CSLA 3.6: Object Factory intro
We have been working the past few days on getting the new ObjectFactory up and running with CSLA 3.6. The approach we are taking is that we really want to abstract the Data Access Layer away from the Business Object so that we have some flexibility in the future to change the way we access the database. At the current time we are planning on using NHibernate as our ORM tool, one especially nice feature is that it is able to load your objects automatically so no more LoadProperty(SomeProperty) or ReadProperty(SomProperty) a million times. This one thing alone makes this very appealing to me.
 |
This diagram above is a representation of the basic architecture of the system. What we are basically doing is using the Repository pattern to keep data access decoupled from the business object through the use of the ObjectFactory and StructureMap. This is a setup very similar to what Frank is using in his project. However we are not using DTO objects in between the Repository and the business object. We might change this later but for now we will let the ORM load the objects directly.
I am currently working on implementing the ProjectTracker from the CSLA examples. My goal with this is to accomplish several things:
- Learn the ins and outs of the ObjectFactory and how it works with StructureMap, NHibernate, Mock Objects, TDD, etc..
- Provide a working drop in replacement to the “stock” ProjectTracker.Library, there shouldn’t be any changes required to the UI or anywhere else.
- Provide the source code so others can learn from what has been done and make contributions to make it work better or smoother and have it provide a model for people to base their frameworks off of.
So stay tuned because I will be adding more information here as I get closer, should be real soon. I will probably setup the source code on Google code or codeplex, if I do that let me know if you want to make contributions to the project.