Wikipedia:Disabling edits by unregistered users and stricter registration requirement

This is a proposed policy change by user:OneGuy
Similar proposals have been made in the past. You may want to read the discussions there:
meta:Anonymous users should not be allowed to edit articles
Wikipedia:Disabling edits by unregistered users and stricter registration requirement (vote archive)

Proposal

edit

Only registered users may edit, except at some sort of special section. Registration would be by ISP-based email, or any e-mail service where someone cannot create hundreds of accounts (i.e. a Hotmail, Yahoo, Gmail, etc). People who don't have a valid non-free email should be able to talk to admins and get an account by some other form of verification, or if just admins are satisfied by examining IP history that it's not a sockpuppet, banned user, or a known troll.

Justification

edit

The reason is the inability of admins/arbitrators to deal effectively with trolls, POV pushers, and spammers. As an example, 168.209.97.34 is a transparent proxy for the largest ISP in South Africa. If, say, everyone agrees that for many months someone using that IP has been violating wiki policies of personal attacks, non-NPOV, and 3 revert rules, there is no effective way to deal with the abuser without affecting all users who use that IP. If admins/abitrators block that IP, they block many other users. This is also a problem in other cases when IPs are assigned automatically by ISP. They will have to block a range of IPs to block one user. Since this place is growing rapidly, soon wikipedia will have thousands of trolls, trouble makes, POV pushers, violators of wiki policies, advertisers, spammers, etc. They will annoy enough good editors, who will see this place as a waste of time, that most of them would leave.

Common objections

edit

People like anonymity, they will not register

edit

Registering with a user ID makes a user more anonymous than not registering. If someone edits without an ID, their IP can be seen by everyone. That is less anonymous than being a registered user. If someone knows the user IP, they knows the user's location and ISP, university, or employer. It's also possible to get hacked if the user has open ports. At least anyone can ping bomb/flood the user, slowing down their computer if not knocking the user off completely, especially if the hacker has a fast connection like T1 and the user is on dial up. In other words, you become anonymous to most people on wiki once you register, not when you edit unregistered.

People prefer free emails like hotmail because they like anonymity

edit

Google/Hotmail/Yahoo emails do not make you anonymous. If you send an email via yahoo etc, the IP is added into the headers by these free services. Anyone can find your location by looking at the headers. If you register using a real email, your email would remain hidden to everyone (including admins) except a very few developers.

Not everyone has email other than these free Google/Hotmail/Yahoo emails

edit

Exceptions can be made for people who don't have a valid non-free email if they talk to admins, or post in the special section, and if admins are satisfied by examining IP history or some other form of verification that it's not a sockpuppet or a known trouble user.

Many unregistered users contribute valuably to wiki, why stop them?

edit

If registration is made a requirement that doesn't mean wiki will lose these contributors. Most likely these people would register and continue to contribute. They edit without registering because they can. That doesn't mean they would not contribute if they would have to register, especially if it is explained that registration makes them more anonymous, as explained above. Most people would still contribute by registering, but even if wiki loses a few good editors, the benefits outweigh the harms. Far more valuable editors would leave this place if they see it as a useless site infested with trolls, POV pushers, personal insults, and helpless admins/arbitratos who cannot block people because blocking dynamic IPs could potentially block everyone who is using that ISP, university, public library, internet cafe, etc. As an example, in early 90s on USENET alt.sex.* newsgroups used to have real people contributing. Once the spam became a problem, these groups died real quick. Everyone left except the spammers. Some other newsgroups went through this later. Once the trolling and personal insults became common, almost all valuable posters left. Potentially without stricter policies, this is most likely the future of wiki, a place filled with trolls, POV pusher, people debating each other endlessly, and helpless admins/arbitrators overwhelmed by all these trolls. Blocking a range of all these IPs would not be a valid solution then. These IPs cannot be blocked for longer than a day or two because it would effect thousands of other users. Requiring a valid email is by far better and effective way to deal with the problem.

Forcing registration means it won't be Wikipedia anymore

edit

Everyone would still be able to contribute. The changes would only help dealing with trolls and trouble users who discourage most good editors, potentially driving them away. This change would not stop anyone from contributing.

