Throw Away Your Bug Tracking System | Hacker News Reader

Throw Away Your Bug Tracking System

wiki.apidesign.org

27 pointslocopati21 comments

Kind of inflammatory post that boils down to one reason--there's no point in having a bug tracking system if you're not going to address the bugs.

I disagree with this article. Especially if you're open source, you gain a lot by having a bug tracking system:

1. It's another way to get user feedback. Are people excitedly opening lots of new enhancements, or frustratedly reporting lots of crashes? Building community is extremely important.

2. It's a way to MEASURE the priority of bugs. Are 30% of your users running into this bug and piling on with "me too" comments? Then you'd better fix it. If you keep closing bugs because you dont want to fix them, you wont see that feedback.

3. It's a way to organize your own thoughts and priorities. Is the new feature X that you want REALLY more important than fixing the existing functionality? How broken are things currently?

There's nothing wrong with leaving a bug open and making the comment that "it will be a while before we fix this" I'm more likely to submit a patch in that case, as a user.

Basically, this article is about optimizing the developer experience rather than the user experience. This seems to be a problem endemic in open source software anyway, which is why so few people use it.

"Basically, this article is about optimizing the developer experience rather than the user experience."

Did you ever use Bugzilla? Nothing is more detrimental to the user experience...

Users would rather send support e-mails or use phone support. For the lack of that in open source project, keep bug trackers simple. E.g. Github's bugtracker is quite ok, it is fine for writing a bugreport and reacting to it. But doesn't have all the bells and whistles that makes it complicated for users.

The article is not specifically about bugzilla, but issue trackers in general.

1. So few people use open source?

2. Closed source is better at optimizing for user experience?

> So few people use open source?

Compared to closed source? As a function of time spent interacting? Yes.

> Closed source is better at optimizing for user experience?

Manifestly. Firefox is the only real counterexample I can think of (where development is conducted in the open, not where source code dumps are made available).

As a user I find the way some of those bug tracking systems are used infuriating. This seems to happen to me often:

  - You submit a bug or find one that's active and comment on it
  - there is no activity
  - then vendor posts needsinfo
  - info posted soon after
  - then ignored for while longer
  - reassigned
  - ignored for while longer
  - then closed because release is obsolete, need retesting in new release

This is mostly in RedHat's bug tracker, but elsewhere also.

My personal favorites are effective game-breakers for branch-of-users-X, possibly more, which should take minimal effort to fix (unless the architecture is horribly screwed up), that get 100+ submissions as duplicates, reduced test cases, discussions, pasted diffs hoping someone, please God anyone with commit privileges will look... and they get ignored for years.

Looks like in cases like that the bug tracker is making the underlying problems more visible. In that case, removing the bug tracker is not the way forward.

Oh most definitely, it's far better than not having one. I just think that bugtrackers are a bit broken themselves.

Why not help out triaging and fixing bugs then?

At points I've submitted patches or very clear descriptions but they get ignored still. In other projects there is often a steep learning curve to get even a small patch in, like in Mozilla, particularly if it's a language you don't normally use. I think in most projects it's simply not feasible to drop in to do a couple of patches you really have to involved long term.

The only exception to this are security bugs where I've avoided the bug tracker and typically got fast responses and fixes.

OK, join the project and put the patches in yourself.

http://fedoraproject.org/wiki/PackageMaintainers/Join

OK, how do I join RedHat Enterprise? which is what we actually use and pay money for therefore where I find the bugs.

I probably should join Mozilla, Ubuntu, KDE and Gnome too - that's going to be a lot of catching up to do just to fix one bug in each.

Can you see what I'm saying? feedback is valuable to those projects but they are ignoring it, if the only way people can get a bug fixed is by spending 10s or 100s of hours getting involved, there is something wrong.

If you contact Red Hat support you'll get immediate help. If you have a support contract you shouldn't be filing bugs in Bugzilla, since the support team will do that for you, and chase the developers to get things fixed. Your support contract will make all this clear.

That's not been my experience, it's like in bugzilla but your oscillating between "waiting on customer" and "waiting on redhat" and the people your talking to don't really understand the problems and so have to ask someone else everytime.

There was an issue in Django's tracker that someone wrote a patch for, and a few of us volunteered to rewrite if given some direction. It's still waiting for a design decision, though.

I've seen this happen too often.

As a user I'd rather see a bug tracked even if it's not addressed. Sure, it's frustrating to find a bug that's received no attention, but having the bug tracked in a central location is one level of feedback better than nothing at all.

As an end user, it's very helpful for me when a bug has been tracked even if it's not addressed. When I run into some kind of weird behavior, I can just check the bug tracker. If someone else has already reported the bug, it saves me a significant amount of troubleshooting.

It's also pretty typical that users will post workarounds and patches in a bug report. As you might imagine, this is hugely useful on a number of levels. It makes it easier for project maintainers to fix the bug in question, and end users with the problem are able to quickly resolve the issue.

This came to mind: http://www.jwz.org/doc/cadt.html

But that said, I feel for this guy. Big backlogs with low priority enhancements masquerading as bugs are no fun.

There are some pretty big open source projects running without bug trackers, they just keep an eye on the mailing list, and if a problem keeps cropping up, it might get handled.

Don't call it a bug tracking system. Call it a todo list.

Reply on news.ycombinator.com