[tomg]

> For example, if someone asked: "can you unbatch those 1k RPCs?"

"Can we do _________?" is the worst question to ask me as an engineer. Like, yes, we probably _can_ do literally almost anything, with enough hours and money. But what's the budget? What really is _______?

Eg: If it's "analytics to a page", what do we want to track? Where does the data go? How do people view the data? I need to know the answer to those questions, plus the "budget", before even entertaining a yes or no answer. Which I think sounds like I'm beating around the bush by asking, but I'm not! :)

[avisser]

One thing I love at my current job is people embracing the X/Y question terms. People will frequently ask on Slack, "How do I do X?" and immediately follow up with "I'm asking because in implementing process Y, it seems like I need an X."

Helps so much getting to the root of the issue.

[spydum]

omg yes. state the underlying goal to achieve and detail how you intend to go about it and seek help. i dont know how many problems i helped solve only to find out the underlying problem was already solved in another way.

ive learned to ask that as my first follow up question for random requests for help: is this the real problem or are you trying to solve a different higher level problem

[theshrike79]

This problem is as old as time and was really prevalent in IRC channels relating to software a few decades ago.

Some friendly individual put up a site to explain the issue to newbies: https://xyproblem.info

[nerdawson]

Yup, questions like those are a trap. A direct answer is going to lead to being asked to build some backwards solution to a problem.

My response to that type of question tends to be along the lines of "what are you hoping to achieve?"