Related
I'm pondering the design of a C# library, that will have several different high level functions. Of course, those high-level functions will be implemented using the SOLID class design principles as much as possible. As such, there will probably be classes intended for consumers to use directly on a regular basis, and "support classes" that are dependencies of those more common "end user" classes.
The question is, what is the best way to design the library so it is:
DI Agnostic - Although adding basic "support" for one or two of the common DI libraries (StructureMap, Ninject, etc) seems reasonable, I want consumers to be able to use the library with any DI framework.
Non-DI usable - If a consumer of the library is using no DI, the library should still be as easy to use as possible, reducing the amount of work a user has to do to create all these "unimportant" dependencies just to get to the "real" classes they want to use.
My current thinking is to provide a few "DI registration modules" for the common DI libraries (e.g a StructureMap registry, a Ninject module), and a set or Factory classes that are non-DI and contain the coupling to those few factories.
Thoughts?
This is actually simple to do once you understand that DI is about patterns and principles, not technology.
To design the API in a DI Container-agnostic way, follow these general principles:
Program to an interface, not an implementation
This principle is actually a quote (from memory though) from Design Patterns, but it should always be your real goal. DI is just a means to achieve that end.
Apply the Hollywood Principle
The Hollywood Principle in DI terms says: Don't call the DI Container, it'll call you.
Never directly ask for a dependency by calling a container from within your code. Ask for it implicitly by using Constructor Injection.
Use Constructor Injection
When you need a dependency, ask for it statically through the constructor:
public class Service : IService
{
private readonly ISomeDependency dep;
public Service(ISomeDependency dep)
{
if (dep == null)
{
throw new ArgumentNullException("dep");
}
this.dep = dep;
}
public ISomeDependency Dependency
{
get { return this.dep; }
}
}
Notice how the Service class guarantees its invariants. Once an instance is created, the dependency is guaranteed to be available because of the combination of the Guard Clause and the readonly keyword.
Use Abstract Factory if you need a short-lived object
Dependencies injected with Constructor Injection tend to be long-lived, but sometimes you need a short-lived object, or to construct the dependency based on a value known only at run-time.
See this for more information.
Compose only at the Last Responsible Moment
Keep objects decoupled until the very end. Normally, you can wait and wire everything up in the application's entry point. This is called the Composition Root.
More details here:
Where should I do Injection with Ninject 2+ (and how do I arrange my Modules?)
Design - Where should objects be registered when using Windsor
Simplify using a Facade
If you feel that the resulting API becomes too complex for novice users, you can always provide a few Facade classes that encapsulate common dependency combinations.
To provide a flexible Facade with a high degree of discoverability, you could consider providing Fluent Builders. Something like this:
public class MyFacade
{
private IMyDependency dep;
public MyFacade()
{
this.dep = new DefaultDependency();
}
public MyFacade WithDependency(IMyDependency dependency)
{
this.dep = dependency;
return this;
}
public Foo CreateFoo()
{
return new Foo(this.dep);
}
}
This would allow a user to create a default Foo by writing
var foo = new MyFacade().CreateFoo();
It would, however, be very discoverable that it's possible to supply a custom dependency, and you could write
var foo = new MyFacade().WithDependency(new CustomDependency()).CreateFoo();
If you imagine that the MyFacade class encapsulates a lot of different dependencies, I hope it's clear how it would provide proper defaults while still making extensibility discoverable.
FWIW, long after writing this answer, I expanded upon the concepts herein and wrote a longer blog post about DI-Friendly Libraries, and a companion post about DI-Friendly Frameworks.
The term "dependency injection" doesn't specifically have anything to do with an IoC container at all, even though you tend to see them mentioned together. It simply means that instead of writing your code like this:
public class Service
{
public Service()
{
}
public void DoSomething()
{
SqlConnection connection = new SqlConnection("some connection string");
WindowsIdentity identity = WindowsIdentity.GetCurrent();
// Do something with connection and identity variables
}
}
You write it like this:
public class Service
{
public Service(IDbConnection connection, IIdentity identity)
{
this.Connection = connection;
this.Identity = identity;
}
public void DoSomething()
{
// Do something with Connection and Identity properties
}
protected IDbConnection Connection { get; private set; }
protected IIdentity Identity { get; private set; }
}
That is, you do two things when you write your code:
Rely on interfaces instead of classes whenever you think that the implementation might need to be changed;
Instead of creating instances of these interfaces inside a class, pass them as constructor arguments (alternatively, they could be assigned to public properties; the former is constructor injection, the latter is property injection).
