<< Is COBOL Dead? or will there be COBOL code on starship Enterprise? | Home | Business Rules in Legacy Modernization >>


COBOL Re-host vs COBOL Translation (to Legible Java or C#)

(Lets hope re-hosters keep their test routines, they'll be needed again)

Recently there has been a lot of activity in the Re-hosting world, as evident in quarterly reports of MicroFocus as well as the big rehosting companies.

But why would anyone rehost as oppose to translate? lets look at the pros and cons of each solution ...

Re-hosting solutions “at best” are non-intrusive: not a single line of code needs to be changed. For example, recompile IBM COBOL in MicroFocus, get rid of the Mainframe, and run on a Unix box. Or compile the COBOL into Java byte-code and run it on any other box. Keep everything in COBOL and don’t need to re-train all the developers (well, the ones that haven’t retired yet).

But then, many of the re-hosted solutions will require lots of code changes. To the best of my knowledge, re-hosting a Unisys COBOL (DMS and DMS-II) or HP COBOL with TurboImage in MicroFocus COBOL cannot be done non-intrusively, and will require a lot of code changes.

Intrusive or not, the re-hosted system has to be tested. And here lies the problem. Testing will be the longest phase of the whole project. Often there will be no existing test cases, and the test cases will need to be  re-developed in due course of the project.

Lets recap:

  • Rehosting is at best quick (no code changes), but has the same “testing” requirements as translation into a new language such as Java (provided it is a maintainable translation).
  • Rehosting and Re-translation projects will have the same ball-park cost figures – since the biggest influence on project cost is the long “Testing” phase.
  • Rehosting will reduce the running costs by changing platform (same as translation)
  • Rehosting will reduce the need for retraining existing development team. But most of the COBOL developers that I know have long since learned newer languages such as Java and C#.

On the last point, recently an organization told us that they have 18 COBOL developers and some 250 java developers. The re-hosting appetite will be diminishing in the next few years.

Let us hope the organizations which undertook re-hosting have not mis-placed the test scripts.

Add a comment Send a TrackBack