Question Ban: Should it be permanent?

Question

Currently, new users who repeatedly ask extremely low quality questions, as defined by various hidden metrics, eventually get a "question ban". They can still use the site, but they are not permitted to ask questions.

This is enforced at the IP level, and is permanent.

The argument has generally been that the threshold of the hidden metrics are so low that it only catches the absolute worst of the worst. The people who cannot be educated as to how to ask questions properly or to do even basic research. I believe Jeff is on record as saying that they simply do not want such people here.

I find that there are a few problems with this.

The first is purely technological: an IP address does not uniquely represent a human. Every time there is a perma-question-ban, there is a tiny chance that someone else adopts that IP address in the future, comes to the site, and is told that their question is not wanted, despite never having been here. Something as simple as moving into a new house can do that.

Yes, it hasn't happened yet. But every perma-question-ban increases the chance of it just a bit. As the number of these bans increase, it will eventually happen. Maybe Stack Exchange will have been overtaken by some new technology that makes it even easier to get questions answered. Or whatever. But given things as they are, there is the small chance that someone will be question banned through no fault of their own.

That is, admittedly, a minor point. A far more significant point is this: people change.

Generally, low-quality questions are asked for one of the following reasons:

1. The person is rather immature. They don't really know any better and aren't interested in being corrected by others. This typically afflicts the young, though admittedly there are some surprisingly immature netcitizens out there.
2. The person is ignorant. They don't know enough about what they're talking about to be able to ask actual questions from the community. Them getting question-banned generally requires repeated attempts to engage the community in their ignorance, so some degree of immaturity or pig-headedness is often a factor.
3. The person is simply a buffoon.
4. The person isn't a native English speaker.

The thing is, #1 and #2 are all solvable. Not by us of course; we still have standards and a website to maintain. But these are all solvable, as with many things, simply by the passage of time.

Consider the immature asker. Most of the time, this is a young person, perhaps 13-15 years of age, who picked up some programming book and maybe did a few tutorials online. They have a little knowledge, and we all know how dangerous a little knowledge can be. Thus, they encounter a problem and come here to drop a question on us.

A bad question. Maybe a lot of bad questions.

But here's the thing: that 13-year-old is going to grow up. Sometime around the 17-20-year-old range, they start to understand more about the subject matter. They understand more about how to ask a question, how to do research on their own, etc. They are now capable of asking real, viable questions.

But they can't. All thanks to a couple of incidents from years ago that they can barely remember (assuming their IP address hasn't changed, of course).

My point is that people who ask bad questions don't always stay that way. People grow up, learn more, and learn to speak English better.

What would happen if we went from permanent question ban to, say, 1 year? Would it negatively impact the value of the ban?

I don't think so. Or if so, then not by much. If a bad user is question banned from the site, then they're gone. They leave; generally, question banned users don't hang around. And odds are, they're not going to be sitting there, counting the days until their ban from that site they asked 3 questions on a year ago expires. They got their hint and are gone, probably to some other community.

The only people who would return are those who either were directed here by a search or were interested in joining back up again for other reasons. If a year has passed, it is now possible that such persons have reformed on their own. And it is possible that they have not.

Let's say that 25% of the people who were question banned return after one year and ask again. And of those 25% of people, half of them actually ask reasonable questions. In which case, the question-ban-for-a-year system will only be 12.5% less effective than the current system.

To be fair, both of these numbers are invented out of nothing. But putting a time limit on the question ban would have the effect of allowing a person (who manages to hold onto their IP address) to mature and become a productive asker of the community.

So the question is this: should we allow for the possibility of a former bad asker to mature on their own, by imposing a time limit to the question ban? Would making the question ban finite (but still lengthy) weaken the system too much, or are we missing something without it.

Answer 1

If they start providing awesome answers the question ban will automatically be removed.

Having looked at a lot of these users odds of a mythical reform approach 0.

Also, the IP level bans only affect new users, established users are not impacted by this.

