A few weeks ago Cobol celebrated its 25th birthday and the committee that has guided its development looked forward to 25 more years, during which it will become widely used on micros, the committee members predict.

These two books demonstrate the tide is already turning that way. They also take different starting points, reflecting the likelihood that micro owners who take to Cobol will be split between new users and members of existing data processing departments.

Kevin Sullivan presents his book as a 'first read' for anybody wanting to look the language over before committing themselves, while Norman Stang aims particularly at Cobol programmers from mainframe environments.

Hence Mr. Sullivan illustrates his text with programs developed and tested on a Tandy Model III with Ryan-McFarland's RMCobol, and the micro environment is taken for granted. Mr. Stang devotes an early chapter to the micro system, using CP/M as a 'typical operating system'.

The structures of the books also differ; a Cobol program runs from Identification Division to Stop with the undeviating monotony of a sacred ritual but books about Cobol apparently offer more scope. But the subject matter is the same and it's treated in the same fashion. Mr. Sullivan's book, thanks to the typeface, looks livelier but this is a superficial distinction. Perhaps when you're trying to explain a Cobol Data Division there is only one way of going about it.

Mr. Sullivan's book is better endowed with appendices and the text is generally easier on the eye. Mr. Stang's scores on sample programs. Both go to some trouble to explain why Cobol should find a niche at the micro end of the business - which of them you prefer will probably depend on your own familiarity with computing.

One final proviso - you should take with a pinch of salt Mr. Sullivan's assertion that his book can be used without a machine to practise on. Joseph Cooper's performances at the dummy keyboard are very impressive but he'd surely use a real piano if there was one handy.

DG