapsulation as it makes sense. Refactoring tools in most popular IDEs these days makes which approach you use (adding encapsulation vs. taking it away) much less relevant than it used to be.

Singletons: good design or a crutch? [closed]

Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 8 years ago.
Improve this question
Singletons are a hotly debated design pattern, so I am interested in what the Stack Overflow community thought about them.
Please provide reasons for your opinions, not just "Singletons are for lazy programmers!"
Here is a fairly good article on the issue, although it is against the use of Singletons:
scientificninja.com: performant-singletons.
Does anyone have any other good articles on them? Maybe in support of Singletons?
In defense of singletons:
They are not as bad as globals because globals have no standard-enforced initialization order, and you could easily see nondeterministic bugs due to naive or unexpected dependency orders. Singletons (assuming they're allocated on the heap) are created after all globals, and in a very predictable place in the code.
They're very useful for resource-lazy / -caching systems such as an interface to a slow I/O device. If you intelligently build a singleton interface to a slow device, and no one ever calls it, you won't waste any time. If another piece of code calls it from multiple places, your singleton can optimize caching for both simultaneously, and avoid any double look-ups. You can also easily avoid any deadlock condition on the singleton-controlled resource.
Against singletons:
In C++, there's no nice way to auto-clean-up after singletons. There are work-arounds, and slightly hacky ways to do it, but there's just no simple, universal way to make sure your singleton's destructor is always called. This isn't so terrible memory-wise -- just think of it as more global variables, for this purpose. But it can be bad if your singleton allocates other resources (e.g. locks some files) and doesn't release them.
My own opinion:
I use singletons, but avoid them if there's a reasonable alternative. This has worked well for me so far, and I have found them to be testable, although slightly more work to test.
Google has a Singleton Detector for Java that I believe started out as a tool that must be run on all code produced at Google. The nutshell reason to remove Singletons:
because they can make testing
difficult and hide problems with your
design
For a more explicit explanation see 'Why Singletons Are Controversial' from Google.
A singleton is just a bunch of global variables in a fancy dress.
Global variables have their uses, as do singletons, but if you think you're doing something cool and useful with a singleton instead of using a yucky global variable (everyone knows globals are bad mmkay), you're unfortunately misled.
The purpose of a Singleton is to ensure a class has only one instance, and provide a global point of access to it. Most of the time the focus is on the single instance point. Imagine if it were called a Globalton. It would sound less appealing as this emphasizes the (usually) negative connotations of a global variable.
Most of the good arguments against singletons have to do with the difficulty they present in testing as creating test doubles for them is not easy.
There's three pretty good blog posts about Singletons by Miško Hevery in the Google Testing blog.
Singletons are Pathological Liars
Where Have All the Singletons Gone?
Root Cause of Singletons
Singleton is not a horrible pattern, although it is misused a lot. I think this misuse is because it is one of the easier patterns and most new to the singleton are attracted to the global side effect.
Erich Gamma had said the singleton is a pattern he wishes wasn't included in the GOF book and it's a bad design. I tend to disagree.
If the pattern is used in order to create a single instance of an object at any given time then the pattern is being used correctly. If the singleton is used in order to give a global effect, it is being used incorrectly.
Disadvantages:
You are coupling to one class throughout the code that calls the singleton
Creates a hassle with unit testing because it is difficult to replace the instance with a mock object
If the code needs to be refactored later on because of the need for more than one instance, it is more painful than if the singleton class were passed into the object (using an interface) that uses it
Advantages:
One instance of a class is represented at any given point in time.
By design you are enforcing this
Instance is created when it is needed
Global access is a side effect
Chicks dig me because I rarely use singleton and when I do it's typically something unusual. No, seriously, I love the singleton pattern. You know why? Because:
I'm lazy.
Nothing can go wrong.
Sure, the "experts" will throw around a bunch of talk about "unit testing" and "dependency injection" but that's all a load of dingo's kidneys. You say the singleton is hard to unit test? No problem! Just declare everything public and turn your class into a fun house of global goodness. You remember the show Highlander from the 1990's? The singleton is kind of like that because: A. It can never die; and B. There can be only one. So stop listening to all those DI weenies and implement your singleton with abandon. Here are some more good reasons...
Everybody is doing it.
The singleton pattern makes you invincible.
