25

If you've read the front page recently, you're well aware that there's a large percentage of total crap on it. Indeed, more than a third of the questions posted in the last 60 days have been closed or deleted. I'm led to believe this is quite a high number among Stack Exchange sites.

This post is about one possible way to reduce the number of of poor quality questions:

Remove the ability of unregistered users to ask questions.

This has been done for some time on SO, Programmers and probably some other sites, with varying outcomes.

A look at our recent statistics shows that 57% of questions posted by unregistered users are "bad" questions. And our resident statistician informs us that making this change could reduce our bad question rate by up to 13%.

But shouldn't we make it as easy as we can to ask questions?

It should be easy to ask a question, but my feeling has been for some time that it is much too easy to ask a question here. We're not meant to be a site which will Google the answers for you or write your configuration files line-by-line or walk you through the finer points of plugging in a network cable.

Requiring registration may discourage poor quality drive-by questions from people who won't even return to improve their question or accept an answer. For those who do register, it gives them a moment in which they may hopefully begin to KEEP CALM and WRITE A BETTER QUESTION. Ideally.

Is this all you've got? This won't have much of an effect.

No, this won't have a large effect on the rate of poor quality questions. But every little bit helps, and I have other ideas which I'll also be proposing in the days ahead.

  • 8
    Before y'all get too excited over the possibility of a 13% reduction in bad questions, read this... – Shog9 Sep 14 '13 at 1:34
  • 3
    Before y'all get too excited over the possibility of a 13% reduction in bad questions, read this too (specifically, see now tiny 13% is). – voretaq7 Sep 14 '13 at 5:20
  • 2
    The 57% actually represents just 4 questions per day. This is dwarfed by the 36 questions per day from registered users. – Iain Sep 14 '13 at 7:16
  • 1
    Does this actually solve the problem on SO however? I see a lot of people there asking bad questions with an account they clearly created just to ask question. Sometimes it looks like the same person is creating multiple accounts so they can ask the same question more than once. I don't know what the solution is, but I don't think requiring an account is effective. I wonder if it should be mandatory to have a non-trivial amount of rep before being allowed to ask a question. Maybe you have to answer questions and get upvoted first, or have a high rep on some other stack exchange site. – Abhi Beckert Sep 15 '13 at 3:19
  • I never said it was a complete solution to all our bad questions. At best it is only a small part of an overall solution. – Michael Hampton Sep 15 '13 at 3:21
  • 1
    @MichaelHampton: Perhaps if you presented the whole set of solutions rathyer than drip feeding them ... – Iain Sep 16 '13 at 16:34
  • @Iain Because it's much easier to make small changes and measure their impact than to make a whole lot of large changes and have no idea what the results are. – Michael Hampton Sep 16 '13 at 16:36
  • 1
    @MichaelHampton: That's not what I asked; but I don't see how you can measure this when it's really noise. – Iain Sep 16 '13 at 16:38
  • 2
    Requiring registration will not change the number of bad questions. You will just be forcing bad users to create bad accounts to ask bad questions. – Cypher Sep 16 '13 at 17:40
  • 2
    Even if 100% of the questions from un-registered users were good questions, there's a stigma associated with being unregistered. Seeing user9324901 screams "I don't really give an ish about your site or your members, just fix my issue." While there may be benefit for future views and a possible accept, it carries the air of a perfunctory question, blasted out to multiple Q&A sites in hopes of an answer. It's human nature to feel more connected to a user here with a "real name" vs. machine generated anonymity. – TheCleaner Sep 16 '13 at 20:28
  • @TheCleaner: Personally I don't care if someone is using a usernnnnnn name or something else but I do care about the quality. Having a usernnnnn name isn't a good indicator of poor quality. – Iain Sep 17 '13 at 11:23
  • I'm not sure if this question or one of the other questions about "the future of SF" is most relevant, but there's an interesting - and very similar - discussion going on on English.SE about dealing with their version of "questions from non-professionals." They went so far as to start another SE site, specifically for people learning English to ask beginner questions, so that only the more advanced questions would stay on the original English.SE Now they're not sure that worked as well as they wanted and are discussing merging the two sites! – Ward Sep 20 '13 at 16:43
  • Here's their meta discussion: meta.english.stackexchange.com/questions/4167/… – Ward Sep 20 '13 at 16:44
  • Let's grant that requiring registration will have a minuscule effect on the quality of questions. On the +ve side, all the evidence from voretaq7 shog9 and michael-hampton is clear: requiring registration does NOT reduce the number of questions. i.e. all the challenges to effectiveness of this solution are also arguments against potential negative impact on "ease of asking questions". I'd say let's implement a registration requirement. But set our expectations right - we won't be doing it for reducing bad question but for other benefits - community, participation, tracking, bans, etc – OC2PS Sep 26 '13 at 14:53
  • @Shog9 I think it's time to un-defer and revisit this. – Michael Hampton Jun 11 '14 at 16:18
27

Yes, you should have to register to ask on SF.

