So I'm making this web application in PHP, and I wanted to make a decent model layer (as much as possible) with Domain Objects and Data Mappers, all orchestrated by Services.
And now I'm thinking, how should I create my objects?
Factories?
But then:
I'd have to use reflection to actually enforce if an object implements a certain interface (DataMapper usually. The others don't have set methods they need to implement; I don't want to rely on names, because naming conventions change).
I'd have to defy the law of demeter by passing the factories into the ServiceFactory which would then only pass it to the service it creates.
The new keyword?
But then:
I'd have problems testing the code in isolation.
Feels too tightly coupled.
Anything else?
I don't know, that's why I'm asking! :P
What should I do? Is there a better approach to this?
Brian Vanderbusch's comment hints at the best way. You should use Dependency Injection to inject the mapper you need. However, to reduce coupling you should also type hint the most basic type, usually an interface.
How you actually inject the Data Mapper into your (presumably) model layer doesn't really matter. A Dependency Injection Container can store all the metadata about which classes need which data mappers and automatically inject the dependencies for you. However, without using a DIC, the top level of your code should, at the most basic level, look like this:
$model = new ProductsModel(new ProductsDataMapper(new Db('localhost', 'u', 'p', 'd')));
of course the mapper and database would almost certainly be shared between other objects in the real world and you'd pass in the references to the existing instances. If you need multiple mappers in the model layer, pass them all in as constructor arguments.
The key here is to build the entire object graph at the very top level of the application rather than worrying about locating them after the fact. This is the premise of inversion of control, which you can think of as building the structure from inside out.
Related
I've been investigating a lot about dependency injection theory and it makes great sense except for the fact that it seems to introduce complexity/bloat in certain scenarios.
Firstly, it is assumed that a DI Container instance is never passed into any object. Bad practice and all that...
For example: Consider a "user" class that relates to all other business objects created in the system by/for a specific user.
If a user instance is to be deleted, all related objects (i.e. records, files, images, etc.) also have to be deleted. Does this mean that an instance of each and every dependency is to be injected into the user instance to allow for the lookup and deletion of all related objects? i.e. an instance of ImageMapper (or ImageMapperFactory) must be passed in in order to delete all images created/uploaded by the user instance being deleted?
If not, is such a scenario a good example of where a service locator should be used?
I find it repeated over and over again in the articles & tutorials that a programmer must avoid using "new" within any class like the plague. However, is this really feasible for instances of controllers that may need to create a variety of views or the like?
A concrete example of adherence to the SOLID mantra, at least as far as DI is concerned, would be appreciated... One that makes it clear how one would either stuff instances of ALL required dependencies into a controller class or how best instances of such dependencies would be located or created?
I don't know PHP but I'll try to explain some things so your question doesn't get neglected, so bear with me. ;-)
I find that dependency injection works best by separating your application in layers:
Presentation Layer - Some (user) interface that provides interaction with your application.
Business Logic/Service Layer - The actual behavior of your application.
Persistence/Data Access Layer - The layer that stores/reads data from a back-end (try searching for repository pattern).
These layers all transfer data to each other in a structured way, using domain models. These models can be for example a user and an image.
When your application logic states that a user's images are to be deleted when a user is deleted you'll want this in the middle layer (Business Logic). The business logic layer calls the data access layer in order to delete entities.
Let's demonstrate using Python (I hope you can read that).
# Define a class that manages users.
class UserService:
# A constructor accepting dependent DAL instances.
def __init__(self, user_repo, image_repo):
self.user_repo = user_repo
self.image_repo = image_repo
def delete(self, user):
self.user_repo.delete(user)
self.image_repo.delete(user.profile_picture)
The repo arguments that go in the constructor are instances of repository classes, this is dependency injection (constructor injection).
The important thing to remember is that an object doesn't instantiate another object, it accepts instances of objects.
I have some confusion between those two patterns:
singleton
Check if the instance exists return it, or create a new one.
Registry
Check if the instance exists return it, or create a new one and store it.
What the difference between them?
Both are about instance control. The difference is that Singleton allows only one instance of a given class while Registry holds a 1-1 map of keys to instances. Typically, the key is (or represents) a class and the value is an instance of that class.
For example, Code Igniter framework holds a registry with an instance of each library/model/controller/helper you load and returning those same instances every time.
Level of access and complexity
One fun filled caveat I always love, is that a registry pattern is utilizing a singleton to store/retrieve a dynamic list of objects.
Singleton Pattern
Singleton pattern is a pattern by which you establish a single point of access and methodology. While many people believe it is an anti-pattern, I assure you it is only because of the level they work at and type of programming they do.
