The problem is, there are very few Frank Lloyd Wrights, fewer wealthy patrons with the vision to give you free reign and, as yet, no equivalent of the architectural drawing for specifying your software design.

The architect can calculate--or in most cases, just consult a pre-calculated table to obtain--the size of his beams and know they will be sufficient for the task.

The software architect can have no such confidence. As yet, we have no way to accurately predict the loads we must cater for, and no way to quantify them if we settle upon an estimation.

Be wary of any suggestion that Project Management is a stepping stone to either job security or job satisfaction. Project management in software requires much the same set of skills as project management in any other field. It is about enforcing the whims of customers and upper management upon the daily lives of the do-ers, and insulating each from the other. A good filing system, time management, people skills, and a thick skin, are infinitely more important and useful than any amount of familiarity with Knuth or Turing.

It may be a stepping stone to upper management for the lucky few that excel at it, but if your joy is in programming, realise that you will rapidly leave it behind.

I agree that the future is likely to be gluing together canned solutions, especially as the demand for software increases, but most of these bread & butter jobs will rapidly move off-shore to countries with lower costs of living. The project management jobs will rapidly follow them.

However, the canned solutions will still need to be devised and written, and these R&D type jobs are where the money will be and these will tend to remain in the western societies for a while to come. The number of these R&D type positions is likely to remain pretty static whilst the rest of the software industry grows, and the people that will get these jobs are those that take inovative and unconventional views of the problems to be solved and are open to exploring uncharted waters.