In "Creativity as an exact science" Altshuller presented his theory of the S-curve as follows: "Life of technical systems (as well as other ones, biological, for example) can be plotted as an S-curve, which shows how main characteristics of a system (its power, performance, speed, the number of copies manufactured, etc.) change over time. ... In "the childhood" a technical system evolves slowly. Then comes the time of growing up and maturing when the system improves fast and is applied in mass numbers. Beginning some point the speed of development starts decreasing, old age begins."
It is obvious that in this theory two different things are confused and mixed: change in the speed of a system's development and change in the values of the system's main characteristics. It is true that initially system changes slowly, then its change accelerates, then it stops accelerating, then it decelerates and eventually comes to stagnation. Thus, what really evolves along the S-curve is only THE SPEED OF A SYSTEM'S CHANGE. All other system characteristics may or may not evolve along the S-curve !
For example, the cost of a technical system is also one of its main characteristics. However it does not evolve along the S-curve ! In the case of computers it even drastically dropped (rather than grew) as their development accelerated !
Characteristics/parameters of many other systems also do not always exibit any inclination to evolve along the S-curve !
It is worth mentioning that Altshuller's theory of the S-curve was never properly presented in the TRIZ-journal. Instead of "main characteristics" its authors talk about either "performance" or "functional capability" or "ideality" or "aggregated value". According to Michael Slocum articles system's performance evolves along the S-curve. According to Ellen Domb, "functional capability" of a system evolves along the S-curve. According to Darrell Mann, "the aggregated value" of a system evolves along the S-curve. According to Zlotin, the degree of ideality of a system evolves along the S-curve.
Although "performance" is a pretty vague term and Slocum means by it whatever he wants, he cannot be absolutely arbitrary and choose a completely irrelevant parameter. Thus, his approach is risky since he can be caught by the arm on a particular example. In this regard, Darrell Mann's, Domb's and Zlotin's approaches are safer. "Functional capability" and "the aggregated value" are absolutely vague terms and "the degree of ideality" is not calculable 1 to avoid any responsibility. Has anybody ever calculated, for example, the degree of ideality of a car to speak about evolution of this parameter along the S-curve ? Nobody ! Is it at all possible to calculate the degree of ideality of a car ? Not at all !
It is better to speak about quality instead. Although it is also hard to calculate quality uniquely, some "approximations" can be proposed (as if something that does not have a unique value can be approximated !). In the case of cars, one of such "approximations" is
Quality = (Payload x Distance x Speed)/(Cost of car x Cost of gas x Cost of maintenance)
Indeed, the heavier payload car can take (e.g. the more passengers) - the better; the farther it can deliver it - the better; at the higher speed it can transport it - the better; the higher cost of the car - the worse; the more gas spent on transportation - the worse; the more maintenance is required per the distance - the worse 2.
There could be other parameters bearing on the quality. Such as comfort, for example. Or parameters mentioned in the publication of the National Highway Traffic Safety Administration. Or parameters mentioned in the publication of the National Center for Education Statistics.
Although quality cannot be calculated uniquely, whatever way you calculate it, in many cases it does not evolve along the S-curve !
Altshuller never collected and never published any data supporting his concept of the S-curve. He never plotted the dynamics of any characteristic of any system. He simply did not have data for it. He was a fantast. And all his theory was a fantasy or educated guess at best !
 See my paper "On the new contribution of Michael Slocum to the concept of ideality" in Vol. 2 No. 5 of the journal to learn why ideality is not calculable.
 Please note that unlike in formula of ideality, in formula of quality characteristics are multiplied rather than added.