COTS is not a panacea.
Ask the US Military about parts obsolescence or security and quickly the issues COTS creates will appear.
http://www.militaryaerospace.com/blo...t-go-away.html
https://www.cigital.com/papers/download/ses.pdf
Now the FIRST challenge of a project too big in a time too small really does have benefit from COTS. It allows under-resourced teams to deliver on something they couldn't otherwise deliver without some serious pain.
On the other hand - personally I think it comes a bad time in their careers.
When one can continuously get financial resources from other people and direct them towards COTS vendors instead of learning the more base principals from the moment they were first old enough, and responsible enough, to understand those principals I think we might be robbing some of the students, in the long term, the value of tactile experience (even if it's hard work) early on.
To this concept let's theorize in direct relation to Dean Kamen himself. Would Dean have had the opportunity to found FIRST if his work did not show unique determination early on to acquire the resources to materialize on his ideas?
I won't deny that when I was younger I certainly used TTL chips which are COTS digital circuits and still maintain a large surplus of them. However they were inexpensive and you had to work to build something from them. The knowledge I acquired from learning how to integrate their functions together was invaluable when moving into programmable logic. While I certainly do not miss hours bread boarding and wire-wrapping circuits - I do often see the price people pay for not having that experience. Simple things to me like: delay lines, are confounding to a new generation of people who never saw a race state up close and personal. So they connect macro cells together so many years later (college and later) than when I started working with TTL at age 8 and can't understand why the resulting responses are unstable.
I think sometimes that in order to make FIRST appear ever more impressive we are trading the illusion of hard experience for the quick delivery of something that looks cool using COTS. Something that justifies more investment into it because it looks cool. Something that might not be delivering at the educational level what the casual onlooker might be thinking it is delivering.
I consider it something very much like 'my kid is a computer genius' syndrome. Where every generation looks at their own personal reference for 'genius' and assumes that their child reached the equivalent proficiency without out realizing that they are bootstrapped on the COTS of the last people who did it. Yes your child was able to write a web based accounting system. On what amounts to a supercomputer from when I was a child with a language that would be appalling inefficient on what was the practical computer when I was a child. The math is not that much more complicated and the protocols and languages were given to them basically for free. I'll issue a challenge here: anyone can write their own protocol for an IP network using UDP. In the financial industry at any place where latency is the determining factor between success and failure writing non-TCP protocols is often the tool of choice. Yet in FIRST we often seem to run away from UDP screaming because TCP 'just works' and you 'don't have to do that work'. This makes TCP basically COTS. Yes using TCP saves the user from the effort to make their work reliable on an average IP network. At the price of the user very likely not actually understanding how it even works - however FIRST is not exactly the average IP network - so what you have here is a perfect example of not understanding why the easy way may not actually be the best way.
It is very cool to watch - but lots of people own cars today - and lots of people can't change a tire properly. So the proposition becomes we expose everyone to the robotics technology like cars and hope that this makes more shining stars because it is accessible. However it is accessible within basically 4, 6 week build seasons and after that - as an adult - you now have to pay for continuing access often at a dear price (see rising college debt). Now I do see this is where the Maker community helps. Where we network our skills and resources together at a reasonable price to break this high cost consequence to not getting the basics in sooner (I spent many hours at NextFAB taking a vast number of their classes and I can clearly see the very wide gap in knowledge when people approach a subject in those required classes). However there is still a large social gap between the two. I see where coming back from being a student to a mentor can help as well. However there are many topics on ChiefDelphi as to the risk of coming back as a mentor and people often can't because the demands of college on their resources are very high.
This is a devil's due. We are advancing our primary cause of FIRST with COTS but we may not be advancing the educational goal one might casually think we are advancing.