Always bet on the Machine – the fate of software development
by
Charlie Bess
If you look back forty years ago, there were hundreds of thousands of telephone operators in the US working for AT&T alone. Now the function has been almost totally automated. Before the 1950s, "computer" was a job code (usually doing actuarial work), not a device that would be recognized by most individuals.
More recently, we've seen the shift from shirts and shoes being made by hand, to today, where they are made by machines. Even more critical and precise, activities like eye surgery and Prostatectomy are being turned over to robotics.
I believe there are changes taking place in the IT space that will push the envelope of what people can perform without assistance. Some of these are:
The move to multi-core and specialized processors (the age of abundance in computing). It is very difficult for people to write parallel code. There is quite a bit of work taking place in this area. Just as robotics help "good" surgeons do "great" work, the average programmer can use automated assistance in the assembly of great programs.
The move of the edge of the enterprise out into more finely grained data elements will require more and more interfaces to be created. Automated techniques will amplify the capabilities of the available resources to meet those new interface needs.
Modeling and simulation will increase the confidence in the experiments and changes to business processes. With the use of techniques like genetic algorithms, they'll even aid in optimizing the performance of business models. Just like in the NASA antenna design competition, they'll likely make adjustments we'd never dream of.
Although there is a Luddite tendency in us all, we must realize that betting on the machine over the long haul is a sure thing. I've blogged before (a few years back) about the ever increasing capabilities of computers and the inevitable outcome.
