If you're in a team and a programmer gives you an interface with create, read, update and delete methods, how do you avoid type switching?
Quoting Clean Code A Handbook of Agile Software Craftsmanship:
public Money calculatePay(Employee e)
throws InvalidEmployeeType {
switch (e.type) {
case COMMISSIONED:
return calculateCommissionedPay(e);
case HOURLY:
return calculateHourlyPay(e);
case SALARIED:
return calculateSalariedPay(e);
default:
throw new InvalidEmployeeType(e.type);
}
}
There are several problems with this function. First, it’s large, and when new
employee types are added, it will grow. Second, it very clearly does more than one thing.
Third, it violates the Single Responsibility Principle7 (SRP) because there is more than one reason for it to change. Fourth, it violates the Open Closed Principle8 (OCP) because it must change whenever new types are added. But possibly the worst problem with this
function is that there are an unlimited number of other functions that will have the same
structure. For example we could have
isPayday(Employee e, Date date),
or
deliverPay(Employee e, Money pay),
or a host of others. All of which would have the same deleterious structure.
The book tells me to use the Factory Pattern, but in way that it makes me feel that I shouldn't really use it.
Quoting the book again:
The solution to this problem (see Listing 3-5) is to bury the switch statement in the
basement of an ABSTRACT FACTORY,9 and never let anyone see it.
Is the switch statement ugly?
In reality, the employee object should have its own calculate pay function that will give you the pay. This calculate pay function would change based on what type of employee it was.
That way it is up to the object to define the implementation, not the user of the object.
abstract class Employee
{
public abstract function calculatePay();
}
class HourlyEmployee extends Employee
{
public function calculatePay()
{
return $this->hour * $this->pay_rate;
}
}
class SalariedEmployee extends Employee
{
public function calculatePay()
{
return $this->monthly_pay_rate;
}
}
When you build the Factory, THEN you do the switch statement there, and only once, to build the employee.
Lets say Employee was in an array, and the type of employee was held in $array['Type']
public function buildEmployee($array)
{
switch($array['Type']){
case 'Hourly':
return new HourlyEmployee($array);
break;
case 'Salaried':
return new SalariedEmployee($array);
break;
}
Finally, to calculate the pay
$employee->calculatePay();
Now, there is no need for more than one switch statement to calculate the pay of the employee based on what type of employee they are. It is just a part of the employee object.
Disclaimer, I'm a minor, so I'm not completely positive on how some of these pays are calculated. But the base of the argument is still valid. The pay should be calculated in the object.
Disclaimer 2, This is PHP Code. But once again, the argument should be valid for any language.
You can totally remove the switch by using a Map of some kind to map the type of an employee to it's corresponding pay calculator. This depends on reflection and is possible in all languages I know.
Assuming the pay calculation is not a responsibility of an employee, we have an interface PayCalculation:
interface PayCalculation {
function calculatePay(Employee $employee);
}
There's an implementation for each category of employee:
class SalariedPayCalculator implements PayCalculation {
public function calculatePay(SalariedEmployee $employee) {
return $employee.getSalary();
}
}
class HourlyPayCalculator implements PayCalculation {
public function calculatePay(HourlyEmployee $employee) {
return $employee.getHourlyRate() * e.getHoursWorked();
}
}
class CommissionedPayCalculator implements PayCalculation {
public function calculatePay(CommissionedEmployee $employee) {
return $employee.getCommissionRate() * $employee.getUnits();
}
}
And the pay calculation would work something like this. Reflection becomes important for this to look at an object and determine it's class at run-time. With this, the switch loop can be eliminated.
public class EmployeePayCalculator implements PayCalculation {
private $map = array();
public function __construct() {
$this->map['SalariedEmployee'] = new SalariedPayCalculator();
$this->map['HourlyEmployee'] = new HourlyPayCalculator();
$this->map['CommissionedEmployee'] = new CommissionedPayCalculator();
}
public function calculatePay(Employee $employee) {
$employeeType = get_class($employee);
$calculator = $this->map[$employeeType];
return $calculator->calculatePay($employee);
}
}
Here we are initializing the map in the constructor, but it can easily be moved outside to an XML configuration file or some database:
<payCalculation>
<category>
<type>Hourly</type>
<payCalculator>HourlyPayCalculator</payCalculator>
</category>
<category>
<type>Salaried</type>
<payCalculator>SalariedPayCalculator</payCalculator>
</category>
...