If someone is bent on trolling, he will find others ways

edit

Yes, but it would reduce the problems significantly. It would be far easier to deal with a few persistent trouble users who will try to seek "other ways."

You only encourage the trolls by making it a challenge for them

edit

Not doing anything won't make them "leave." The spammers on USETNET didn't stop spaming just because there was no one to stop them. Some of the best newsgroups with least flame wars are moderated groups. Dealing with the problem doesn't usually make the problem worse.

If everyone is registered, vandals would be hard to catch

edit

Some people have objected that they look for edits by nonregistered users to catch vandalism. How are they going to catch vandalism if everyone is registered? This won't be a problem because vandalism would sharply decrease. If a registered user is doing the vandalism, he could be easily dealt with by a warning and then a few days ban.

But some of the worst users are registered users, not unregistered users

edit

This change would make it easier to deal with bad registered users. As an example, after RfC was posted against a user for personal insults, the abuser simply gave up that ID. He can create another ID a few weeks later and start the abuse again. These problems cannot be dealt without requiring registration with real email as suggested. Since this place is growing, these problems would grow rapidly when more trolls join this place. It could potentially become a site filled with only trolls and POV pushers and everyone leaving.


Poll

edit

The poll result was approximately 8 supporting and 55 opposing. The results have been archived here. Maurreen 19:38, 12 Dec 2004 (UTC)

Comments

edit

Obviously proposals such as this do not command majority support, as we already know. I would however like to see some of the people who oppose it making some constructive suggestions on how to deal with the problems OneGuy has quite correctly identified, particularly the infestation of Wikipedia by anonymous idiots who drive away serious editors. Please don't just tell us that the Wiki process is sacred and can't be modified, as that is not an answer. The answer has to be framed in terms of the product and not just the process. Adam 01:30, 10 Dec 2004 (UTC)

The problem with trolls is the antisocial behaviour of the persistent registered users. Particularly those fond of dealing with others by casual personal abusiveness. How does this proposal deal with such people? - David Gerard 22:46, 10 Dec 2004 (UTC)
This proposal deals with such people by allowing admins to impose long term ban on these accounts because the accounts would be associated with valid e-mails. Right now admins cannot ban people for longer than a day or two because banning dynamic IPs blocks everyone who is using that ISP, library, internet cafe etc. OneGuy 22:57, 10 Dec 2004 (UTC)
I agree - it's like bashing your head again a brick wall trying to get rid of the current inadequate system while people belive it to be sacred when it's not - just like flags and wilderness are not sacred. PMA 04:13, 10 Dec 2004 (UTC)
I'd like to see existing structures tightened up and used properly before inventing new ones. It seems to me that there's no will here for the long-term or permanent blocking of nuisance editors. I can personally think of five editors that any reasonable person would call persistent vandals, extreme POV pushers and troublemakers. They're blocked briefly for violating 3RR, then they're back again, but the issue of them being totally inappropriate editors is never addressed.
The question will be asked: Who's to judge who is totally inappropriate? I would like to see something like the following: (1) A strengthening of the description of the NPOV and "no original research" policies, so that all perceived loopholes are closed. I think they're good policies, they're coherent and they work well when people understand them, but lots of editors don't, and as things stand, the policies are not enforced so why should troublemakers respect them? (2) A large and very visible reference to these policies should be placed permanently on the front page so that no editor can claim they didn't know about them; (3) There should be a committee of clever, trustworthy, reasonable editors with a variety of expertise so they're in a position to judge the quality of most articles -- and not elected, but appointed (on a voluntary basis) by the Foundation -- who are given the power to act swiftly and decisively in enforcing NPOV and "no original research," without having to offer reasons and justifications to the community or to the individual editors who are blocked. There can be a right of appeal, but only to the committee itself. And the blocks should be long-term, not 24 hours, because blocking someone for 24 hours is not a deterrent. It's fairly obvious when blocked users pop up again using sockpuppets, because they turn up in their old haunts and the editors who complained about them in the first place recognize them. And arguably if those editors don't recognize them, then the blocked editor has clearly learned his or her lesson and so there might be no harm in that person staying on under a different name. (I know there's a more general problem with anonymous IP address vandalism, but I feel it's less of a problem than the small number of registered known troublemakers.)
The creation of a non-elected committee with these powers might seem harsh, but it's a lot better than changing the open nature of the community. All people have to do to avoid coming to the committee's attention is read very carefully the NPOV and no original research policies, then stick to them. It's not hard. Slim 04:55, Dec 10, 2004 (UTC)