None of this presupposes the existence of any DI library, and it doesn't really make the code any more difficult to write without one.
If you're looking for an example of this, look no further than the .NET Framework itself:
List<T> implements IList<T>. If you design your class to use IList<T> (or IEnumerable<T>), you can take advantage of concepts like lazy-loading, as Linq to SQL, Linq to Entities, and NHibernate all do behind the scenes, usually through property injection. Some framework classes actually accept an IList<T> as a constructor argument, such as BindingList<T>, which is used for several data binding features.
Linq to SQL and EF are built entirely around the IDbConnection and related interfaces, which can be passed in via the public constructors. You don't need to use them, though; the default constructors work just fine with a connection string sitting in a configuration file somewhere.
If you ever work on WinForms components you deal with "services", like INameCreationService or IExtenderProviderService. You don't even really know what what the concrete classes are. .NET actually has its own IoC container, IContainer, which gets used for this, and the Component class has a GetService method which is the actual service locator. Of course, nothing prevents you from using any or all of these interfaces without the IContainer or that particular locator. The services themselves are only loosely-coupled with the container.
Contracts in WCF are built entirely around interfaces. The actual concrete service class is usually referenced by name in a configuration file, which is essentially DI. Many people don't realize this but it is entirely possible to swap out this configuration system with another IoC container. Perhaps more interestingly, the service behaviors are all instances of IServiceBehavior which can be added later. Again, you could easily wire this into an IoC container and have it pick the relevant behaviors, but the feature is completely usable without one.
And so on and so forth. You'll find DI all over the place in .NET, it's just that normally it's done so seamlessly that you don't even think of it as DI.
If you want to design your DI-enabled library for maximum usability then the best suggestion is probably to supply your own default IoC implementation using a lightweight container. IContainer is a great choice for this because it's a part of the .NET Framework itself.
EDIT 2015: time has passed, I realize now that this whole thing was a huge mistake. IoC containers are terrible and DI is a very poor way to deal with side effects. Effectively, all of the answers here (and the question itself) are to be avoided. Simply be aware of side effects, separate them from pure code, and everything else either falls into place or is irrelevant and unnecessary complexity.
Original answer follows:
I had to face this same decision while developing SolrNet. I started with the goal of being DI-friendly and container-agnostic, but as I added more and more internal components, the internal factories quickly became unmanageable and the resulting library was inflexible.
I ended up writing my own very simple embedded IoC container while also providing a Windsor facility and a Ninject module. Integrating the library with other containers is just a matter of properly wiring the components, so I could easily integrate it with Autofac, Unity, StructureMap, whatever.
The downside of this is that I lost the ability to just new up the service. I also took a dependency on CommonServiceLocator which I could have avoided (I might refactor it out in the future) to make the embedded container easier to implement.
More details in this blog post.
MassTransit seems to rely on something similar. It has an IObjectBuilder interface which is really CommonServiceLocator's IServiceLocator with a couple more methods, then it implements this for each container, i.e. NinjectObjectBuilder and a regular module/facility, i.e. MassTransitModule. Then it relies on IObjectBuilder to instantiate what it needs. This is a valid approach of course, but personally I don't like it very much since it's actually passing around the container too much, using it as a service locator.
MonoRail implements its own container as well, which implements good old IServiceProvider. This container is used throughout this framework through an interface that exposes well-known services. To get the concrete container, it has a built-in service provider locator. The Windsor facility points this service provider locator to Windsor, making it the selected service provider.
Bottom line: there is no perfect solution. As with any design decision, this issue demands a balance between flexibility, maintainability and convenience.
What I would do is design my library in a DI container agnostic way to limit the dependency on the container as much as possible. This allows to swap out on DI container for another if need be.
Then expose the layer above the DI logic to the users of the library so that they can use whatever framework you chose through your interface. This way they can still use DI functionality that you exposed and they are free to use any other framework for their own purposes.
Allowing the users of the library to plug their own DI framework seems a bit wrong to me as it dramatically increases amount of maintenance. This also then becomes more of a plugin environment than straight DI.
So I'm in the middle of rafactoring a small to medium sized Windows Forms application backed by a SQLite database accessed through NHibernate. The current solution contains only an App Project and Lib Project so it is not very well structured and tightly coupled in many places.