</payCalculation>
I read it somewhere, that if you're using a switch, then it's suspect that there's too much variation. And when we have too much variation, we should try to encapsulate the variation behind an interface, thereby decoupling the dependencies between objects. Having said that, I think that you should try to create an SalaryType lightweight base class object that will encapsulate this type of logic. Then you make it a member of class Employee and rid yourself of the switch construct. Here's what I mean in a nutshell:
abstract class SalaryType
{
function calculatePay() {}
}
class CommissionedType extends SalaryType
{
function calculatePay() {}
}
class HourlyType extends SalaryType
{
function calculatePay() {}
}
class SalaryType extends SalaryType
{
function calculatePay() {}
}
class Employee
{
private $salaryType;
public function setType( SalaryType emp )
{
$this->salaryType = emp;
}
public function calculatePay()
{
$this->salaryType->calculatePay();
}
}
Btw, a lot of your example code does not seem very "PHP-ish". There are no return types in PHP nor is there really any type safety. Keep in mind also that PHP is not truly polymorphic, so some of the polymorphic behavior found in typical type-safe languages may not work as expected here.
Related
Basically, I want to override a parent class with different arguments. For example:
class Hold<T> {
public var value:T;
public function new(value:T) {
set(value);
}
public function set(value:T) {
this.value = value;
}
}
Then override that class, something like:
class HoldMore extends Hold<T> {
public var value2:T;
public function new(value:T, value2:T) {
super(value);
set(value, value2);
}
override public function set(value:T, value2:T) {
this.value = value;
this.value2 = value2;
}
}
Obviously this will return an error, Field set overloads parent class with different or incomplete type. Is there a way around this? I tried using a public dynamic function, and then setting set in the new() function, but that gave a very similar error. Any thoughts?
This is just a complement to #stroncium's answer, which is totally correct.
Here is an example how it could look like:
class Hold<T> {
public var value:T;
public function new(value:T) {
set(value);
}
public function set(value:T) {
this.value = value;
}
}
class HoldMore<T> extends Hold<T> {
public var value2:T;
public function new(value:T, value2:T) {
super(value);
setBoth(value, value2);
}
// you cannot override "set" with a different signature
public function setBoth(value:T, value2:T) {
this.value = value;
this.value2 = value2;
}
}
alternatively, you could use an array as parameter or a dynamic object holding multiple values in order to "set" them using the same method, but you loose some of the compiler's type checking.
If you wrote the base class you could add an optional argument to it, this would be a workaround though, not directly what you want to do.
In the current state it totally won't work. There is not only 1 problem, but few of them:
Type T is meaningless in context of this new class, you should either use some concrete type or template this class over T.
You can not change the number of arguments of function when overriding it. However you can add another function(with a different name) to accept 2 arguments and do what you want (which is the way you would use in most languages, by the way).
I don't really understand how you see a contravariance problem there. The actual problem is that haxe doesn't support function overload. (It actually does, the function signature is name + full type, but that's not what you would want to write nor support, and is mostly used for js/java externs.)
Unfortunately the language doesn't allow it.
I have an aggregate that includes the entities A, AbstractElement, X, Y and Z. The root entity is A that also has a list of AbstractElement. Entities X,Y and Z inherit from AbstractElement. I need the possibility to add instances of X, Y and Z to an instance of A. One approach is to use one method for each type, i.e. addX, addY and addZ. These methods would take as arguments the values required to create instances of X, Y and Z. But, each time I add a new type that inherits from AbstractElement, I need to modify the entity A, so I think it's not the best solution.
Another approach is to use an abstract add method addAbstractElement for adding AbstractElement instances. But, in this case, the method would take as argument an instance of AbstractElement. Because this method would be called by entities located outside of the aggregate, following DDD rules/recommandations, are these external entities authorized to create instances of AbstractElement? I read in the Eric Evans book that external entities are not authorized to hold references of entities of an aggregate other than the root?
What is the best practice for this kind of problem?
Thanks
From Evan's book, page 139:
"if you needed to add elements inside a preexisting AGGREGATE, you might create a FACTORY METHOD on the root of the AGGREGATE"
Meaning, you should create a factory method on the root (A) which will get the AbstractElement's details. This method will create the AbstractElement (X/Y/Z) according to some decision parameter and will add it to its internal collection of AbstractElements. In the end this method return the id of the new element.
