A pretty picture. For more pretty pictures, see here.

The S-curve in business development

This is a story about two isolated takes on the same subject: mechanical engineering simulation. These two takes either cannot or will not acknowledge each other. I act in the belief that one day they will embrace. My favorite metaphor is that of Ford Model T. After years of denial and mistakes, Ford and the other automakers ended up acknowledging that they were in the same business.

As a newcomer, I am in a position to build bridges and promote this development instead of fighting it. I believe that many others will eventually see a benefit in doing the same. However, I do not expect you to eat the elephant in one piece. Therefore, I intend to write a number of articles and put them here. And now for the S-curve:

Many sources note that as a function of time,

  • adoption of new technology (also called diffusion of innovations)
  • revenue from a certain business operation
  • usefulness of whatever you have learnt as a person

as well as other properties follow the same stages of Ferment-Takeoff-Maturity illustrated below:

To the basic S-curve plot, I have added my estimate of how the two major approaches to mechanical engineering simulation should be positioned. If I am right, the open source approach will experience big bangs for the buck in the near future, while the license-bound approach will achieve less and less for each dollar spent.

Similar events have happened many times before in business history. Time will show if the license-bound world will embrace the open source world smoothly like Microsoft did when their time was up. Or if the big vendors will go the IBM way and ignore the new market conditions as long as possible.

Whatever happens during the next years, I consider this a safe bet: The license-bound vendors will during the transition period charge their long-time, faithful customers similarly to the way the established phone companies have charged their long-time, faithful customers.

For a more light-hearted take on the above issue, look here.

See, the point is this:

When it comes to big time number crunching, hardware capacity available to ordinary people has increased dramatically during the last decade. At any particular instance in time, the license fee will be proportional to the amount of numbers crunched. But as more numbers are crunched, the intellectual property (IP) content of each additional numeric operation will decrease towards zero.

The upward surge of open source software is in part due to the fact that when simple formulae are repeated again and again on a large array of CPUs, the undeniable y-axis advantage of today's big vendors is of limited value. Their performance advantage mainly resides in a more coherent user interface, a generally better documentation and an established user base.

This means that when planning the next number crunching project, many engineers and their sponsors may see an advantage in going from licensed software to open source software. Doing so, they will probably spend more money on harnessing the necessary software tools (calling me should do the trick). But the daily overhead of license fees will disappear.

If you - as a rule of thumb - expect software license costs associated with a given capacity scale-up to be twice as high as the associated hardware costs, open source software may cut your marginal upgrade costs down to one third!

Possibly with better scripting capabilities, too...

Creative Commons License
Website for the defunct company Simxon by Kim Ravn-Jensen is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Permissions beyond the scope of this license may be available at kim@ravn-jensen.dk.