In addition to that, our registration page should have a subset of help/on-topic on it. Then there's no reasonable excuse for people to have no clue.

12

I don't think you should have to register to ask questions.


Frankly the speed bump of registration (especially for the developer types, who probably already have a Stack Overflow account so they can code-by-clipboard for their day jobs) is barely a seam in the road.

I would much rather see an old-school copy-protection style shibboleth based on the about page for unregistered users ("According to https://serverfault.com/about what is the third type of question you should not ask on this site?") -- A correct answer being required in order to post your question.
At least they'd have to read it when they copy and paste it.

  • 8
    If only we could cast the ones that fail into the Chasm of Doom: "What... is your favourite colour?" "Blue! No, Green! Aaaaaaahhhh!!!" – Ward Sep 14 '13 at 5:59
  • 3
    I think your answer means the exact opposite of what you mean to say, because of the double negative in the first sentence. Is your answer in opposition to MDMarra's, or not? – MadHatter Sep 14 '13 at 7:18
  • 6
    I don't not can't understand your first sentence. – Wesley Sep 14 '13 at 18:05
  • 1
    @MadHatter: Bro. Do you even diagram sentences? – Scott Pack Sep 16 '13 at 3:58
  • 1
    What double negative? I see no kitten here. :-) – voretaq7 Sep 16 '13 at 15:26
6

I think requiring registration could help to prevent fly-by questions but only if done like Voretaq said with a q&a for first time users of the site. There's that fine line though between being welcoming to potential long-term experts and keeping the riff-raff out.

Personally, I like the idea of requiring new accounts to be asked a few basic questions (and agreeing to them via checkmarks or T/F or whatever, before allowing their first question to be posted. These questions could be something like the following:

STOP! Before posting your first question you need to answer the following:

  1. There are other SE sites about IT, so please make sure you are on the right one. Did you read the About page? Specifically the sections on what you can and cannot ask on this site?
  2. We don't like fly-by questioners. Do you intend on accepting an answer to your question if it is valid?
  3. SF doesn't like long discussions back and forth. Does your question make sense and does it have examples of what you've already researched or tried?

I also find the site interesting as of the last few years and wonder if others feel the same.

Personally I see a large void "in the middle" of the sysadmin q&a. Perhaps that's because Superuser, Unix, Engineering, and other "sister" sites exist within SE. What I mean is most of the questions seem to fall into the:

check out this crazy, obscure scenario only really relevant to me that Google and others have no answer for

OR omghi2u I am new to IT

OR the I'm from SO and know very little about sysadmin stuff, help! questions.

I don't see a lot of questions/answers that the community has deemed mega-upvote worthy in 2013 vs. years gone by. Has the industry simply matured?

It appears that only 5 of the top 250 "upvoted" questions are from this year, the rest from the past. Have we simply answered 99% of the most common "server admin" questions at this point and the current crop of questions is relegated to either crap, shopping, lack of search, or in currently rare cases legitimate questions that make even the experts here have to pause and think it through or even throw up our hands?

Are we the minority voice hoping to hold onto a site that should embody professionalism and expertise when the rest are wanting a dumping ground of mediocrity and "in case of emergency break glass for simple and quick online IT support", where we are relegated to being the "wise old sages in the Comms Room"?

  • 1
    I wouldn't put a whole lot of stock on the voting being synonymous to 'good' questions. The voting aspect to this site is more of a social experiment than a true measure of quality. There's also the volume aspect to consider. With so many questions being posted, visibility of a question goes from seen to forgotten really quickly, which also affects votes, number of answers, etc. – Cypher Sep 17 '13 at 0:38
  • @Cypher: Serverfault gets just 113 questions/day so they hang around for long enough for a view to be formed on their quality. Contrast this with Stack Overflow which gets 7k questions/day and questions are gone before you can see them. – Iain Sep 17 '13 at 11:26
  • @Cypher I don't put a lot of stock on questions with small votes, but questions with even 10+ votes on this site nowadays is very likely to be a good question. Ward even made the argument a while back that voting is essential to this site's progress/upkeep. – TheCleaner Sep 17 '13 at 12:58
4

This isn't a bad idea, but... I'm skeptical that it'll do much of anything when it comes to decreasing the number of low-quality questions here. It certainly didn't do much for Progse.