Answer 2

I think a time restriction is needed. And 1 year is still too long.

To my opinion, a ban on IP level is not effective. I can tell only for Germany, but here you will only have a permanent IP as bigger company or organization (or if you are willing to pay for it). And I am not sure if the problematic users are coming from there.

Additionally my company has only one or two IP addresses and uses NAT and there are about 100 developers, I know at least one other colleague that is a SO user (a good one so I am not afraid that our IP will be banned). But what will happen if the company will get some day, e.g. an inexperienced student, who manages to get banned on SO from his workplace?

Most other (private) users, who are customers of the big internet providers, will get, at least every 24 hours, a new IP address.

So, for pure technical reasons, I would say don't ban IP addresses (at least not for ever).

Answer 3

The issue with the IP addresses possibly representing another person at some point in the future is a weak argument:

1. Take the number of people on the internet
2. Then take the fraction of those people who have IPs coinciding with IPs blocked by Stack Exchange
3. Then take the fraction of those people who are interested in the information Stack Exchange has to offer
4. Then take the fraction of those people who also try to ask a question
5. Now compare that number to the number of questions asked per day on Stack Overflow

I'm willing to bet the farm that the final number is exceedingly small: so small it's a largely theoretical concern that doesn't add any weight to the argument.

But in all honesty, Stack Overflow doesn't need more questions: it needs more answers. And this theoretical population isn't blocked from providing those.

Which brings us to your other argument, on providing a way out for people who made some bad choices early on in life and are now potentially functioning members of the Q&A community.

Putting aside extreme cases where perma-bans are actually permanent and unfixable (I mention it only because I'm sure SE has such a thing and I don't want to make a universal claim, not because that's what we're talking about), when a person is blocked from asking any more questions via the automatic system, they can get out of it on their own: they just need to start providing high-quality answers.

If a person is reachable and they just had a bad run at first, they will almost certainly dig their way out of the question ban hole after a few high-quality answers. The deeper the hole, the longer it'll take, but it will happen.

Of course, there's the flipside to this: if they provide low-quality answers, or never bother to switch to answer-mode, they'll never be able to ask questions again.

That's fine: being a user on Stack Exchange is a social contract. People are supposed to pay it forward: someone who is incapable or unwilling to provide useful answers in exchange for getting their questions answered isn't someone we want here. If they change their mind or find the ability, they of course can come back into the fold.

If they don't, that's fine too: they'll find some other community that'll allow them to leech off their experts.

Answer 4

I'm fairly sure a permanent ban can be lifted via an email to the team. This would allow the non-buffoon to demonstrate their newly found non-buffoonery, or the new occupant of a house to explain what an idiot the previous resident was. Nothing is really permanent in computer land.

Answer 5

The ban can be lifted by giving legitimate answers. I've been banned before, and after answering some questions and getting upvotes, the ban was lifted.

But banning an IP is a pain, if you have several programmers working in an office who all use meta, the IP ban would punish those who didn't need the ban.

So, should people have to make more reputation in stackoverflow in order to ask questions on meta? IP bans are issued so that people can't just make new accounts to ask questions, but... what if the requirement of reputation earned in stackoverflow (the reputation required to use meta) was increased so that even if you made a new account, you had to work before you could use meta again?

That solution would sort of be a punishment to newer users who haven't done anything wrong, but I see it as a lesser evil. Slapping around IP bans is something that I think shouldn't be done.

Answer 6

Another issue that arises with IP banning (and it has happened before) is when a corporate IP gets banned, affecting more people than just the buffoon. It doesn't happen very often though (it is a similar problem to when some users find they cannot ask any more questions in a certain timeframe).

In addition to a set amount of time, personally I would throw another metric into the mix before unbanning them: they must have answered x number of questions and received a minimum of y number of upvotes (and possibly z number of accepted answers). This metric shows not only an increase in knowledge (and maturity), it also promotes a better and more balanced use of the site.