I think it's important that the following be stated: nothing lasts forever. One day Wikipedia will fall into disuse and die. Perhaps soon, perhaps not. We should be using our energy to ensure that the work we are putting into this project is protected, and continues to live in one form or another. Until that day, wikipedia stays the way it is. Otherwise, it's not wikipedia. -- Sean Kelly 07:36, 10 Dec 2004 (UTC)

I don't think my proposal would radically change wikipedia. I am not proposing any kind of censorship or privilege usage by some editors. I only offered this as a solution against persistent POV pushers and trolls (not vandals who leave right away but the persistent, in most cases registered, trouble users). If you have a better suggestion on how to deal with the problems I mentioned (banning dynamic IP -- large ISP proxies, etc., see below), I would accept that solution OneGuy 08:30, 10 Dec 2004 (UTC)

Actually, I don't think the problem has been spelled out here adequately, which is part of the problem. Are we trying to deal with casual vandals, or persistent trolls? The former is irritating, but hardly a major drain on the project. Most of the latter are registered users, or would quickly register if required to do so.

I don't see the clear benefit of either of these proposals. We get lots of valuable contributions from anonymous users, including many who stop by and notice something that needs fixing, but would not take the time to register. Attempting to ban free e-mail providers is a really bad idea. First, it's incredibly impractical given the enormous number of free e-mail providers on the web. Secondly, it would be a major stumbling block, or at least a serious inconvenience, to valuable contributors who use free e-mail services.

I might be more open to restricting anonymous edits to some number per day, or some number per article, but I'd have to think about it. Mostly, though, I think what we need is more aggressive action to block vandals and trolls, anonymous or not.

RadicalSubversiv E 08:22, 10 Dec 2004 (UTC)

Of course I am talking about persistent trolls/POV pushers, not occasional vandals. Yes in most cases they are registered, but how do you block them? How do you stop them from creating many more accounts? How do you deal with the issue of proxy servers used by large ISPs? Blocking dynamic IPs doesn't work because that effect thousands of innocent users. When in future you have thousands of persistent trolls/POV pushers, you will have to block a large chunk of IPs. You cannot block the IPs for more than a day or two, making this tactic useless against persistent trolls/POV pushers. The email requirement will significantly reduce their ability to create accounts and help in getting rid of them (at least most of them). If there is a better solution, please post it here OneGuy 08:46, 10 Dec 2004 (UTC)
Thousands? At the moment I would be suprised if with have a hundred persistent trolls. If the number of trolls increases by an order of magnitude then why don't you think that the number of regular users and admins will incease by an order of magnitute. IP blocks are for the most part not that important. Mosts of the antivandle and anti POV stuff is done by editing and reverting rather than by IP blocks which can always be got round if a person is determined enough. Most if not all of the POV pushers I have run into were less anonymous than I am. Geni 10:45, 10 Dec 2004 (UTC)
Unfortunately in some cases you do need to ban people. Editing and reverting doesn't always work. Imagine 3 POV pushers, each registered with three IDs (making total 9 POV pushers), trying to push a POV into an article. This would create a lot of edits/reverts/flame wars. If you cannot block/stop these people, don't you think that would dishearten some very good editors who have worked on the article? Wouldn't those editors reach the conclusion that the place is a waste of time and a joke? Wouldn't that harm wiki? These are the questions that need to be considered OneGuy 14:30, 10 Dec 2004 (UTC)

Alternative Suggestion

edit

I don't think anyone opposing above have responded to the problems I mentioned. People seem to have some weird fantasy about "freedom for anyone to edit." They don't know the power of trolls. I remember in early 90s a discussion about whether the government would censor alt.sex.* newsgroups. The government didn't do anything because they didn't have to. The spammers destroyed these newsgroups. Now, the question is how do you deal with persistent abusers, the sockpuppets, trolls, and POV pushers? What good is RfC against a user if he can simply give up that ID and create a new one? (this has happened, see the abuse here. This guy was a registered user, but now he just gave up that ID because of RfC against him). If people can simply change IDs like that, what good is RfC? And blocking a range of IPs is not a solution because you can't block dynamic IPs or transparent proxy servers for more than a day or two. No one dealt with the problems above while opposing. I have already responded to comments posted by opposing sides in "Common objections" section. No need to repeat them.

