The other day I was on a conference call with my client desperately trying to convince them to let me start converting their application over to using MonoRail. The current implementation is XML and XSLT and a custom front controller implementation and hand written data access layer full of stored procedures. I put together a comparison of using MonoRail vs. the existing framework and I estimated it would take 10x less code to implement the same features. During the debate they asked me how I would feel about replacing a ton of code I wrote, and I tried to convince them that I would be ecstatic!

The first example that came to mind is Castle ActiveRecord. On my last big project we used regular NHibernate. We had a boat load of mapping files, and we wrote what Dave later extracted as NHibernateRepository. We also implemented our own validation engine which Dave later extracted as EVIL. A week ago while I was working up this example I started using Castle ActiveRecord and it does everything we tried to do, only better. There are no mapping files, you get the full benefits of NHibernate, and it encapsulates all of the logic in the model. Then I started using the built in validations and was again impressed, they do a better job then EVIL. As soon as I got through that example application I realized I would never use NHibernateRepository or EVIL again, and I was excited!

Now, some people would keep using what they wrote. This is called Not Invented Here syndrome. But I would love to use something that is Not Invented Here over something that is Invented Here. (NIH > IH) Why? Well for one Castle ActiveRecord is in my opinion better. Another good reason is that more than one team is working on and using the code. Another good reason is that when I leave the project people will have forums and other developers they can ask for help, etc. etc.

Due to scheduling I lost the debate with my client, but I plan on vigorously renewing it in two months.

-James