E.W.Dijkstra Archive: The fruits of misunderstanding (EWD 854)
Some things, it seems, can only happen in the world of computing. The design of a new product is announced and the world is full of high expectations, but, when the product is ready and people start to use it, it fails to meet in a most blatant manner its main objective. It is a failure and one would expect it to recognized as such. Nothing of the sort: the original main objective is kindly forgotten and the vast majority believes the product to be a great success.
You may have another example in mind, I was referring to COBOL, whose main objective was to make the professional programmer superfluous and which is now the main vehicle for 3 out of 4 professional programmers. In the name of "ease of programming", "readability", "understandability" its verbosity was considered a virtue; in practice that same verbosity turned out to be one of it striking defects.
COBOL was not an isolated case. When in the 60's the promise of salvation by means of "higher-level" programming languages did not materialize, the slow response in batch processing was identified as the culprit that made efficient debugging impossible, and in the name of the fast response needed for debugging, "interactive programming" was introduced on a large scale. Its spread seems to have been totally unaffected by the fact that, in the mean time, controlled experiments had demonstrated that the beneficial effect of interactive debugging facilities on "programmer productivity" was at best marginal.
