Many companies are beginning to realize that the rate of mainframer retirement is only increasing with few new experts in the pipeline. While jobs are plentiful, many members of the workforce do not know COBOL or other legacy languages. They are simply not being taught any longer. And yet, legacy languages are the heart of the applications that drive a large portion of many enterprise businesses.
It is easy to try to ignore the problem given the solutions which are most often advocated. In a few cases, legacy functions can be replaced with commercial software, but you give up customization and branding. You could rewrite these applications in modern coding languages, like Java, and run them on servers, but that is high cost, high risk, and potentially could lead to performance issues with formerly well-behaved applications. Some have tried outsourcing, but that is not a panacea.
There is a better way. What if you could make your remaining mainframe staff more efficient? If you could automate a large part of the work they did, allowing them to be more productive as well as making the work more interesting, that would be a much more useful solution than the ones referenced above.
It turns out that many analysts spend much of their time simply reading old code, trying to understand it. This understanding is critical to impact analysis, which is the art and science of understanding and documenting the impact of making a change to the system. There are myriad problems with this approach. You have to understand the code in order to determine what the consequences of a change might be. Legacy code is hard to read, with data structures strewn around and the code underlying business logic chopped up and inserted in various parts of the module. Documentation, if you even have it, is probably out of date. The result is that people spend hours trying to analyze code when there are better things to do.
The solution is to automate this analysis. A great solution would have to have auto-discovery where components and business rules can be detected and documented. You would want a repository of the results that show things like the screens customers see and a list of all system components. Even better, you would like to see the relationships of every component to another. What you want at the end are high-level maps that allow you to drill down through this information so the answer you seek is presented to you.
CM First has just acquired such a solution, evolveIT. This product is in use in a large number of enterprise, mainframe-focused companies by DevOps professionals to speed their analysis of legacy code. Using CM evolveIT, developers are not only able to perform quality impact analysis, they are also modernizing these legacy applications, easily creating web and mobile apps, using the cloud, or even being able to replatform these applications if needed. When time permits, CM evolveIT can also help you eliminate some technical debt and improve code quality. And finally, you will have complete and accurate documentation of your system.
The new capabilities provided by CM evolveIT enhance CM First initiatives around modernization and mobile-enablement. Learn more about the core capabilities here.