In any case, here is an alternative suggestion. Maybe wiki software should have two levels. The articles can be moved between level 1 and level 2 by the admins. When an article is moved to level 2, then only users who have registered with a valid e-mail can edit the article. I don't think this would solve the problem though. Just a thought.

When you have thousands of trolls and POV pusher (not to mention spammers), I don't believe admins would be able to deal with these problems. Blocking dynamic IPs is not a solution. If people see the site infested with trolls, personal insults, and POV pushers, with no credibility, that will eventually drive away far many good users, as happened to many unmoderated newsgroups. Why won't it happen here? No explanation was given by people opposing OneGuy 04:59, 10 Dec 2004 (UTC)

It hasn't happened here, that's the point. You're anticipating something that might never happen, and yet you're prepared radically to alter (and, many would argue, destroy) the nature of the community just in case it might. Trolls and vandals are a problem but nowhere near the point where Wikipedia is "infested". I don't mind your idea of "level two" articles. I feel the structure of the Featured Article process could be strengthened a bit, and then a block put on articles that pass, where only the majority authors and members of a Featured Article committee are allowed to edit them. In summary, I believe that existing structures should be tightened and existing policies enforced before we start inventing new ones. Slim 05:20, Dec 10, 2004 (UTC)
No, it does happen here but is not as extensive yet. I gave an example of abuse here. The admin failed to block him because the abuser came back with a different IP. He also was once a registered user, but after RfC was posted against him, he became "anonymous" again. What good is RfC if people can change IDs? Surely it happens here. The other example I gave was that I posted some evidence of abuse by another user 168.209.97.34 here. Wikipedia:Requests for arbitration/168.209.97.34/Evidence. Someone else posted more evidence against the guy on the same page, but 168.209.97.34 is a proxy server for the largest ISP in South Africa. The evidence posted by the second guy might not have been against the same person. There is no effective solution for dealing with this guy without harming other users who use that ISP. These problems are going to grow as wikipedia grows. No solution was offered by the opposing view. OneGuy 05:33, 10 Dec 2004 (UTC)
And whats to stop people changing ISPs if they are really persistant? Geni
In that case those few can be blocked again if they continue the abuse. The problem right now is the ease with which anyone can create many accounts (or have no account), but there is no way to block/stop them. OneGuy 13:45, 10 Dec 2004 (UTC)
You're being incredibly naive. I can think of several ways to circumvent the ISP emails only regulation without expending any money or serious effort. Johnleemk | Talk 18:54, 10 Dec 2004 (UTC)
Of course, but it would be a lot of work for you do that. You will be banned once again if you get caught abusing using a different email. Not all POV pushers will have patience to go through that much trouble of finding email accounts. If they do abuse again, they will get blocked again. Right now those POV pushers/trolls have nothing to worry about. They can't be banned or stopped alt all from creating sock puppets and pushing POVsOneGuy 19:24, 10 Dec 2004 (UTC)
Hahaha...you're kidding, right? The kind of trolls this proposal is intended to handle don't mind spending time or money to harass us. Determined trolls will always be one step ahead of a technical solution. Besides, once persistent sockpuppeters and POV pushers are handled by the arbcom, admins generally have carte blanche blocking them and their sockpuppets. If a block covers people other than the troll, then it lasts only for 24 hours, by which time the troll should have moved on to another IP. The right thing to do now is to speed up dispute resolution and arbitration so admins will have this freedom to tackle confirmed trolls. Johnleemk | Talk 19:51, 10 Dec 2004 (UTC)
No, I am not "kidding." There are many POV pushers/trolls on wiki. Go to Arab-Israel or religious pages. The proposal would deal exactly with these people. It will solve the problems I mentioned. Arbcom can't solve the problems of transparent proxies. I gave an example of abuse by a user 168.209.97.34. Wikipedia:Requests for arbitration/168.209.97.34/Evidence. Someone else posted more evidence against the guy on the same page, but 168.209.97.34 is a proxy server for the largest ISP in South Africa. The evidence posted by the second guy might not have been against the same person. There is no effective way of dealing with this guy without blocking other users who use that ISP. The proposal that I offered deals exactly with problems like these OneGuy 20:21, 10 Dec 2004 (UTC)
I'm well aware of these POV pushers, having dealt with some before. Even if your proposal solved the problem, that doesn't change the fact that we lose as many, if not more, editors than we would without implementing it. Proxies pose a problem, but this is why blocks on them generally last only 24 hours. That would probably be more effective than blocking Troll X who can reregister an email address and continue his antics. The number of editors disenfranchised is still smaller than those who would be harmed by POV pushers or the banning of anons. Johnleemk | Talk 21:03, 10 Dec 2004 (UTC)
When you block the proxy IP of a large ISP for 24 hours, you block everyone who is using that ISP for 24 hours. If the e-mail criterion is used, you can ban that e-mail for a much longer period than 24 hours without harming anyone else on that ISP. If the POV pusher is back with a different e-mail account, he will get banned again until he runs out of e-mails. Not all POV pushers have access to infinite e-mails accounts. 12:57, 11 Dec 2004 (UTC)
POV pushers won't run out of emails. They can register another domain (there are quite a few free domain registrars out there. There are also quite a few services that provide free email hosting for website owners with their own domain. Put these two together, and what do you get?). They can move to another free email provider. What you fail to grasp is that there are many ways to get free emails. It's impossible to block every free email provider in the world. I can think of at least one British ISP that also offers free email addresses. How do you handle that? I found it through a Google search while I was hunting for a new email account. For every free email provider we block, there will be another left available for a troll to use. Playing "hunt the troll" is a waste of time. Johnleemk | Talk 17:26, 11 Dec 2004 (UTC)
Not everyone (not even most) POV pushers have infinite amount of e-mails. It's a lot of work to keep finding new e-mails. A POV pusher/abuser who only knows how to login to internet and type certainly won't go to the extreme of finding infinite e-mails. The problem would be significantly reduced. To claim otherwise is just pure nonsense OneGuy 17:53, 11 Dec 2004 (UTC)
I don't need to be some uber-nerd to know how to use Google. Johnleemk | Talk 18:58, 11 Dec 2004 (UTC)

