14

I just found this in the review queue: Does it matter that wlan0 has changed its name to eth1 after I upgraded to 12.04? It has a couple of votes to close as off-topic, I assume because somebody posted a link to this bug report and they thought:

I can see the word Launchpad on the page!
It must be a bug!
Let's close this now!

I'm all for closing flat-out bugs as bugs but there are a few problems with this bug report:

  • It's a really bad bug report. The report blames udev because the user found a udev configuration file. People who know what udev does will know it just parrots what it's told to do. The reason udev is calling this interface eth1 instead of wlan0 is because the Kernel told it to.

  • The problem actually lies in the gradual change in Broadcom drivers. As somebody who was doing a lot of kernel compiling around the 12.04 release, I've seen how violently the naming convention has flipped around but anybody upgrading with the right hardware probably would have seen this too.

  • The problem isn't actually a problem. No applications I've seen have given two hoots whether or not wireless hardware was called eth*.

And that's all the answer needed to say. "It's a driver quirk but it shouldn't give you problems".

"But Oli, not everybody knows the sordid history of Broadcom drivers like you do!"

Sure, but everybody voting has the ability to read that bug thread. By comment #4 you can see it might not be udev. The mere fact it's "confirmed" means nothing. That just means more than a couple of people say they see the same behaviour. The plural of anecdote is not data.

And it's clear to see there's no actual resolution. There's no diagnosis churn and there's no indication whether this is something to worry about. It doesn't answer the question at hand.

A person came to us asking if they had a problem and we're telling them to shut up and watch a dead bug report. That's not good enough.

If you don't have the technical je-ne-sais-quoi to make determinations like these, perhaps you shouldn't be voting either way. That's fine. We don't expect everybody to know everything, just skip it.

If you think you do, please make sure you're reviewing bug reports as well as the questions.

And [almost] lastly, just because the answer is "Look at this bug report", it doesn't mean the question is off-topic or a bug report.

If it's a known bug, a good answer linking to the bug, explaining the problems (and answering sub-questions) is better than voting to close as a bug report. It's more helpful to the user and it'll be better for attracting future searchers with the same issue (and therefore stopping dupes).

Edit: "But we've always closed bug reports..."

Yes and no. Historically we closed bug reports. Issues where people would come onto the site and say "Firefox doesn't work" as "not a real question" because they weren't questions. Those sort of shouty posts are still fair game. Tell them to go and file a bug report in the right place using the OT reason.

What I have been talking about here (and to a degree what Braiam has distilled into just a few words in his answer) is some people are targeting "things suspected of being bug behaviours" and are trying to get them off the site.

Well over 90% of all our questions here are in some way related to a bug of some sort. A design flaw, missing drivers, PEBCAK-through-poor-design-or-documentation. If we followed this and closed everything that fell into the gaping chasm of suspected bugs, we'd have twelve questions left and half as many users.

By all means push people to Launchpad, just don't forget we're a community support site. We fulfil our goals by helping people, not having the most conditioned answer rate.

And this also helps us. It's not just about attracting people here, it's resolving their problems. If they have a problem and Google brings them to an OT-Bug question simply, what do they do next? Is there a LP bug attached to it? Do they know what that means? Does that even help (cite this one... which doesn't). If that process hasn't answered the problem to some degree, they're going to ask somewhere. It doesn't matter where but that's the moment where we've just failed.

An answer with a documented explanation wins. Explains the problem. Explains where the problem is being handled. Answers periphery questions.

