I see a long fruitful life ahead for COBOL. COBOL remains the best implementation language for business applications and will retain that distinction for the foreseeable future.

COBOL is the dominant language for business applications. Over 80 percent of the world’s business run on COBOL (Gartner Group, ‘97). Almost every major industry from financial applications to manufacturing real-time system relies on COBOL. The investment in COBOL technologies, staff and hardware, is estimated to be greater than 5 trillion dollars (Micro Focus, Sept. ‘97). Over 180 billion lines of COBOL code are in use today, with an estimated 5 million new lines added per year (Gartner Group, ‘97). COBOL applications include:

There is no single language that meets all programming requirements on all platforms. The issue is function and fit. A good solution must provide great performance and fit well with hosting infrastructures. COBOL, CICS and DB2 are good examples. Some solutions while not as good a fit will have a following because of functionality. For example, Visual Basic from Microsoft provides an extensive toolset. The fit, however, is limited to select platforms. 

The solution selection process boils down to "risk" then "functionality." A review of most businesses depict a complex intermingling of applications and implementation vehicles with varying mixes of assembler, COBOL, C, C++, PL/1, Visual Basic and Java, among others. 

In the 80s the SHARE Language Futures Task Force recommended to IBM that they provide more support for multi-language programming. The requirement was based on the typically extensive mainframe inventory of languages. The recommendation addressed issues of interoperability - applications written in different languages should work together seamlessly. The requirement that programming languages work together, and do so efficiently, is truer today than ever. A major difference today is that the host has changed from a single platform to a multi-tier environment.

The IT community wants answers to, "how do I maximize what I have?" and "how do I get from here to there"? They are, in many respects, different sides of the same coin. 

Once the millennium problem is solved the next major issue will be the Internet. The driving force is profit. Business seeks to decrease costs while increasing performance. Unfortunately, hidden costs particularly software maintenance costs can be life changing (and not in a good way). 

Business will approach the Internet in a variety of ways. Some companies will choose to embrace seemingly better functionality over fit by taking the "baby/bathwater" approach, attempting to rewrite legacy code in more "Web-friendly" languages such as C++. Others will prepare by moving to Web-enabled turnkey solutions, at great cost and greater risk. Still others will get from "here" to "there" by leveraging existing code bases and programmer skills.

COBOL continues to change to meet user-demands supporting an ever more diverse application base. Traditional "Pseudo-Conversational" CICS applications continue to exist. Today, they are being complimented with client/server and network applications. COBOL has always played a major role in transactional subsystems and will continue to do so in the future.

COBOL investments are substantial and the decision about what should be done boils down to these options: 

1. Throw the investment away in a staged or "cold turkey" approach.
2. Add functionality to the investment using a GUI builder.
3. Provide a bridge to increased Web-functionality and fit by adding Java capabilities.

Business software has evolved over decades and is seldom considered "throw away". Software applications are more and more being regarded as corporate assets, critical to business success. Software tools must fit within the software development infrastructure and allow for evolutionary rather than revolutionary growth in development and maintenance. "Technology for technology’s sake" has been, and continues to be, IT’s Achilles’ heel. 

The Gartner Group estimates that there are over 3 million COBOL programmers, far exceeding any other programming language. It represents billions of dollars in code, training, and education. The business knowledge of an experienced programming staff represents an even greater asset. Business can access new platforms and leverage existing COBOL skills bringing fit and functionality goals within reach with minimal disruption to existing programming infrastructures.

Tried and true applications that are line-mode, text oriented report-driven work well for running many applications; we have done it for decades and it works. A need for visual environments and data requests from non-traditional users is placing new demands on existing systems. Business needs to maximize information gathering, delivery, and deployment by getting it to the people who need it, when they need it, in the form they need it. 

Taking information to new heights means adding functionality to COBOL. Alternatives for adding functionality to existing systems can be separated into three categories: front-, middle- and back-end improvements.

A new front-end – Front-end platform independent graphical user interfaces can provide access to legacy applications and reduce maintenance and support costs associated with porting legacy applications to new platforms.

1. Create a Windows-only solution with a tool designed to take advantage of   functions and features available in that environment. 
2. Create a hybrid solution with COBOL and HTML/CGI 
3. Create a Java based front-end that relays messages to the terminal application. 
4. Use PERCobol and write once and deploy anywhere.

Server Options

Create server applications capable of supporting multiple diverse clients. This middleware ideally should enable the deployment of applications on any platform.

1. Write  Java applications that can be interfaced with legacy programs through special constructs. Existing legacy applications continue to reside on the mainframe and/or mid-range servers throughout the enterprise. Access to new platforms is achieved with new applications written in Java. Communication with existing server-based applications is done through special APIs. 

2. Write COBOL applications using one of the multi-platform COBOL compilers. These COBOL applications can then be linked to existing COBOL programs. An API or callable interface between programs allows communications between older applications residing on the server and the new applications.

Back-end Options

Back-end (data access, transaction processing and messaging) – Remote COBOL application access to data via CICS, and/or MQSeries messaging.

1. Write "glue" code
2. Use a COBOL compiler that supports these technologies such as PERCobol.