Another Alternative Suggestion

edit

For any large group of computers behind a proxy, NAT or otherwise effectively block a large community of both valid editors and vandals, why not just require registrations on THOSE ips? -- AllyUnion (talk) 15:41, 10 Dec 2004 (UTC)

This is possible, though there is still the problem of dynamic IPs used by most ISPs. In any case, this suggestion might reduce the problem to some degree. OneGuy 18:28, 10 Dec 2004 (UTC)
I agree with this proposal, especially with AOL in mind. It is impossible to tell AOL users from each other due to their web proxies. All I want is a way to tell legitimate AOL users from illegitimate ones. Rhobite 20:46, Dec 10, 2004 (UTC)
Is it just me or is "legitimate AOL user" a contridiction in terms. :P seriously though, i am yet to find a problem with annon users. The only probs ive found are with logged in users, and in that case i think that lots of short bans taking immediate effect is the only solution. You know Pavlov's dogs and all that. The bellman 10:40, 2004 Dec 11 (UTC)
We are talking about how to permanently and more effectively ban logged users who are POV pushers/trolls or have sock puppets. OneGuy 13:03, 11 Dec 2004 (UTC)

Assessing the premises

edit

OK, let's look at the premises behind this proposal, and see if they stand up to scrutiny. As many others have mentioned, there are 2 classes of vandal that will be affected here: the "casual vandal", and the "persistent ban evader"; both need considering, because this proposal affects both, even if it is primarily aimed only at the latter.

