I have recently evaluated a few code generators and (Object Relational Model) ORM frameworks in .NET, to my surprise, there were quite a few of them all with there own unique sets of pro’s and con’s and designed for various stages of development.

Why do we need an ORM/Code generator?

The main purpose here is to eliminate writing repetitive and static code; this will lead to a higher quality product, developed in a much shorter time. 

The repetitive/static code I am referring to is CRUD (Create, Read, Update and Delete), CRUD is common and is required by almost all applications, it is repetitive and error prone, Crud requires lots of testing time which can be reduced if code is automatically generated and it can take a lot of time (like weeks) if the database used is huge.

The Evaluation 

As mentioned, there are many ORM/code generators out there, each with there own pro’s and con’s, these are the features I looked for when selecting our ORM.

• Support for hierarchies and relationships,
• Support for grouping and aggregate functions,
• Support for transactions,
• Database independent,
• Concurrency,
• Performance and footprint,
• Learning curve and simplicity with which it can be integrated into code,
• Up to date documentation,
• Online support,
• Reasonable price.

End Result 

 
My evaluation has lead to an ORM and code generator that complement each other well, satisfy all of the above requirements, and has now been implemented as the core foundation of our development framework. 

 
Implementation has been successful and development time and quality of the product has been noticeably improved within the first week of implementation. This has a knock on effect that affects the clients directly, as greatly reduced development time means cheaper pricing and a happier client.

Nadir Kamdar