News

Comparison of Migrations in .NET and in GeneXus

By Enrique Almeida, in the Developing from the Trench blog.



In Case Study: Migrating a VB6 Large Application to .NET I read the following about the migration of an application with 950,000 code lines to .NET from an ERP: (words in bold are my own additions)

The entire ERP application was migrated in 9 months by 3 developers totaling “3,650 developer-hours to migrate the code, 3,400 hours for code review and refactoring, and 1,300 hours for testing.” The code review was requested because the code would need further development in the future and none of the original developers were available. Total cost: 750,000 Euros, much less than the required one for a customized ERP. The process evolved in phases: when a module was done, it would be integrated with the rest of the VB application until the entire application would have been migrated to .NET.

I always enjoy articles with figures, because they enable real life comparisons within our industry.

In 2004 we underwent a "similar" migration, with a slightly larger ERP (with a few more lines of code), from GeneXus Visual FoxPro to GeneXus Java.

The change implied a significant amount of work because working on the screen with Visual FoxPro was different from Java, and certain rules were triggered at different times.

In all, 14 of us worked on it (some of us with a lesser participation) for 5 months, totaling 2,300 hours of work (including tests and the solution of errors from reprogramming) at a cost of approximately USD 23,500.

Classifications of the tasks performed cannot be subject to comparison because they differ, but I believe they can be illustrative anyway. Our figures were:

Task
Hours 
Programming
1595
Error fixing 
112
Testing
420
Meetings
 25
Installations/Documentation

75
Other
 67
Total
2294

Despite the fact that our migration from GeneXus/Visual FoxPro to GeneXus/Java was a big job, costs, as well as the number of hours and the time taken by the project were less than what it took to the group of European developers. And this is something else to bear in mind when comparing costs of maintaining, for lengthy periods, applications that imply high maintenance costs.

Just in case anyone needs further details, the case study has been published in the Microsoft site.