By Megan Oster
In January of this year, IBM announced Enterprise COBOL V5.2, the second release of its new COBOL compiler technology. The key difference between the two versions is improvements made to the back end. So what are some details you should know about these changes? Below are some highlights, as outlined by speaker Tom M. Ross of IBM at the 2015 Orlando SHARE conference.
- More use of decimal floating point instructions for PACKED-DECIMAL data. It offers fixed point precision, which makes it a great data type, especially for financial institutions. Now, COBOL can convert to decimal floating point in two instructions – reducing CPU.
- Path length is no longer an accurate way to count instructions. The goal used to be making each path as short as possible. But now, vector instructions can work on many characters at the same time, which means that a longer instructions path is actually preferred. IBM plans to incorporate vector instructions further in the future.
- Features such as ZONEDATA, AMODE 24 support, VLR and MAP ease migration efforts from earlier COBOL compilers.
- Message compiler options alert users of potential coding errors – for example, if a programmer tries to use a period in a code. Periods are problematic because users can delete them with a simple click of the space bar without generating a syntax error. The result? An incorrect algorithm and a nearly untraceable error.
- Access to z/OS JSON services from COBOL, which fulfills the need for generic web services client support. The new z/OS Client Toolkit JSON parser enables COBOL programmers to parse or create a JSON document that interfaces with clients such as mobile. This development is highly valuable, as it externalizes http and https client functions in a user-friendly, generic fashion for users in almost any z/OS environment.
With all of these improvements and more, COBOL V5.2 is already carving out a sound niche for itself in the industry.