Best Regards,
Itzik Saban
A few comments. As the previous answerer said, it's a good practice to use a factory method. If you can avoid it, never create objects out of the blue. Usually, it's a pretty big smell and a missed chance to make more sense out of your domain.
I wrote a small example to illustrate this. Video is in this case the aggregate root. Inside the boundaries of the aggregate are the video object and its associated comments. Comments can be anonymous or can have been written by a known user (to simplify the example, I represented the user by a username but obviously, in a real application, you would have something like a UserId).
Here is the code:
public class Video {
private List<Comment> comments;
void addComment(final Comment.Builder builder) {
this.comments.add(builder.forVideo(this).build());
// ...
}
}
abstract public class Comment {
private String username;
private Video video;
public static public class Builder {
public Builder anonymous() {
this.username = null;
return this;
}
public Builder fromUser(final String username) {
this.username = username;
return this;
}
public Builder withMessage(final String message) {
this.message = message;
return this;
}
public Builder forVideo(final Video video) {
this.video = video;
return this;
}
public Comment build() {
if (username == null) {
return new AnonymousComment(message);
} else {
return new UserComment(username, message);
}
}
}
}
public class AnonymousComment extends Comment {
// ...
}
static public class UserComment extends Comment {
// ...
}
One thing to ponder on also is that aggregate boundaries contain objects and not classes. As such, it's highly possible that certain classes (mostly value objects but it can be the case of entities also) be represented in many aggregates.
This question may appear to have been answered before but I have been unable to find exactly what I need. Here is my situation:
// Base class
interface IAnimal {};
public abstract class Animal : IAnimal{}
// Derived classes
interface IDog {}
public class Dog : Animal, IDog { }
interface ICat { }
public class Cat : Animal, ICat { }
interface ITiger { }
public class Tiger : Animal, ITiger { }
interface ILion { }
public class Lion : Animal, ILion { }
// Collection Classes
interface IPets { }
public class Pets
{
IDog dog = new Dog();
ICat cat = new Cat();
}
interface ICircus { }
public class Circus
{
ITiger tiger = new Tiger();
ILion lion = new Lion();
}
I would like to create the collections at run time in an generic Event class by reading in a list animals from xml that would make up the collection. What would be the correct way to accomplish this?
Thanks in advance.
This is kind of an answer to my own question. Maybe this will help others.
I chose a very generic example to illustrate my situation because I have uses for this in many places in Windows Forms, XNA and Silverlight that are all very different.
When I used the Activator, I found out that it assumes the executing assembly. My method is in a library so I had to load a different assembly. Next I had to make sure that I had the right namespace. My base class is in a library and the derived classes are in another namespace so this will require refactoring to properly create the list.
Another problem I found was that the Activator assumes a constructor with no parameters. In my test case all my derived classes are XNA game components with a parameter of type Game.
Have to do some refactoring to test out the interfaces and how the game objects are to interact.
Will be back to this list when I have something further.
Does this sort of example help? (It's from some of my code I happened to have handy.) The key point here is the use of reflection in Activator.CreateInstance(...).
public static List<dynamic> LoadChildEntities(XElement entityElt)
{
var children = new List<dynamic>();
foreach(XElement childElt in entityElt.Elements("entity"))
{
// Look up the C# type of the child entity.
string childTypename = "MyNamespace." + Convert.ToString(childElt.Attribute("type").Value);
Type childType = Type.GetType(childTypename);
if(childType != null)
{
// Construct the child entity and add it to the list.
children.Add(Activator.CreateInstance(childType, childElt));
}
else
{
throw new InvalidOperationException("No such class: " + childTypename);
}
}
return children;
}
If you want a list of IAnimal instead, it wouldn't be too tricky to change.
I'm struggling with implementing a factory object. Here's the context :
I've in a project a custom store. In order to read/write records, I've written this code in a POCO model/separated repository:
public class Id { /* skip for clarity*/} // My custom ID representation
public interface IId
{
Id Id { get; set; }
}
public interface IGenericRepository<T> where T : IId
{
T Get(Id objectID);
void Save(T #object);
}
public interface IContext
{
TRepository GetRepository<T, TRepository>()
where TRepository : IGenericRepository<T>
where T:IId;
IGenericRepository<T> GetRepository<T>()
where T:IId;
}
My IContext interface defines two kind of repositories.
