Questions with no concrete answer?

Question

Let's say I want to create a robot to follow a line, but make this as fast as possible.

Can I actually ask a question like: "How can I make a line following robot as fast as possible".

I would say that the question is way too broad, but what if I really were wondering this or wanted to discuss about it? And it's not really an answer which has a 'solution' but more a discussion/design question. Can this be asked on SE?

Should I:

• Just ask the question, it's a good question?
• Not bother SE people with design questions, just try it out yourself, if you get stuck, come back here.
• Not bother SE people with design questions, just try it out yourself and ask a question if you have to further optimilize something like: "How to read sensors faster?"
• Divide it in smaller/more basic (and probably design) questions: "Which motor is fastest (servo/DC)?", "How to make a fast (line-follow) robot chassis? (maybe not for arduino SE)" and "How to use 'PID' controlling in arduino"
• Do more research before asking said question

Answer 1

An answer doesn't have to be concrete, but it should be at least a "lightly packed gravel" answer. As for your question, you're walking on the line. If you put a ton of research into your question, figure out most of the parts of the design, and ask a well-written question about the parts you cannot figure out, it's probably not going to be closed. However, if you just add two sentences to the body, it's a lot more likely to be closed. The point is, we cannot tell if your question is too broad without seeing it.

What do I do then? Let's break it down. First, figure out your design objectives. You're never going to achieve the fastest, so figure out how fast is good enough. How much can you spend? How big will this be? Will it have any cargo or other sensors?

Then, let's look at a bit more specific stuff:

• Chassis: Steel? Plastic? Wood? Candy canes? Look at your tools and resources and pick that out. This is essential in how much this will weigh
• Sensor System: Do you have the budget for an array of sensors so you can build better algorithms? How will you mount them? How fast can you sample them?
• Motors: How will you drive them? How will you power them? Servo, stepper, or brushed? Closed or open loop feedback?
• Board: How many things do you need to attach? Power usage constraints? How will you communicate with it? Do you need an ARM board?

You get the point. If you didn't specify enough information, I could tell you to try one of these if you don't specify, and I doubt that you want to attach something explosive to your Arduino! :)

The thing is, design questions are really hard to answer and they take a long time to get answered. If you put the time into your question, people will be a lot more willing to answer and you're a lot less likely to get your question closed. Furthermore, you get more specific answers that're more useful to you and you get a sense of pride when you realize that you designed most of this without any help. Just remember: we're here to help you along the way, not do everything for you.