Quote:
Originally Posted by bEdhEd
Andy, I don't take your post as a bash, so please don't take mine as a bash either.
|
Haha, no worries, I don't. I was just a little worried that the discussion would turn into a flame war before the _real_ discussion started, so I was a bit defensive (plus I'd just finished shoveling after the snowpacolypse so...) . Reading the discussion now (and holy cow it's exploded over the last day) I'm really stoked with where this discussion is going.
I'm seeing lots and lots of great opinions, and even a sort of consensus about COTS parts.
As I Read further, I'm starting to see my viewpoint change a bit too with the way people frame the issue.
This is one of my favorite replies so far
Quote:
Originally Posted by JesseK
In developing new products, I simply hope that suppliers do not encourage sacrificing resiliency in problem solving for convenience in executing a solution. Students who are taught that mindset are terrible Engineers, expecting where to be told a solution and the complaining when it doesn't "just work by pressing the button". "Business is Business" is a cop-out when it goes that far.
|
And I think that's the center of my worry too. Problem solving is important and you can't start problem solving by trying to design an airliner, you've got to start small. Where better to start that process than at the lowest (reasonble
) level possible?
For example, to jump into software-land. We're using Java now, and Java has a hozillion great libraries for everything under the sun, INCLUDING lots of fantastic FRC open-source projects that we could leverage. This is a LOT like COTS mech/electrical parts.
I let my students use these libraries, even if (and when) they do find them on their own, but under one condition. I get them to understand the concepts and reasons WHY that library existed.
NAVx MXP is a great example; No one is going to expect every student to understand sensor fusion, and kalman filtering, on top of a robust I/O protocol. That's Crazytown. I'm sure some particularly bright students might be able to get it. I don't expect them to reverse-engineer anything, but I do help them understand what a Filter is, and why it's important.
A yearly project for new students is to build a simple complementary filter that eats gyro data & a single magnetometer and produce a smoothed heading. This is a simple (mathematically too) project that helps them understand just what's happening under the hood. Once that project is done, the concepts are understood, I give the go-ahead and they pull in the NAVX libraries and navigate away! No re-designing the wheel, the guys at kauailabs did a fantastic job, better than we can expect to in 6 weeks, so by all means lets leverage that.
That is how I feel about larger COTS Parts. If you use them, that's great, just make sure the backup knowledge is there, the why, the question to the answer.
Thanks again guys for the fun discussion