Working on bugs: hurdles to bug squashing

Monday August 9, 2010, Will Kahn-Greene | share this (mastodon)

Note: This is an old post in a blog with a lot of posts over a long span of time. The world has changed, technologies have changed, and I've changed. It's likely this is out of date, the code doesn't work, the ideas haven't aged well, or the ideas were terrible to begin with. Let me know if you think this is something that needs updating.

There are two big hurdles to get through when starting on a bug before I'm in a position where I can fix it. I think the bulk of the time spent on most bugs is in getting through these two hurdles. If we could somehow reduce the amount of time and energy I spend on getting through the two hurdles, we can increase the speed that I can fix bugs.

The first hurdle: what is the problem?

I think most people confuse the question "what is the problem" with "what's causing the problem". These are two distinct questions. The first one can only be answered by whoever reported the problem and this hurdle is a communication hurdle. I spend a lot of time trying to coach people into explaining the problem they have in a way that's coherent and complete.

It's not uncommon to get bug reports along the lines of "something's wrong and Miro doesn't work". I'm psyched someone has taken the time to report the issue, but there isn't enough detail here for me to do anything. In order for me to do something with this bug, I spend time walking the reporter through giving me enough details such that I have a good idea as to what the problem is. Because this requires communicating back and forth, this can take anywhere from a few hours to a few weeks to transpire.

Because it takes so long and requires me to ask 20 questions, I'm pretty sure most bug reporters find this phase very frustrating. Some question and answer volleys are really frustrating for me, too. I want to help--I just don't know what the problem is.

Assuming we get past that hurdle, then the second hurdle comes into play....

The second hurdle: what is the cause of the problem?

This hurdle is sometimes difficult because I can't reproduce the problem. I may not have the right context (e.g. "Problem XYZ happens to all people in Ecuador") or the right equipment (e.g. "Problem XYZ happens to people using Windows XP SP2 with video card Foo"). I'll read through code and work with the reporter to try to figure out the cause. Sometimes it works and sometimes it doesn't. When it doesn't it's really frustrating.

Most bugs aren't like that, though. I'm able to either reproduce the context and equipment or I'm able to use a reasonable facsimile. Then figuring out the cause is entirely on my shoulders and I can work through it like I do most things: code spelunking, research, Google searches, talking with co-workers, talking with upstream developers, ....

How can you help?

The following helps a lot:

use an email address that you respond to
be as specific as you can in your bug report description
attach your Miro log file
be patient with me when I ask my 20 questions to figure out what the problem is
be prompt with responses--the longer it takes to turn around a question and answer, the longer it takes to get the bug fixed
thank people who help get your bug fixed: other people who comment on the bug providing additional details, people who worked on the bug, people who tested the bug fix--it takes a community to build an application

Guess what....

This isn't specific to me or Miro! This is true of all FLOSS applications that you use: we're all working hard to build the best applications ever. Doing the above helps everyone.

Want to comment? Send an email to willkg at bluesock dot org. Include the url for the blog entry in your comment so I have some context as to what you're talking about.