Finding a more objective way to identify questions that are too broad

Question

This is an attempt to settle a discussion started by Mourdos almost a year ago. I would like to know what others think of this. There is the full version here but I made some modifications to the proposition, see below.

I believe this could help people write their questions and help users determine if the questions should be left open or closed. Even if the questions fall into these categories, it's always required to have some kind of scope, otherwise it could still be too broad. Overall, this is just a general guideline and it's not because a question doesn't meet these criteria that it should automatically be closed. Hopefully, we might be able to improve it, in order to make it even more useful.

We can classify most questions into categories. Using 3 things to classify them.

• A is the START
• B is the RESULT
• X is the CHANGE or process

Giving these types of questions:

1. I have A and X, give me B
2. I have B and X, give me A
3. I have A and B, give me X

Or in plain English:

1. I have START and CHANGE, give me RESULT : "What happens if I do this?" (Good)
2. I have RESULT and CHANGE, give me START : "How was it before?" (Good)
3. I have START and RESULT, give me CHANGE : "How would I cause this?" (Good)

The other combinations are not mentioned: having only one of these and asking for the other two is always too broad. It's typical of unclear or incomplete questions.

Other possibilities

4. I have either START or RESULT : "Is this plausible?"

   Asking information about something specific can be on-topic as reality-check but can sometimes lack basic research. A lot of questions asking about a good concept but with no tomorrow falls into this category

    

Answer 1

1. I have START and CHANGE, give me RESULT : "What happens if I do this?" (Good)

I think this is where we have most of our problems. "Suppose on Earth the following event had never happened -- what would change?" is usually still too broad. If you're asking this kind of question you probably need to scope that change -- to biology, to societal development (which might need to be further scoped), to technological development... something.

Answer 2

I think the biggest problem is that some of these are our most well voted questions and yet they really should fall into the too broad category. But because they are interesting and well voted we are loathe to close them...not to mention they can be fun to answer.

Scope is what these questions most often need. I enjoy the alternate world scenarios but people both asking and answering need to keep this in mind and the community needs to learn to close these questions even if we like them.

It often seems like we will close a question once, a small edit will be made and the question gets re-opened...but the question is still way way to broad.

TL;DR I guess I am personally torn on what we should do with these questions. When the are well asked and interesting I feel like they should be in scope...when they aren't as interesting I feel like they should be closed. I think the site carries that bias from us users and we are inconsistent.

That inconsistency should be resolved but I am not completely sure if there is an objective way to do that...

Answer 3

Starting with:

Using Vincent's very helpful framework, let's consider what happens when we add constraints:

• A is the START
• B is the RESULT
• X is the CHANGE or process

Giving these types of questions:

I have A and X, give me B

I have B and X, give me A

I have A and B, give me X

Or in plain English:

I have START and CHANGE, give me RESULT : "What happens if I do this?" (Good)

I have RESULT and CHANGE, give me START : "How was it before?" (Good)

I have START and RESULT, give me CHANGE : "How would I cause this?" (Good)

Where Clause

I think all we need to add here is a WHERE (condition) clause that specifies constraints on the scope of the answer. "Within" is another word that could be used in place of "where".

I have START and CHANGE, give me RESULT within CONSTRAINTS on START and CHANGE: "What happens if I do this within this specific context?" (Better)

I have RESULT and CHANGE, give me START within CONSTRAINTS on RESULT and CHANGE: "How was it before?" (Better)

I have START and RESULT, give me CHANGE within CONSTRAINTS on START and RESULT: "How would I cause this?" (Better)

Examples

Vacuum Creatures - Suggested constraints: The OP doesn't state any criteria on which a "best" solution will be chosen such as "lowest energy requirements" or "simplest propulsion system". More details about the creature in general beyond a vague "fungus-like".

Sounds is faster than light - Suggested constraints: a proposed set of the laws of physics in this world, most importantly. Second, since light is fundamental to pretty much everything in how the universe operates, the OP will need to further narrow down the scope of covered warfare to something more restricted than "all known wars have been fought, but the technologies and war tactics used during them were drastically different."

New Power Sources - Suggested constraints: The OP mentions nothing about how the new power source is supposed to behave beyond "more powerful than nuclear" (antimatter is the only thing more powerful than nuclear). There aren't any statements of how hard the physics in this world are.

Discussion

This is exactly like doing a query on a large database. Without any constraints, all rows/answers come back; which in WB speak is "too broad". But with the addition of sufficient constraints, an entire universe of answers resolves down to one or two answers.