cost factor of refactoring to the Spring Framework

[mwinter]

Hello everybody,

I’m currently working on my masters thesis. I’m working on the cost factors of refactoring an EJB application to an application, which uses the Spring Framework and Hibernate. My main goal is to determine which properties of the application and the used techniques have the biggest impact on the refactoring costs. Although my focus is on the costs of refactoring an EJB application to an application that uses the Spring Framework and Hibernate, I’m also interested in more general experiences. I would appreciate it if anyone could provide me with information!

My questions are:
What are your experiences with integrating new technology in applications?

Which technique(s) did you apply to what kind of application?

Which techniques did the application use before and after the refactoring?

Which parts of the application had to be adjusted?

Which part of the application took most of the time to adjust?

What kinds of problems did you encounter?

Which part(s) of adjusting the application took you longer then expected?

What were the main cost factors of applying the new technology to the application?

Thanks already!

With regards,

Martin Winter

[katentim]

Haven't addressed all our questions, but hope this helps...

I’m working on the cost factors of refactoring an EJB application to an application, which uses the Spring Framework and Hibernate

You probably want to separate discussion on different EJBs, and when it actually makes sense regardless of cost. For example you can almost always do without Entity beans, but if you use multi-threaded asynchronous JMS messages consumption, you'd probably stick with EJB MDBs (unless ou can wait for Spring 1.3).

My main goal is to determine which properties of the application and the used techniques have the biggest impact on the refactoring costs.

Size of the application. If you have a monolith that works, you'd have a hard time justifying a refactor.