The Stripy Strudel's Journal

My recent IT Allegory is easy to misintepret in two ways. Or, to put it that way, to conceive two ideas out of it that I didn't intend to convey, and that I don't consider correct in the general case. Specifically:

1. It is wrong to develop tools that implement fundamental generalizations. One should make small and simple tools to solve the given problem, and no more than that.
2. It is wrong to make small and simple tools to solve the given problem, and no more than that. One should develop tools that implement fundamental generalizations.

I don't deem either of these a universally correct approach, and it's not to demonstrate the advantages of one approach over the other that I wrote my article for. I wrote it to convey a different idea:

• It is wrong to adapt a system universal for one class of problems to solve problems of another class. One loses the advantages from the universality and gets the overhead from the overall solution complexity.

This is something that not many would argue about, at least it's much less questionable than the other statement comprising my point of view:

• For every class of problems, one should undertake a detailed analysis and then develop a system implementing the fundamental generalization for this particular area.

A corollary is that development under strong time pressure harms the product quality while possibly (in the case of a commercial product) contributing to its better marketing success. По-русски: IT-аллегория: послесловие