The former is for standard objects with only get/save methods, the later allows me to define specifics methods for specific kind of objects. For example :
public interface IWebServiceLogRepository : IGenericRepository<WebServiceLog>
{
ICollection<WebServiceLog> GetOpenLogs(Id objectID);
}
And it the consuming code I can do one of this :
MyContext.GetRepository<Customer>().Get(myID); --> standard get
MyContext.GetRepository<WebServiceLog, IWebServiceLogRepository>().GetOpenLogs(myID); --> specific operation
Because most of objects repository are limited to get and save operations, I've written a generic repository :
public class BaseRepository<T> : IGenericRepository<T>
where T : IId, new()
{
public virtual T Get(Id objectID){ /* provider specific */ }
public void Save(T #object) { /* provider specific */ }
}
and, for custom ones, I simply inherits the base repository :
internal class WebServiceLogRepository: BaseRepository<WebServiceLog>, IWebServiceLogRepository
{
public ICollection<WebServiceLog> GetByOpenLogsByRecordID(Id objectID)
{
/* provider specific */
}
}
Everything above is ok (at least I think it's ok). I'm now struggling to implement the MyContext class. I'm using MEF in my project for other purposes. But because MEF doesn't support (yet) generic exports, I did not find a way to reach my goal.
My context class is looking like by now :
[Export(typeof(IContext))]
public class UpdateContext : IContext
{
private System.Collections.Generic.Dictionary<Type, object> m_Implementations;
public UpdateContext()
{
m_Implementations = new System.Collections.Generic.Dictionary<Type, object>();
}
public TRepository GetRepository<T, TRepository>()
where T : IId
where TRepository : IGenericRepository<T>
{
var tType = typeof(T);
if (!m_Implementations.ContainsKey(tType))
{
/* this code is neither working nor elegant for me */
var resultType = AppDomain.CurrentDomain.GetAssemblies().SelectMany(
(a) => a.GetTypes()
).Where((t)=>t.GetInterfaces().Contains(typeof(TRepository))).Single();
var result = (TRepository)resultType.InvokeMember("new", System.Reflection.BindingFlags.CreateInstance, null, null, new object[] { this });
m_Implementations.Add(tType, result);
}
return (TRepository)m_Implementations[tType];
}
public IGenericRepository<T> GetRepository<T>() where T : IId
{
return GetRepository<T, IGenericRepository<T>>();
}
}
I'd appreciate a bit of help to unpuzzle my mind with this quite common scenario
Not sure if I've understood you correctly, but I think you're perhaps over complicating things. To begin with, make sure you've designed your code independent of any factory or Dependency Injection framework or composition framework.
For starters lets look at what you want your calling code to look like, this is what you said:
MyContext.GetRepository<Customer>().Get(myID); --> standard get
MyContext.GetRepository<WebServiceLog, IWebServiceLogRepository>().GetOpenLogs(myID);
You don't have to agree with my naming choices below, but it indicates what I undertand from your code, you can tell me if I'm wrong. Now, I feel like the calling would be simpler like this:
RepositoryFactory.New<IRepository<Customer>>().Get(myId);
RepositoryFactory.New<IWebServiceLogRepository>().GetOpenLogs(myId);
Line 1:
Because the type here is IRepository it's clear what the return type is, and what the T type is for the base IRepository.
Line 2:
The return type here from the factory is IWebServiceLogRepository. Here you don'y need to specify the entity type, your interface logically already implements IRepository. There's no need to specify this again.
So your interface for these would look like this:
public interface IRepository<T>
{
T Get(object Id);
T Save(T object);
}
public interface IWebServiceLogRepository: IRepository<WebServiceLog>
{
List<WebServiceLog> GetOpenLogs(object Id);
}
Now I think the implementations and factory code for this would be simpler as the factory only has to know about a single type. On line 1 the type is IRepository, and in line 2, IWebServiceLogRepository.
Try that, and try rewriting your code to simply find classes that implement those types and instantiating them.
Lastly, in terms of MEF, you could carry on using that, but Castle Windsor would really make things much simpler for you, as it lets you concentrate on your architecture and code design, and its very very simple to use. You only ever reference Castle in your app startup code. The rest of your code is simply designed using the Dependency Injection pattern, which is framework agnostic.