Singleton rhymes with "win" (or "fun" depending on your accent).
I think there is a great misunderstanding about the use of the Singleton pattern. Most of the comments here refer to it as a place to access global data. We need to be careful here - Singleton as a pattern is not for accessing globals.
Singleton should be used to have only one instance of the given class. Pattern Repository has great information on Singleton.
One of the colleagues I have worked with was very Singleton-minded. Whenever there was something that was kind of a manager or boss like object he would make that into a singleton, because he figured that there should be only one boss. And each time the system took up some new requirements, it turned out there were perfectly valid reasons to allow multiple instances.
I would say that singleton should be used if the domain model dictates (not 'suggests') that there is one. All other cases are just accendentally single instances of a class.
I've been trying to think of a way to come to the poor singelton's rescue here, but I must admit it's hard. I've seen very few legitimate uses of them and with the current drive to do dependency injection andd unit testing they are just hard to use. They definetly are the "cargo cult" manifestation of programming with design patterns I have worked with many programmers that have never cracked the "GoF" book but they know 'Singelton' and thus they know 'Patterns'.
I do have to disagree with Orion though, most of the time I've seen singeltons oversused it's not global variables in a dress, but more like global services(methods) in a dress. It's interesting to note that if you try to use Singeltons in the SQL Server 2005 in safe mode through the CLR interface the system will flag the code. The problem is that you have persistent data beyond any given transaction that may run, of course if you make the instance variable read only you can get around the issue.
That issue lead to a lot of rework for me one year.
Holy wars! Ok let me see.. Last time I checked the design police said..
Singletons are bad because they hinder auto testing - instances cannot be created afresh for each test case.
Instead the logic should be in a class (A) that can be easily instantiated and tested. Another class (B) should be responsible for constraining creation. Single Responsibility Principle to the fore! It should be team-knowledge that you're supposed to go via B to access A - sort of a team convention.
I concur mostly..
Many applications require that there is only one instance of some class, so the pattern of having only one instance of a class is useful. But there are variations to how the pattern is implemented.
There is the static singleton, in which the class forces that there can only be one instance of the class per process (in Java actually one per ClassLoader). Another option is to create only one instance.
Static singletons are evil - one sort of global variables. They make testing harder, because it's not possible to execute the tests in full isolation. You need complicated setup and tear down code for cleaning the system between every test, and it's very easy to forget to clean some global state properly, which in turn may result in unspecified behaviour in tests.
Creating only one instance is good. You just create one instance when the programs starts, and then pass the pointer to that instance to all other objects which need it. Dependency injection frameworks make this easy - you just configure the scope of the object, and the DI framework will take care of creating the instance and passing it to all who need it. For example in Guice you would annotate the class with #Singleton, and the DI framework will create only one instance of the class (per application - you can have multiple applications running in the same JVM). This makes testing easy, because you can create a new instance of the class for each test, and let the garbage collector destroy the instance when it is no more used. No global state will leak from one test to another.
For more information:
The Clean Code Talks - "Global State and Singletons"
Singleton as an implementation detail is fine. Singleton as an interface or as an access mechanism is a giant PITA.
A static method that takes no parameters returning an instance of an object is only slightly different from just using a global variable. If instead an object has a reference to the singleton object passed in, either via constructor or other method, then it doesn't matter how the singleton is actually created and the whole pattern turns out not to matter.
It was not just a bunch of variables in a fancy dress because this was had dozens of responsibilities, like communicating with persistence layer to save/retrieve data about the company, deal with employees and prices collections, etc.
I must say you're not really describing somthing that should be a single object and it's debatable that any of them, other than Data Serialization should have been a singelton.
I can see at least 3 sets of classes that I would normally design in, but I tend to favor smaller simpler objects that do a narrow set of tasks very well. I know that this is not the nature of most programmers. (Yes I work on 5000 line class monstrosities every day, and I have a special love for the 1200 line methods some people write.)
I think the point is that in most cases you don't need a singelton and often your just making your life harder.
The biggest problem with singletons is that they make unit testing hard, particularly when you want to run your tests in parallel but independently.
The second is that people often believe that lazy initialisation with double-checked locking is a good way to implement them.
Finally, unless your singletons are immutable, then they can easily become a performance problem when you try and scale your application up to run in multiple threads on multiple processors. Contended synchronization is expensive in most environments.