Singleton patterns are used at lower levels to manage strict resources. For instance, while most languages have libraries that wrap keyboard input, it is effectively a singleton pattern as it has very strict input/output and only one can exist, regardless of how many keyboards you have.
Testing Singleton patterns
One of the reasons people feel singleton patterns are anti-patterns, is they are in general untestable in a unit test system. It is not possible to even mock a singleton in most languages, as it is fixed to a specific identity.
For this reason the singleton should be minimal, and a wrapper used to interface it.
You can then inject the wrapper and it can be mocked to used a fake singleton mock.
Registry Pattern
The registry pattern is where you establish a single point of access, but not a single methodology. This very registry access will have at least 3 methods: put, find, and delete. These methods rely on a singleton pattern to work.
The map is testable, as it is not isolated to your use, and while the registry itself is not actually unit testable in connection to the units that consume it, you can utilize mocks of everything the registry is intended to contain.
However, if you utilize IoC in a given architecture consuming it, you can inject the registry from the container registry, allowing a mocked registry to be injected instead.
Final note
Inversion of control is vastly popular these days, however big parts of it are often not learned or ignored. Some frameworks boil it down to mostly the factory pattern for most things, abstracting the rest away.
I would strongly encouraging most to toss out everything they know about IoC.
Then learn SOLID.
Afterwards read a book on design patterns for OOD. (Note: If the book has pros and cons of the patterns, it is good, if it does not show the cons, I would be very wary of the book.)
While these terms were coined relatively recently, the principles were learned in one manner or another early on, and in fact became the basis of Object Oriented Design.
The patterns most object oriented developers use successfully are all based on SOLID principles.
A reference to SOLID, that is laid out and summed up.
https://www.digitalocean.com/community/conceptual_articles/s-o-l-i-d-the-first-five-principles-of-object-oriented-design
As a final suggestion: Even if you are not an object oriented developer, it is very useful to learn SOLID and OOD, as must system architecture follows similar patterns when integrating a system together.
In addition I would not suggest you assume that one language/pattern/philosophy fits best for all needs. Keep learning, keep an open mind. The more you know, the better the solutions you can bring to the table.
A Registry differs in that its principal purpose is to allow you to navigate to associated objects. See Martin Fowler.
The main difference between registry and singleton is that singleton allow us to create a single instance of a class at a time, whereas registry allows us to create multiple instance of the same class.
Registry Definition: When you want to find an object you usually start with another object that has an association to it, and use the association to navigate to it. Thus, if you want to find all the orders for a customer, you start with the customer object and use a method on it to get the orders. However, in some cases you won't have an appropriate object to start with. You may know the customer's ID number but not have a reference. In this case you need some kind of lookup method - a finder - but the question remains: How do you get to the finder?
A Registry is essentially a global object, or at least it looks like one - even if it isn't as global as it may appear.
Singleton Definition: Sometimes it's important to have only one instance for a class. For example, in a system there should be only one window manager (or only a file system or only a print spooler). Usually singletons are used for centralized management of internal or external resources and they provide a global point of access to themselves.
Based on these definitions their usage are completely different.
Why should I use Zend_Registry instead of the Singleton pattern?
My coworker and me recently had a discussion about this. His point was that we should use Zend_Registry for all consistent objects, but I wanted to use the singleton pattern, since Zend_Registry just does the same, but wrapped.
I have a problem with code like this:
$list = Zend_Registry::get('database')->getList($sql);
Since theres a chance that database isn't in Zend_Registry. In cases of lazy loading, I would have to make my own registry with information about specific objects in my system. Like if the database takes specific parameters on loadtime, then it would have to know this.
I would instead use getInstance and then all the code would reside in the same object. Does that make sense?
Even though you phrased your question as an either/or, might I suggest a third alternative?
I try to avoid both singletons and Zend_Registry wherever possible, since they function, in effect, as globals. When a segment of code can reach into the global ether - via a call to a singleton or a global registry - to get something it needs, it creates a hidden - or at least, a non-explicit - dependency that makes things harder to debug and unit-test.
In contrast, I try to follow dependency injection advice, paraphrased as: "Give a component what it needs. Don't make it find what it needs."
I find that for most entities for which I might feel I need a registry/singleton - db connections, loggers, etc - I can create them at Bootstrap, store them in the Bootstrap registry and inject them into my controllers, usually during init() using $this->getInvokeArg('bootstrap')->getResource('myResource'). Only controllers reach back into the Bootstrap. Then, any models or services that need these dependencies get them passed-in explicitly by the controller, either via constructor or by setter injection.