I started off with a structure like in this answer but ran into some problems down the road.
DB initialization:
Since the code building the NHibernate SessionFactory is in the DAL and I need to inject an ISession into my repositories, I need to reference the DAL and NHibernate in my Forms project directly to be able to set up the DI with Ninject (which should be done in the App Project / Presentation Layer right?)
Isn't that one of the things I try to avoid with such an architecture?
In an ideal world which projects should reference eachother?
DI in general:
I have a decently hard time figuring out how to do DI properly. I read about using a composition root to only have one place where the Ninject container is directly used but that doesn't really play well with the current way NHibernate Sessions are used.
We have a MainForm which is obviously the applications entry point and keeps one Session during its whole lifetime. In addition the user can open multiple SubForms (mostly but not exclusively) for editing single entities) which currently each have a separate Session with a shorter lifetime. This is accomplished with a static Helper exposing the SessionFactory and opening new Sessions as required.
Is there another way of using DI with Windows Forms besides the composition root pattern?
How can I make use of Ninjects capabilites to do scoped injection to manage my NHibernate Sessions on a per-form basis (if possible at all)?
Terminology:
I got a little confused as to what is a Repository versus a Service. One comment on the posted answer states "it is ok for the repository to contain business-logic, you can just call it a service in this case". It felt a little useless with our repositories only containing basic CRUD operations when we often wanted to push filtering etc. into the database. So we went ahead and extended the repositories with methods like GetByName or more complex GetAssignmentCandidates. It felt appropiate since the implementations are in the Business Layer but they are still called repositories. Also we went with Controllers for classes interacting directly with UI elements but I think that name is more common in the Web world.
Should our Repositories actually be called Services?
Sorry for the wall of text. Any answers would be greatly appreciated!
Regarding 1:
Yes and no. Yes you would prefer the UI Layer not to be dependent on some specifics of x-layers down. But it isn't. The composition root is just residing in the same assembly, logically it's not the same layer.
Regarding 2:
Limit the usage of the container. Factories (for Sessions,..) are sometimes necessary. Using static should be avoided. Some Frameworks however prevent you from using the ideal design. In that case try to approximate as much as possible.
If you can currently do new FooForm() then you can replace this by DI or a DI Factory (p.Ex. ninject.extensions.Factory). If you have absolutely no control on how a type is instanciated then you'll need to use static to access the kernel like a service locator and then "locate" direct dependencies (while indirect dependencies are injected into direct dependencies by the DI container).
Regarding 3: i think this is somewhat controversial and probably often missunderstood. I don't think it's really that important what you call your classes (of course it is, but consistency across your code base is more important than deciding whether to name them all Repository or Service), what's important is how you design their responsibilities and relationships.
As such i myself prefer to extract filters and stuff in the -Query named classes, each providing exactly one method. But others have other preferences... i think there's been enough blog posts etc. on this topic that there's no use in rehashing this here.
Best practice to implement for situation like yours is to use MVP design pattern. Here its the architecture that i can offer to you.
MyApp.Infrastructure // Base Layer - No reference
MyApp.Models // Domain Layer - Reference to Infrastructure
MyApp.Presenter // Acts like controllers in MVC - Reference to Service, Models,
MyApp.Repository.NH // DAL layer - Reference to Models, Infrastructure
MyApp.Services // BLL Layer - Reference to Repository, Models
MyApp.Services.Cache // Cached BLL Layer(Extremely recommended) - Reference to Services, Models
MyApp.UI.Web.WebForms // UI Layer - Reference to all of layers
I will try to do my best to explain with the example of basic implementation of 'Category' model.
-Infrastructure-
EntityBase.cs
BussinesRule.cs
IEntity.cs
IRepository.cs
-Models-
Categories(Folder)
Category.cs // Implements IEntity and derives from EntityBase
ICategoryRepository.cs // Implements IRepository
-Presenter-
Interfaces
IHomeView.cs // Put every property and methods you need.
ICategoryPresenter.cs
Implementations
CategoryPresenter.cs // Implements ICategoryPresenter
CategoryPresenter(IHomeView view, ICategorySevice categorySevice){
}
-Repository-
Repositories(Folder)
GenricRepository.cs // Implements IRepository
CategoryRepository : Implements ICategoryRepository and derives from GenricRepository
-Services-
Interfaces
ICategorySevice.cs
AddCategory(Category model);
Implementations
CategorySevice.cs // Implements ICategorySevice
CategorySevice(ICategoryRepository categoryRepository ){}
AddCategory(Category model){
// Do staff by ICategoryRepository implementation.