Singletons have their uses, but one must be careful in using and exposing them, because they are way too easy to abuse, difficult to truly unit test, and it is easy to create circular dependencies based on two singletons that accesses each other.
It is helpful however, for when you want to be sure that all your data is synchronized across multiple instances, e.g., configurations for a distributed application, for instance, may rely on singletons to make sure that all connections use the same up-to-date set of data.
I find you have to be very careful about why you're deciding to use a singleton. As others have mentioned, it's essentially the same issue as using global variables. You must be very cautious and consider what you could be doing by using one.
It's very rare to use them and usually there is a better way to do things. I've run into situations where I've done something with a singleton and then had to sift through my code to take it out after I discovered how much worse it made things (or after I came up with a much better, more sane solution)
I've used singletons a bunch of times in conjunction with Spring and didn't consider it a crutch or lazy.
What this pattern allowed me to do was create a single class for a bunch of configuration-type values and then share the single (non-mutable) instance of that specific configuration instance between several users of my web application.
In my case, the singleton contained client configuration criteria - css file location, db connection criteria, feature sets, etc. - specific for that client. These classes were instantiated and accessed through Spring and shared by users with the same configuration (i.e. 2 users from the same company). * **I know there's a name for this type of application but it's escaping me*
I feel it would've been wasteful to create (then garbage collect) new instances of these "constant" objects for each user of the app.
I'm reading a lot about "Singleton", its problems, when to use it, etc., and these are my conclusions until now:
Confusion between the classic implementation of Singleton and the real requirement: TO HAVE JUST ONE INSTANCE OF a CLASS!
It's generally bad implemented. If you want a unique instance, don't use the (anti)pattern of using a static GetInstance() method returning a static object. This makes a class to be responsible for instantiating a single instance of itself and also perform logic. This breaks the Single Responsibility Principle. Instead, this should be implemented by a factory class with the responsibility of ensuring that only one instance exists.
It's used in constructors, because it's easy to use and must not be passed as a parameter. This should be resolved using dependency injection, that is a great pattern to achieve a good and testable object model.
Not TDD. If you do TDD, dependencies are extracted from the implementation because you want your tests to be easy to write. This makes your object model to be better. If you use TDD, you won't write a static GetInstance =). BTW, if you think in objects with clear responsibilities instead classes, you'll get the same effect =).
I really disagree on the bunch of global variables in a fancy dress idea. Singletons are really useful when used to solve the right problem. Let me give you a real example.
I once developed a small piece of software to a place I worked, and some forms had to use some info about the company, its employees, services and prices. At its first version, the system kept loading that data from the database every time a form was opened. Of course, I soon realized this approach was not the best one.
Then I created a singleton class, named company, which encapsulated everything about the place, and it was completely filled with data by the time the system was opened.
It was not just a bunch of variables in a fancy dress because this was had dozens of responsibilities, like communicating with persistence layer to save/retrieve data about the company, deal with employees and prices collections, etc.
Plus, it was a fixed, system-wide, easily accessible point to have the company data.
Singletons are very useful, and using them is not in and of itself an anti-pattern. However, they've gotten a bad reputation largely because they force any consuming code to acknowledge that they are a singleton in order to interact with them. That means if you ever need to "un-Singletonize" them, the impact on your codebase can be very significant.
Instead, I'd suggest either hiding the Singleton behind a factory. That way, if you need to alter the service's instantiation behavior in the future, you can just change the factory rather than all types that consume the Singleton.
Even better, use an inversion of control container! Most of them allow you to separate instantiation behavior from the implementation of your classes.
One scary thing on singletons in for instance Java is that you can end up with multiple instances of the same singleton in some cases. The JVM uniquely identifies based on two elements: A class' fully qualified name, and the classloader responsible for loading it.
That means the same class can be loaded by two classloaders unaware of each other, and different parts of your application would have different instances of this singleton that they interact with.
Write normal, testable, injectable objects and let Guice/Spring/whatever handle the instantiation. Seriously.
This applies even in the case of caches or whatever the natural use cases for singletons are. There's no need to repeat the horror of writing code to try to enforce one instance. Let your dependency injection framework handle it. (I recommend Guice for a lightweight DI container if you're not already using one).