A hybrid approach to which I do sometimes fall back is to design my service/model classes with getters/setters for these dependencies - getDbAdapter() and setDbAdapter(); getLogger() and setLogger(), etc. The getter lazy-loads from the global registry - whether some singleton or by Zend_Registry, throwing exceptions when they are not where I expect them to be. In that sense, it is similar to what you are suggesting. It does violate the purist dependency injection philosophy. But at least the presence of the getter/setter methods explicitly demonstrates that there is a dependency and allows me to mock it out or to provide non-default implementations.
It does for simple blog or something. Otherwise you're stuck with only one DB instance. And that's NOT what you want in the long run. You may want to connect to other server (to log errors to central db, to import products from someone, ...) or connect as different user (for security reasons - you don't want your API to have access to admin_users table, but you still need to connect to it to check if user is valid in the first place).
You can do one-purpose registers (My_Db_Admin, My_Db_ReadOnly, ...) but that does not make much sense to me. Using registry you're not stuck with one instance. You can create one outside registry and work with it for a while and then trash it ;)
I'm currently rewriting an e-shop - but only the client side, i.e. the CMS remains mostly in tact. I am not using a pre-built framework, as the system has to retain backwards compatibility with the CMS and I have to have full freedom of the structure of code.
The new system is purely MVC based and I have a Bootstrapper which loads controllers based on the current uri and the latter use models for the real work - both with sessions and the database.
tl;dr It's my first project without a pre-built framework.
I am very inexperienced when it comes to design patterns. I know how do most of the popular ones work but have had never put them to use.
Now I am suspecting code smells because all of my models are classes that consist purely of static methods. I can find no advantages of doing them in a different manner. I routinely need some of the methods in various places through out the code. I.e. I need to fetch the logged in user in the main layout, check user rights to see current page in the bootstraper, display user panel by the controller. I'd need to re-instantiate an object each time or keep a global one if I wasn't using statics. There also won't be a need for more than one such class at a time.
I must be missing something, because even though I use OOP, some my classes are just meaningless containers for their methods (and sometimes a couple of private variables). I could have just been using PHP4 and simple functions.
Any comments or advice would be highly appreciated.
EDIT: in spite of all these educated answers, I remain unconvinced. Even though it's most probably because of my lack of experience, I still don't foresee anything going wrong with the current setup. I mean I don't even fathom a situation where I'd have any inconveniences due to the code architecture as it is now. I hope I don't get a harsh lesson when it's too late to change anything...
You are right, it's a code smell and everybody will tell you it's baaaad.
So here I suggest rather to make a self-assessment of the severity of the problem:
Do you have classes with many getter and setter?
Are your static functions like the one below?
If yes, try to move the logic in the class MyClass that will be already way more OO. That's a classic mistake from procedural/scripting world.
static void myMethod( MyClass anObject )
{
// get value from anObject
// do some business logic
// set value of anObject
}
Do you have a lot of global state, such as data you fetch from the current session?
If yes, make an assessment whether you want to change it. The OO way would be to pass the session down the call chain. But in practice, it's convenient to access the session as a global object. But it impedes testability. Try to remove some global state and turn that into regular object that you pass and manipulate in methods.
Make this assessment, and try to identify utility classes, services classes and the business objects. Utility class are helper classes with utility methods (e.g. formatting, conversion, etc.) which can be static. Service class do some business logic but they should be stateless and one instance suffice. Business objects are user, products, article, etc. is where you must concentrate your effort. Try to turn plain data into objects with embed some behavior.
Have a look at should entity be dumb. Even if it's for java, the concepts are general.
EDIT
Here is my analysis based on your comment:
You don't have a domain model with entities. You manipulate the database directly.
What you call your model, is what I call services and is where you perform the business logic that manipulate data. Service classes are stateless, which is correct. As you pointed out in the question, you then either need to constantly re-create them, create one global instance, or use static methods.
The OO paradigm would say that you should try to have a domain model where you map your database with entities. At least have an anemic domain model where entities are dull data container that are loaded/persisted in database. Then the OO paradigm would also say to put a bit of logic in the entities if possible.
It would also say to turn the services into objects to ease composition and reuse. If it was the case you could for instance wrap all services with an interceptor to start/stop transactions or do some security check, which you won't be able to do with static methods.
What you describe (no entities + stateless procedural services) is not considered a great OO design. I would suggest you introduce an anemic domain model at least and DAO. Regarding the sateless procedural services, this is actually the reality of many web applications -- if you don't need more you can stick to it.
My 2 cents
If you are mainly only using static classes then you've really taken out the object out of object oriented programming. I am not saying you are doing things incorrectly, I am saying maybe your system shouldn't lend itself to OOP. Maybe it is a simple app that requires some basic utility functions (sends email, etc). In this case most of your code becomes very procedural.