11
  • 2
    "But Oli, not everybody knows the sordid history of Broadcom drivers like you do!" - Things I am proud to not have to deal with! Dec 16, 2013 at 15:45
  • So.. What are you trying to say? You seem to contradict yourself in the first and last paragraphs.. Who closing things that are only "suspected" bugs? We have been closing things against bug reports (or things, like segfaults, that need to be reported). I understand maybe more people should read those bug reports to double make sure they are relevant to the question, but no one (AFAIK) is closing things because they "smell" like a bug. And why do you think we're doing this because of "answer rate"? This isn't helping the answer rate..
    – Seth
    Dec 29, 2013 at 16:27
  • 1
    And if a question is a bug, there isn't anything we can do but link them the bug report and say "Click the effects me too button". If there is a good workaround we could maybe post it as an answer, but that's rather counterproductive for the way this site works. Maybe we need another canonical question about how bugs work?
    – Seth
    Dec 29, 2013 at 16:28
  • 1
    @Seth I'm trying to draw the distinction between shouty reports (that we used to close as NARQ) and people who ask a question that describes a problem that might be a bug. If the answer to something is "Oh that's this bug...", it should be answered, not closed. (And by answer rate, I think I meant unanswered rate - which closing/deleting/etc does help)
    – Oli Mod
    Dec 29, 2013 at 17:30
  • So if I find a question that describes a bug so that the OP is probably experiencing aforesaid bug I should answer it with a link to the bug report? That's not constructive, not to mention a link only answer (in most cases).
    – Seth
    Dec 29, 2013 at 17:32
  • @Seth If you don't think it answers the question at hand but is relevant, post it as a comment. And at no point have I said just post the URL as an answer. There is plenty of scope to explain the problem back to the OP and explain how it might be fixed.
    – Oli Mod
    Dec 29, 2013 at 17:34
  • So.. I should explain how this bug got created and how the devs will probably fix it?
    – Seth
    Dec 29, 2013 at 17:35
  • If it helps answer the OP's question, yeah. My example here (while a little grumpier than I'd expect you'd need to be) covers the cause of the problem that underscores the bug and more than that, answers their question.
    – Oli Mod
    Dec 29, 2013 at 17:38
  • @Oli but that's a question. Most of these "bugs" aren't questions, they're "This stahped working, how to fix it!" and the truth is it's broken, no one knows when it will get fixed. I understand that we should probably be more careful when closing bugs, but the vast majority should be closed as they are not questions and cannot be reasonably answered.
    – Seth
    Dec 29, 2013 at 17:55
  • @Seth Indeed. I'm all for holding fast on the reports, they're bad posts... But there are a lot of problems related to bugs (help diagnosing, working around, other options, etc) that do ask questions. Like this one.
    – Oli Mod
    Dec 29, 2013 at 18:00
  • It seems like we've finally gotten to a place where we can agree. I wouldn't have closed that question. He's not asking "why did this change" he's asking "Is this change OK" which some bug report does not answer. So yeah.
    – Seth
    Dec 29, 2013 at 18:06

1 Answer 1

3

Ok, my mistake. I corrected it, and I repent about my behavior. Maybe it was late at night, and I just jumped the gun (should stop visiting closing stuff late at night and without coffee). Luckily, nobody died. But I have a problem with this:

If it's a known bug, a good answer linking to the bug, explaining the problems (and answering sub-questions) is better than voting to close as a bug report. It's more helpful to the user and it'll be better for attracting future searchers with the same issue (and therefore stopping dupes).

Here in meta we have discussed ad absurdum that bugs, things suspected of being bug behaviours (even if there isn't confirmed bug), and asking for workarounds for known bugs are off topic. Why? Because they are relevant only while the bug is not fixed. I think that's the last agreement the community reached.

Now, I don't see why we need to attract more visitors (in fact I think we need to atract more answerers) that will see that we allow questions and answers about an specific bug behavior. It poisonous! It shouldn't be the kind of traffic we should be attracting. Our traffic should be:

— "I have a question"
(looks in google)
— "I found the answer in Ask Ubuntu, nice, lets see in what I can help this guys out" and/or "well I didn't found the question, but something similar in Ask Ubuntu, let me ask my own"

As can be seen, the last show the user what we are about, and give an example what kind of questions should be asked. If they see we allow questions about an specific bug, embrace yourself, because we will not stop receiving bug reports/behaviours kind of question.

Is the broken window problem! If we allow one, there will be more. If we close them, users will be snarky about why we closed his if X wasn't closed (I really want to close some "list" kind of questions too), calling us elitist and what not. Ignoring this fact will keep plundering the site quality, we lose good answerers, questions don't get answers, people don't get the reputation to access the review queues and stop the crap, and the cycle repeats until SE decides to intervene, or the site just lose its function.

So, I do not care if a question gets a good answer, if the solution should come from the developers fixing their stuff (note, if the developers offers a solution, ie. install this for my hardware to work, is ok), that question will be closed.

4
  • 4
    I've just edited to account for some of this but in short, I think your definitions are dangerously encompassing. I'm also not sure I'd agree that the community agrees on where the line exists for bugs on AU. I think we all agree that reports are OT but there are a hundred different views on possibly-bug-related questions. Many of them "official".
    – Oli Mod
    Dec 16, 2013 at 14:04
  • 1
    @Oli I don't see what is the problem with being "encompassing", at least with two of the definitions that could apply to the context. Also, I beg to differ in the last statement. There isn't a single "official" reason to why bug problems (reproducible errors) should be accepted. Please, point to previous relevant discussions about this.
    – Braiam
    Dec 23, 2013 at 22:03
  • I think this answer is well intentioned, but I would suggest that if a significant number of questions are bug reports in another form, then askubuntu needs to adapt to that so that some questions can be converted into, or at least link with, bug reports.
    – Thufir
    Oct 4, 2014 at 2:12
  • @Thufir the problem with those kind of question is that their relevance is fleeting. It would be useful for a month max and then forgotten forever. In SE those kind of questions are normally closed and deleted.
    – Braiam
    Oct 4, 2014 at 12:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .