On the lesser point of testing due-diligence, of you have precisely defined the point of diminishing returns in micro-miniaturization. When discrete components continue to reduce in size, beyond the ability to probe with the point of a needle, design layouts naturally make the micro-leap to SoC's as you pointed out. But testing proper functionality of SoC's (to achieve the desired result is either a theoretical analysis, or an empirical effort requiring actual 'build & sample' efforts.
Cabe, regarding your main point, getting the pre-design requirements and specification right BEFORE Design & Build efforts are launched, I only say 'Amen'. This seems like an obvious thing, but most times is still missed because of complexities stretching the hard-line schedule. How many times have we heard the Program Manger direct that "we're slipping – we can address that later",,,,
Great article - it just takes getting management to buy in so that it becomes corporate culture in a world where speed to market is often an overriding factor. If you can prove this:
"Catching errors early will save 10 to 1,000 times the money to solve. Due diligence ahead of time is key."
I worked for a company were we did company-wide QIT training. It focused on doing things right the first time and this certainly falls within that paradigm. While initially we had more expenditure and time costs, it did prove that in the long run it saved money and increased the profit margin as a result.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.