Software development has been in constant crisis since 1968, when the designation, 'Software Crisis' was first used, at a series of famous conferences.

The problem acknowledged then was this of the exponential growth of the need for software, and for complexity, in relation with the shortage of resources, and the inadequation of tools.

Two main venues to cure the issue were thought to be: software factories, and software components (modules).

Then came the Internet, and the period ended in the bubble crisis.

For these 40 years, software has been crawling behind hardware advances, poorly performing on orders of magnitude faster and more powerful machines; never been able to achieve decent quality.

The main reason for this failure is the inadequacy of traditional management recipes: a mix of naivete and cynism. The denial that software was different and required novel approaches.

For centuries, the graal had been to manage knowledge, while the postmodern challenge was in fact to manage ignorance.

For centuries, integrity had been granted by adhesion to processes, and submission to pyramidal organisations, driven by leaders. Now it could only be granted by tools. The difference is a new approach to objectivity.

-- MarcGirod - 24 Oct 2011