There's a letter in the current Datamation so wrong-headed, and so weakly answered that I can't bear to let it slip into the oblivion it so richly deserves. One Stephen Wright of Applied Data Research, Princeton, writes "Surely 1984 should see at least the beginning of the end of stored-program computers... Self-modifying programs are known to be harmful... Mixing of programs and data leads to abominations like VS..."

He was writing about a very good predictive article by Ted Withington, and the comment appended by the latter was so wishy-washy as to be almost as incredible as the letter itself.

If there is one idea beyond the original Babbage dream of the computer that is central, that is absolutely valuable and powerful, that is certain to be perpetuated as long as data is processed and models are built and our messy human culture is operated, it is the stored-program concept. To envisage, let alone approve of, a computer future without it is like thinking of education without writing or printing or of communication without voice links.

What Wright is telling us is that he and his overweening peers are now so clever, so farsighted that they can prescribe character sets and codes, data structures operation sets, storage hierarchies and the details of application programs for the entire future, world without end, amen!

I've been in this nutty racket and its preelectronic forerunner for almost 40 years, and I find its pluralism, its incredible variety, its twists and turns so challenging and so unexpected that I can only see into tommorrow by relying on the eternal verities: pride, greed, the lust for power - and, of course, stupidity. Ever since the hardware boys began to string tubes and resistors together, there have been software insolents who dreamed of remodeling the machines. They range from the simple souls who want everybody to work in APL, or advocate business data processing in Algol, to the mercenaries who are making it possible for manufacturers like Burroughs to seal off their bit-manipulating power from the users and make higher level languages compulsory.

We absolutely must retain access to the fundamental power of the computer. That means no locked-on integument of wasteful software, whether language or operating system. That means retaining the possibility of programming in machine language.

That means insisting on all the options. And that means not denigrating, but glorifying, in the stored-program idea.