If some of this isn't clear, let me know if you'd like me to update this answer with the implementation code of your repositories too.
UPDATE
and here's the code which resolves the implementations. You were making it a bit harder for yourself by not using the Activator class.
If you use Activator and use only one Generic parameter as I've done in the method below, you should be ok. Note the code's a bit rough but you get the idea:
public static T GetThing<T>()
{
List<Type> assemblyTypes = AppDomain.CurrentDomain.GetAssemblies()
.SelectMany(s => s.GetTypes()).ToList();
Type interfaceType = typeof(T);
if(interfaceType.IsGenericType)
{
var gens = interfaceType.GetGenericArguments();
List<Type> narrowed = assemblyTypes.Where(p => p.IsGenericType && !p.IsInterface).ToList();
var implementations = new List<Type>();
narrowed.ForEach(t=>
{
try
{
var imp = t.MakeGenericType(gens);
if(interfaceType.IsAssignableFrom(imp))
{
implementations.Add(imp);
}
}catch
{
}
});
return (T)Activator.CreateInstance(implementations.First());
}
else
{
List<Type> implementations = assemblyTypes.Where(p => interfaceType.IsAssignableFrom(p) && !p.IsInterface).ToList();
return (T)Activator.CreateInstance(implementations.First());
}
}
I know that the Specification pattern describes how to use a hierarchy of classes implementing ISpecification<T> to evaluate if a candidate object of type T matches a certain specification (= satisfies a business rule).
My problem : the business rule I want to implement needs to evaluate several objects (for example, a Customer and a Contract).
My double question :
Are there typical adaptations of the Specification patterns to achieve this ? I can only think of removing the implementation of ISpecification<T> by my specification class, and taking as many parameters as I want in the isSatisfiedBy() method. But by doing this, I lose the ability to combine this specification with others.
Does this problem reveal a flaw in my design ? (i.e. what I need to evaluate using a Customer and a Contract should be evaluated on another object, like a Subscription, which could contain all the necessary info) ?
In that case (depending on what the specification precisely should do, I would use one of the objects as specification subject and the other(s) as parameter.
Example:
public class ShouldCreateEmailAccountSpecification : ISpecification<Customer>
{
public ShouldCreateEmailAccountSpecification(Contract selectedContract)
{
SelectedContract = selectedContract;
}
public Contract SelectedContract { get; private set; }
public bool IsSatisfiedBy(Customer subject)
{
return false;
}
}
Your problem is that your specification interface is using a generic type parameter, which prevents it from being used for combining evaluation logic across different specializations (Customer,Contract) because ISpecification<Customer> is in fact a different interface than ISpecification<Contract>. You could use Jeff's approach above, which gets rid of the type parameter and passes everything in as a base type (Object). Depending on what language you are using, you may also be able to pull things up a level and combine specifications with boolean logic using delegates. C# Example (not particularly useful as written, but might give you some ideas for a framework):
ISpecification<Customer> cust_spec = /*...*/
ISpecification<Contract> contract_spec = /*... */
bool result = EvalWithAnd( () => cust_spec.IsSatisfiedBy(customer), () => contract_spec.IsSatisfiedBy( contract ) );
public void EvalWithAnd( params Func<bool>[] specs )
{
foreach( var spec in specs )
{
if ( !spec() )
return false; /* If any return false, we can short-circuit */
}
return true; /* all delegates returned true */
}
Paco's solution of treating one object as the subject and one as a parameter using constructor injection can work sometimes but if both objects are constructed after the specification object, it makes things quite difficult.
One solution to this problem is to use a parameter object as in this refactoring suggestion: http://sourcemaking.com/refactoring/introduce-parameter-object.
The basic idea is that if you feel that both Customer and Contract are parameters that represent a related concept, then you just create another parameter object that contains both of them.
public class ParameterObject
{
public Customer Customer { get; set; }
public Contract Contract { get; set; }
}
Then your generic specification becomes for that type:
public class SomeSpecification : ISpecification<ParameterObject>
{
public bool IsSatisfiedBy(ParameterObject candidate)
{
return false;
}
}
I don't know if I understood your question.
If you are using the same specification for both Customer and Contract, this means that you can send the same messages to both of them. This could be solved by making them both to implement an interface, and use this interface as the T type. I don't know if this makes sense in your domain.
Sorry if this is not an answer to your question.