Tool-maker supports an approach to modernizing legacy applications it calls "extend and surround" that advances the idea of eating the elephant one bite at a time.
Application modernization is a process that has been ongoing for some time. Even though the momentum has picked up pace recently, there are companies that have been working on this challenge for awhile and have a depth of experience. Two new white papers from tool developer mrc can help clarify the issues a company must face when considering modernization alternatives and help companies select which one of several specific approaches it may wish to take on this journey toward application modernization.
mrc, which stands for michaels, ross & cole, ltd., has a Java-based Rapid Application Development (RAD) tool called m-Power that is popular in the IBM midrange market and has roots in mrc's former 4GL development environment for the AS/400. With m-Power, even business users can create new applications, while seasoned RPG developers can create Java apps without suffering the steep learning curve some experienced a few years back when trying to acquire Java skills. The mrc tool uses prebuilt objects and templates that link to the Web server inherent in the AS/400 or IBM i server.
Apart from having released a new m-Power WYSIWYG screen designer called m-Painter last spring, MRC now has an application or "project exchange" where developers can upload their m-Power templates, external objects, and User-Defined Functions (UDFs—specific programming routes) and share them with other developers. The project exchange likely will greatly enhance the value of m-Power to existing customers if enough people contribute to it and could give m-Power an edge against competing tools. So far, there are 13 project contributions, and the list is likely to grow as even non-m-Power users are invited to contribute to the exchange.
Needless to say, if a company is interested only in a screen-scraping approach to application modernization, then m-Power is not the tool for them. What mrc advocates is an approach to application modernization that is more akin to redevelopment, one it calls "extend and surround." The eat-an-elephant-one-bite-at-a-time approach involves gradually surrounding your old applications with completely new and modern ones. The idea is to modernize different parts of your applications gradually as the need arises without touching other aspects. As one project is complete, you move on to focus on other areas that also need to be modernized.
The approach, says mrc, allows a company to modernize gradually, at a pace it can afford, but have a truly modern set of Web-based applications in the end. Of course, they share the same logic and database as do the old applications. There is no question the approach will take time for the development of these new applications and likely will require having a developer on staff familiar with modern methods and languages, but the task can be greatly simplified by using a tool such as m-Power, the company says. The approach does take much of the risk out of the undertaking compared to summarily rewriting all applications at once and can make the project more affordable for many companies that might be put off by a large dollar estimate.
In its white paper, "Crash Course in Modernization," mrc identifies several basic approaches. While it begins with screen-scraping, the company claims this is not really modernization at all because the underlying application hasn't been changed. While for some this could be the method's greatest advantage, mrc is not enthusiastic about screen-scraping. It does acknowledge, however, that often the IT department may be under so much pressure from upper management to get rid of the firm's text-based interfaces that it's a low-cost and attractive upgrade. However, the company points out that the underlying legacy applications, which more than likely are procedurally based, often cannot fulfill the changing needs of the business indefinitely and may increase maintenance since anytime the core application is changed, the graphical screens and links will also have to be updated.
The second approach is converting an application written in one language to the same application written in another modern programming language. This is done by running the code through a code converter and turning the applications into, say, Java. The solution may solve the problem of becoming dependent upon a dwindling talent pool to maintain old applications but the fact that a company still has legacy applications designed prior to the introduction of e-commerce will eventually be a limiting factor.
For those companies with beaucoup resources, the "rip-and-replace" method, whereby you build completely new applications using modern programming languages, methods, and databases etc., leaves you with applications that address your current and future business needs. They also are likely to be much easier to maintain as they will be more modular. They will be able also to support your growing business. However, throwing out all your old applications does pose significant risks. How much disruption can your organization afford? You will also need a cadre of loyal and skilled developers who know your business or have the ability to train new developers on how the company's business applications must function.
The final approach is to buy out-of-the-box applications that hopefully will work fairly closely along the lines of your business or can be customized to your specific business needs by your staff, a consultant, or the ISV from whence it came. The devil is in the details, however, and it's in the customizing that such projects get expensive, not to mention that many prebuilt application packages are expensive to begin with. Of course, you can always go with an open-source solution, but the customization required for that may also get pricey, and the eventual cost will be hard to estimate.
As Jon Paris of Partner400 says, "There is no silver bullet," but the folks at m-Power have a long history of helping people modernize their legacy applications, and the extend-and-surround approach makes good sense for many companies that don't have buckets of cash to throw at the problem. Whether written in J2EE or RPG, I personally like the idea of ending up with solid, Web-enabled, event-driven applications that support specific business functions. It may take a little longer to get there, but at the end of the journey, you have a modern, quality application that (hopefully) works. You haven't deferred dealing with the underlying problem, and you haven't shifted the task onto someone else's shoulders as you gingerly head off to accept your next career opportunity.
as/400, os/400, iseries, system i, i5/os, ibm i, power systems, 6.1, 7.1, V7,
LATEST COMMENTS
MC Press Online