Our shop designs and create custom software applications for a vareity of vertical industies. We currently use a modified version of the Csla framework for most of our development.
It’s a great framework, supports a vareity of ways to communicate with a database, directly, remoting, WCF and so on. It offers a ton of features, many of which we do not use. The pros of the framework are numerous, the big one being the Rockford Lhotka is a step ahead when it comes to new technology, meaning we don’t have to do the research. The cons of the framework are the fact that you are at the mercy of how the creator implements changes and technology and all the many features that you do not use.
With the advent of Linq-to-Sql we are seriously looking at making to switch, granted a lot of what is generated is purely data access, but by creating partial classes we could extend the data access and provide business logic. We could also create some formal interfaces for working with the business logic. Could use/create our rules manager, and so on. In a nutshell we would be growing our own application framework.
I noticed during Jeff Atwood’s discussion the ASP.NET MVC framework at PDC 2008, he was primarily working with a single project, and it also looked like he was extending Linq-to-Sql with partial classes. This architecture seems to demonstrate the fact that the code is easily maintainable, new features are quickly added and bugs are quickly fixed, and that it performs well … most of the time.
I’m just curious as to what other user’s thoughts are? Am I crazy to abandon our framework for something that I perceive is easier to use and more maintainable?
It would seem that you will be exposed to these same cons with LINQ so keep that in mind when making a change. In any event you should do a complete analysis before making such a leap, perhaps by porting one of the smaller existing apps or a subset of one of the apps as a case study.