If you are dealing with databases you could have a static db class, and a simple business layer, and your php app interacts with your business layer which in turn interacts with your database layer. This becomes your typical 3-tier architecture (some people like to refer to this as 4 t-iers and seperate the actual database from the data layer, same thing).
If you are not really calling methods that require an object than what is the point of all of these static classes, just a question to ask yourself.
One thing you may notice is that if you plan on doing any kind of unit testing with mocking/stubbing you are probably going to have a hard time since static classes and methods are not easy to mock, stub or test.
I would be cautious about using static variables and classes in web applications. If you really must share an instance between users, then it should be ok to have a single instance (lookup "Singleton" design pattern).
However, if you trying to maintain state across pages, you should do this through either utilising the ViewState or by using the Session object.
If you were to have a global static variable, you could have a situation where concurrent users are fighting to update the same value.
Short answer: It's ok but you are foregoing the benefits of OOP.
One reasoning behind using objects is that most of the time there is more than one type of object that performs a role. For example you can swap your DBVendor1 data access object with a DBVendor2 data access object that has the same interface. This especially handy if you are using unit tests and need to swap objects that do real work with dummy objects (mocks and stubs). Think of your objects with the same interface as Lego bricks with different colors that fit together and are easily interchangeable. And you simply can't do that with static objects.
Of course, the increased flexibility of the objects comes at a price: The initialization of the objects and putting them together is more work (like you wrote) and leads to more code and objects that put together other objects. This is where creational design patterns like builder and factory come into play.
If you want to go that route, I advise you to read about dependency injection and using a DI framework.
Technically there is nothing wrong in doing it. But practically you are loosing lot of the benefits of object oriented programming. Also write the code/functionality where it belong to.. for example:
user.doSomeTask()
on the user object makes more sense than
UserUtils.doSomeTask(User user)
Using OOP concepts you abstract the functionality where it belongs to and in future it helps you change your code, extend the functionality more easily than using the static methods.
There are advantages to using static methods. One being that since you cannot inherit them they perform better. But using them all of the time limits you. The whole OOP paradigm
is based on re-usability of base classes thorough the use of inheritance.
I'd like to:
Make commonly required services visible to all classes that need them,
with a minimum of boilerplate, and
without sacrificing testability!
It's a small project and I think DI might be overkill, but maybe I'm wrong? Anyhow, I have been focusing on the ServiceLocator pattern as described by Martin Fowler
In a client class' constructor, I have something like this:
this->db = Locator::getDb();
this->log = Locator::getLogger();
Then the rest of the class' methods access the service through those member attributes, e.g.:
this->fooModel = new fooModel(this->db);
fooItem1234 = this->fooModel->findById(1234);
However I would also like this level of visibility for "model" objects (like fooModel above) because they are accessed from several different places and there is no need to have more than one instance.
So my initial thought was to extend Locator to have a ::getFooModel() but now it seems I'm violating the Open/Closed Principle, since I'll have to modify Locator every time I introduce a new model class.
To satisfy OCP, I could use the Dynamic Service Locator (also described on Fowler's page) however I'm not totally sold on this for the same reasons as him, i.e. it's not explicit enough.
Another solution would be to just make all my models' methods static. So:
fooItem1234 = FooModel::findById(1234);
I like this because it's zero boilerplate. I can just create a new model class and start calling it from anywhere with a single line. But now the model depends on Locator to find its DB connection and I'm not sure how I feel about that. For one, if I ever needed to have two fooModels open on separate database connections, it would be a mess and/or impossible. That said, I don't actually need to do that currently so this option seems a little tempting.
Finally, there's DI. But like I said above I think it might be too much for this little project.
Conclusion: I'm a little stuck here and would appreciate some advice from the gurus of StackOverflow!
Why do you think that DI is overkill for your project? DI patterns such as Constructor Injection is way simpler and cleaner than Service Locator (which I consider an anti-pattern).
I consider Service Locator to be an anti-pattern since it is totally opaque to the user of the API which dependencies need to be in place; thus, one could easily invoke methods on your objects in a context where the Service Locator would throw, and the API gives you absolutely no clue that this is the case.
You don't need a DI Container to use DI. If just have a simple project, you can use what is known as Poor Man's DI where you wire up dependencies manually.
... and there is no need to have more than one instance.
You're mixing apples and oranges. The fact that you only need one instance of a class for an application, is not the same thing as it being a good idea to make that instance globally available. With DI you don't change the cardinality - there is still just one instance. What you change is the scope of variables that address said instance. There's a difference.