}
-Services.Cache-
// It all depents of your choose.. Radis or Web cache..
-UI.Web.WebForms-
Views - Home(Folder) // Implement a structure like in MVC views.
Index.aspx // Implements IHomeView
Page_Init(){
// Get instance of Presenter
var categoryPresenter = CategoryPresenter(this, new CategorySevice);
}
I'm not sure if i got your question correct, but maybe give you an idea:)
I need to fetch data from an external API, only accessible via VPN.
The development/test machine will not always be able to connect to the VPN.
The desired behaviour is to use two different implementations (one that calls the actual external API and one that acts as the real thing but returns dummy data). Which implementation to use will be configured via a flag in web.config
I've tried the IoC containers StructureMap and Unity and they both did the job but they only seem to be applicable for MVC, I'm looking for a generic solution that also works for web forms. And also, isn't it a bit overkill to use them for this isolated design problem!?
Is there a design pattern or best practice approach for this particular scenario?
IoC / dependency injection sounds like the correct approach, but you don't necessarily need a container for a simple scenario. The key is to have classes that depend on the API reference an interface IAPI, and pass it the actual implementation RealAPI or FakeAPI.
public class SomeClass
{
private readonly IAPI _api;
public SomeClass(IAPI api)
{
_api = api;
}
}
Now you should be able to switch out the implementation easily by passing a different object to MyClass. In theory, when you're using an IoC approach, you should only need to bind the interface to the implementation once, at the top level of the application.
isn't it a bit overkill to use them for this isolated design problem!?
They probably are. Those IoC containers only help you when you wrote loosly coupled code. If you didn't design your classes according to the SOLID principles for instance, those frameworks will probably only be in the way. On the other hand, which developer doesn't want to write loosly coupled code? In other words, IoC container solves a problem you might not have but it's a nice problem to have.
StructureMap and Unity [...] only seem to be applicable for MVC
Those ioc frameworks can be used in any type of application (as long as it is written in loosly coupled way). Some types of applications need a bit more work to plug a framework in, but it's always possible. StructureMap and Unity might only have integration packages for MVC, it's quite easy to use them in ASP.NET Web Forms as well.
Is there a design pattern or best practice approach for this
particular scenario?
What you're looking for is the Proxy pattern and perhaps the circuit breaker pattern.
I'm pondering the design of a C# library, that will have several different high level functions. Of course, those high-level functions will be implemented using the SOLID class design principles as much as possible. As such, there will probably be classes intended for consumers to use directly on a regular basis, and "support classes" that are dependencies of those more common "end user" classes.
The question is, what is the best way to design the library so it is:
DI Agnostic - Although adding basic "support" for one or two of the common DI libraries (StructureMap, Ninject, etc) seems reasonable, I want consumers to be able to use the library with any DI framework.
Non-DI usable - If a consumer of the library is using no DI, the library should still be as easy to use as possible, reducing the amount of work a user has to do to create all these "unimportant" dependencies just to get to the "real" classes they want to use.
My current thinking is to provide a few "DI registration modules" for the common DI libraries (e.g a StructureMap registry, a Ninject module), and a set or Factory classes that are non-DI and contain the coupling to those few factories.
Thoughts?
This is actually simple to do once you understand that DI is about patterns and principles, not technology.
To design the API in a DI Container-agnostic way, follow these general principles:
Program to an interface, not an implementation
This principle is actually a quote (from memory though) from Design Patterns, but it should always be your real goal. DI is just a means to achieve that end.
Apply the Hollywood Principle
The Hollywood Principle in DI terms says: Don't call the DI Container, it'll call you.
Never directly ask for a dependency by calling a container from within your code. Ask for it implicitly by using Constructor Injection.
Use Constructor Injection
When you need a dependency, ask for it statically through the constructor:
public class Service : IService
{
private readonly ISomeDependency dep;
public Service(ISomeDependency dep)
{
if (dep == null)
{
throw new ArgumentNullException("dep");
}
this.dep = dep;
}
public ISomeDependency Dependency
{
get { return this.dep; }
}
}
Notice how the Service class guarantees its invariants. Once an instance is created, the dependency is guaranteed to be available because of the combination of the Guard Clause and the readonly keyword.