As you say, every little bit helps, but in this case it could end up hurting almost as much as it helps.

IMHO, the best reason to turn this on is simply to reduce the amount of confusion and support that unregistered users tend to generate / require. That's not something to ignore - as one of the folks handling support requests, I certainly wouldn't mind seeing fewer of them - but it's also an issue that could be and maybe should be addressed in other ways; if nothing else, it's a separate discussion.

So I'm marking this "deferred" for the time being. Let's try something else first, and see what (if anything) happens as a result. We'll revisit this idea on down the road.

  • 3
    I am very frequently seeing unregistered users who lose their cookies and so lose ownership of their content. I'd turn it on for just that reason alone. – Michael Hampton Sep 25 '13 at 21:38
  • All right, it's time to revisit this. I've just had to kill YET ANOTHER crap question from a guest. It's really getting old. – Michael Hampton Nov 19 '13 at 6:11
3

No, I don't believe ServerFault should stop accepting questions from unregistered users, as I do not believe this will reduce the number of poor quality questions. what it will do is increase the number of abandoned accounts and increase the likely-hood of missing out on some "good" questions.

This post is about one possible way to reduce the number of of poor quality questions

I don't agree. There are already mechanisms in place to handle poor quality questions (votes, flags, etc).

Yes, this is after the fact and you are looking to prevent poor quality questions from being asked in the first place. However, I believe the best you can do is offer some information at the question asking stage (which is already being done) and hope the user cares enough to read it and adhere to it.

Unfortunately, many of these people who are asking these poor questions are real people with real deadlines on real projects who could care less about how clean the SF database is, or how "good" we think their question is. They are stuck. They have a question, and need an answer. If this request is implemented, they will create an account that they will probably forget about in 30 seconds. Then they will post their question. It may get down-voted into oblivion, closed, comment-spammed to hell and back. But they will probably just ask again, until either they find the solution for themselves, or someone posts something that helps them solve their problem.

Then they are gone. Probably never to return. Or maybe they do, and have to create an account so they can ask their new question. But they forgot their old account, or probably forgot they ever had an account, so now they create another account.

Repeat.

While I applaud your effort and wish this was something that could help, I think we're trying to solve a problem that can't be solved.

  • 6
    Whist I agree with you regarding the efficacy of this initiative I disagree with your reasoning. We don't come with an SLA. Real people with real projects and real deadlines should contract with real IT support professionals to get the support that they need. – Iain Sep 17 '13 at 6:05
  • Why would you argue "I wouldn't put a whole lot of stock on the voting being synonymous to 'good' questions." in my answer above and then state "There are already mechanisms in place to handle poor quality questions (votes, flags, etc)."? Are you stating that only downvoting is worthwhile? – TheCleaner Sep 17 '13 at 13:00
  • @Iain, if that is all you got out of my post, then I don't believe I made myself very clear. If I have some time, I'll try to be more explicit. – Cypher Sep 18 '13 at 0:53
  • @TheCleaner, No, certainly not. I believe what I was trying to say was that while highly up-voted questions are usually well written and worthwhile, that's not always the case. Voting on the SE sites tends to follow a combination of popularity of the user, timing, social jest, and quality. [continued] – Cypher Sep 18 '13 at 1:01
  • Then there are those questions that would fall under the good category that just don't seem to receive a whole lot of attention, thus, little voting occurs, and thus, they don't get counted in all these statistical reports that folks like to tout around. In other words, whilst the vote-count is certainly helpful, it's not the end-all-be-all of post quality. I hope that was more clear. :-) – Cypher Sep 18 '13 at 1:01
-8

I am a LAN Admin that also handles help desk. I have very little time and truthfully, one of the reasons I have gravitated towards SF is the fact that I can ask a question quickly without logging in. Don't change your policy. I think, for the most part, the system is working well.

  • You already have an account, so this wouldn't affect you. – Michael Hampton Sep 24 '13 at 15:27
  • 2
    @MichaelHampton - unless he is a bozo and saying that he regularly chooses not to log in and ask questions randomly as "new" users each time. – JM4 Sep 24 '13 at 17:05
  • 9
    I personally don't think that logging in with an OpenID (e.g. my Google account) wastes any time at all. It's practically effortless. As a matter of fact, StackExchange is one of very few online entities that handle logins in a way that I like. – Skyhawk Sep 24 '13 at 19:45
  • I haven't "logged into" Stack Exchange in weeks. I'm almost always already logged in and don't have any need to logout. The only time I have to login is when something crazy happens to my browser that invalidates all the cookies. In short, if I need to ask a question, I just ask... – jmort253 Sep 26 '13 at 5:05

You must log in to answer this question.

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