Use Abstract Factory if you need a short-lived object
Dependencies injected with Constructor Injection tend to be long-lived, but sometimes you need a short-lived object, or to construct the dependency based on a value known only at run-time.
See this for more information.
Compose only at the Last Responsible Moment
Keep objects decoupled until the very end. Normally, you can wait and wire everything up in the application's entry point. This is called the Composition Root.
More details here:
Where should I do Injection with Ninject 2+ (and how do I arrange my Modules?)
Design - Where should objects be registered when using Windsor
Simplify using a Facade
If you feel that the resulting API becomes too complex for novice users, you can always provide a few Facade classes that encapsulate common dependency combinations.
To provide a flexible Facade with a high degree of discoverability, you could consider providing Fluent Builders. Something like this:
public class MyFacade
{
private IMyDependency dep;
public MyFacade()
{
this.dep = new DefaultDependency();
}
public MyFacade WithDependency(IMyDependency dependency)
{
this.dep = dependency;
return this;
}
public Foo CreateFoo()
{
return new Foo(this.dep);
}
}
This would allow a user to create a default Foo by writing
var foo = new MyFacade().CreateFoo();
It would, however, be very discoverable that it's possible to supply a custom dependency, and you could write
var foo = new MyFacade().WithDependency(new CustomDependency()).CreateFoo();
If you imagine that the MyFacade class encapsulates a lot of different dependencies, I hope it's clear how it would provide proper defaults while still making extensibility discoverable.
FWIW, long after writing this answer, I expanded upon the concepts herein and wrote a longer blog post about DI-Friendly Libraries, and a companion post about DI-Friendly Frameworks.
The term "dependency injection" doesn't specifically have anything to do with an IoC container at all, even though you tend to see them mentioned together. It simply means that instead of writing your code like this:
public class Service
{
public Service()
{
}
public void DoSomething()
{
SqlConnection connection = new SqlConnection("some connection string");
WindowsIdentity identity = WindowsIdentity.GetCurrent();
// Do something with connection and identity variables
}
}
You write it like this:
public class Service
{
public Service(IDbConnection connection, IIdentity identity)
{
this.Connection = connection;
this.Identity = identity;
}
public void DoSomething()
{
// Do something with Connection and Identity properties
}
protected IDbConnection Connection { get; private set; }
protected IIdentity Identity { get; private set; }
}
That is, you do two things when you write your code:
Rely on interfaces instead of classes whenever you think that the implementation might need to be changed;
Instead of creating instances of these interfaces inside a class, pass them as constructor arguments (alternatively, they could be assigned to public properties; the former is constructor injection, the latter is property injection).
None of this presupposes the existence of any DI library, and it doesn't really make the code any more difficult to write without one.
If you're looking for an example of this, look no further than the .NET Framework itself:
List<T> implements IList<T>. If you design your class to use IList<T> (or IEnumerable<T>), you can take advantage of concepts like lazy-loading, as Linq to SQL, Linq to Entities, and NHibernate all do behind the scenes, usually through property injection. Some framework classes actually accept an IList<T> as a constructor argument, such as BindingList<T>, which is used for several data binding features.
Linq to SQL and EF are built entirely around the IDbConnection and related interfaces, which can be passed in via the public constructors. You don't need to use them, though; the default constructors work just fine with a connection string sitting in a configuration file somewhere.
If you ever work on WinForms components you deal with "services", like INameCreationService or IExtenderProviderService. You don't even really know what what the concrete classes are. .NET actually has its own IoC container, IContainer, which gets used for this, and the Component class has a GetService method which is the actual service locator. Of course, nothing prevents you from using any or all of these interfaces without the IContainer or that particular locator. The services themselves are only loosely-coupled with the container.
Contracts in WCF are built entirely around interfaces. The actual concrete service class is usually referenced by name in a configuration file, which is essentially DI. Many people don't realize this but it is entirely possible to swap out this configuration system with another IoC container. Perhaps more interestingly, the service behaviors are all instances of IServiceBehavior which can be added later. Again, you could easily wire this into an IoC container and have it pick the relevant behaviors, but the feature is completely usable without one.
And so on and so forth. You'll find DI all over the place in .NET, it's just that normally it's done so seamlessly that you don't even think of it as DI.
If you want to design your DI-enabled library for maximum usability then the best suggestion is probably to supply your own default IoC implementation using a lightweight container. IContainer is a great choice for this because it's a part of the .NET Framework itself.
EDIT 2015: time has passed, I realize now that this whole thing was a huge mistake. IoC containers are terrible and DI is a very poor way to deal with side effects. Effectively, all of the answers here (and the question itself) are to be avoided. Simply be aware of side effects, separate them from pure code, and everything else either falls into place or is irrelevant and unnecessary complexity.
Original answer follows:
I had to face this same decision while developing SolrNet. I started with the goal of being DI-friendly and container-agnostic, but as I added more and more internal components, the internal factories quickly became unmanageable and the resulting library was inflexible.
I ended up writing my own very simple embedded IoC container while also providing a Windsor facility and a Ninject module. Integrating the library with other containers is just a matter of properly wiring the components, so I could easily integrate it with Autofac, Unity, StructureMap, whatever.
The downside of this is that I lost the ability to just new up the service. I also took a dependency on CommonServiceLocator which I could have avoided (I might refactor it out in the future) to make the embedded container easier to implement.
More details in this blog post.
MassTransit seems to rely on something similar. It has an IObjectBuilder interface which is really CommonServiceLocator's IServiceLocator with a couple more methods, then it implements this for each container, i.e. NinjectObjectBuilder and a regular module/facility, i.e. MassTransitModule. Then it relies on IObjectBuilder to instantiate what it needs. This is a valid approach of course, but personally I don't like it very much since it's actually passing around the container too much, using it as a service locator.
MonoRail implements its own container as well, which implements good old IServiceProvider. This container is used throughout this framework through an interface that exposes well-known services. To get the concrete container, it has a built-in service provider locator. The Windsor facility points this service provider locator to Windsor, making it the selected service provider.
Bottom line: there is no perfect solution. As with any design decision, this issue demands a balance between flexibility, maintainability and convenience.
What I would do is design my library in a DI container agnostic way to limit the dependency on the container as much as possible. This allows to swap out on DI container for another if need be.
Then expose the layer above the DI logic to the users of the library so that they can use whatever framework you chose through your interface. This way they can still use DI functionality that you exposed and they are free to use any other framework for their own purposes.
Allowing the users of the library to plug their own DI framework seems a bit wrong to me as it dramatically increases amount of maintenance. This also then becomes more of a plugin environment than straight DI.
I’m new to Dependency Injection and had a question/need guidance.
I had an application that used the repository pattern for data access. I used StructureMap to get the correct repository and all worked well.
I have since broken out my model (including the repository logic) into its own assembly and added a service layer. In the interest of DI the service layer class takes an IRepository in its constructor. This seems wrong to me as now all consumers of my model need to know about the repository (at least configure their DI to know which one to use). I feel like that is getting into the guts of the model.
What sounds wrong with this?
An application written to use dependency injection typically configures a single container instance where all the interface/implementation type mappings have been registered at an initialization stage of the application. This would include the registration of the repositories, services, and any consumers of the service within the application.
By resolving the consumers of the service through the container, consumers need only indicate their dependency upon the service, not any dependencies the service might need. Therefore, the consumers of the service will not be coupled to its dependencies (e.g. your repository). This is the benefit of doing dependency injection through a container as opposed to doing manual dependency injection.
If you are designing services to be consumed by other applications in the form of a reusable library then your options will vary depending on the level of flexibility you wish to offer.
If you presume all clients of your library will be using dependency injection, then you will need to provide an appropriate amount of documentation about what types need to be registered within their container.
If you presume all clients will be using a specific container (e.g. StructureMap), then you can ease the registration requirements by providing registries which encapsulate all the specific registration needs for the client.
If you wish to allow your library to be used by clients not using their own dependency injection container then you can provide a static factory which returns the service. Depending on the level of complexity, such a scenario may not require use of a container (for example, if your service is comprised by just a few objects in all). If your library is comprised of a substantial amount of components which need to be composed then you might have factories which resolve the services through their own shared internal infrastructure initialization needs.
I understand your dilemma there Dan, I too spent lots of time wrestling over that in my mind. I believe the way I decided to go forward with was one of best ways to encapsulate all of the concerns and still have easily maintainable loosely coupled objects.
I wrote this blog post specifically about NHiberante but if you look at the repository pattern in implement you can easily change the NH specific code to use your backingstore.
Creating a common generic and extensible NHiberate Repository