  1. Slowing down a user's first edit will cut down on casual vandalism: True - if a user needed not only to log in, but to verify an account, they would likely give up before entering silly vandalism.
    • They would also, however, likely give up before making their first "real" contribution, and I have no doubt we have gained many committed users who originally stopped by to correct some punctuation or grammar. So at the very least this is a double-edged sword. So far, we have been able to avoid risking the drain on new contributors by simply dealing with petty vandalism of this sort as it happens.
  2. Ease of editing allows persistent vandals to flout bans: True - it is a perennial problem on the Internet that someone determined to get round a personal ban will be able to do so by simply not revealing their identity.
  3. IP addresses are not good criteria for blocking: True - this is a many-to-many relationship, since people dynamic IP assignment means 1 user has many IPs, and 1 IP refers to many users; so an IP address cannot be equated to an identity.
  4. E-mail addresses are a good criterion for blocking: False - it is very easy for 1 user to acquire many e-mail addresses.
    1. E-mail addresses, excluding those from certain providers, are a good criterion for blocking: False - although disallowing Hotmail and Yahoo! Mail would make it marginally harder for a banned user to return, it is important to remember that these people know they are flouting a ban; unlike casual vandals, "speed bumps" are not enough to stop them. Enforcing this policy would entail the following:
      • Research, and constantly update, a complete list of all free web-mail providers in the world (note that some, such as Another.com, have literally hundreds of domains to choose from, all of which would need blacklisting)
      • Research, and constantly update, a complete list of throwaway spam domains (e.g. "SpamGourmet", "SpamHotel"); again, some of these offer a choice of domains, which would need to stay up-to-date.
      • Monitor BugMeNot for "fake" registrations, and block them; also watch out for any copycats of this service.
      • Attempt to identify "vanity domains", where a blocked user can create any number of e-mail addresses under their own domain name. Also, some ISPs (e.g. Demon Internet) offer unlimited e-mail accounts under a personal sub-domain (e.g. "<anything>@myusername.demon.co.uk"). Both cases would need manually spotting and blocking completely (with the associated risk of "false positives" causing innocent users to be blocked).
      • Monitor a constant flow of e-mails from people saying "please can I edit even though I don't have an ISP-based inbox"; apart from being a waste of volunteer admins' time, this is exactly the kind of confusing and off-putting process that a wiki is against. (OK, I understand that everything has to change, but think carefully: would you be editing here today if you had to ask for permission to create an account, and then wait for approval before editing?)
      • And even after all that, there would be absolutely nothing we could do about ISPs offering users multiple e-mail accounts - my ISP, freedom2surf, offers 20 inboxes per account, and each of these can have aliases; even if these aliases are included in the limit, I suspect that deleting one frees up a "space" to create a new one. So you'd need to block me at least 20 times before you were truly rid of me, unless you got fed up and blocked everyone who uses f2s. So essentially we're back to the same many-to-many relationship as with the IP address blocking we are trying to replace.

So, to summarise: slowing down every user's first edit may or may not do more harm than good; using e-mail addresses as identification is not as trivial as it might seem; and, in general, wiki is premised on the idea of it being quick and easy to edit. I sincerely believe that many users would not join the project if they felt they had to "jump through hoops" before they could correct their first spelling error. People being banned but returning anyway are always going to be a problem, and making them use a new e-mail address each time isn't going to bother them much. - IMSoP 17:09, 10 Dec 2004 (UTC)

Some valid objections for the first time. Yes, monitoring all the free email services is a lot of work, but if we can write an Encyclopedia, surely we can find people who would monitor and update the list of free email services. As for some ISPs allowing many e-mails, true. Using email as a criterion to block people (instead of IPs) won't solve the problem completely, but it would significantly reduce the problem. Not many POV pushers would have access to many accounts. This would make a lot harder for them to continue the abuse. They would be more careful. Right now we cannot block them at all, nor can we effectively deal with the problem of sockpuppets. The email requirement will reduce these problems significantly but obviously not entirely solve them completely. OneGuy 18:21, 10 Dec 2004 (UTC)
You're saying you can "find people". I notice you've yet to say you'd do it yourself. How much time are you willing to personally commit to such a task? How many people are really going to contribute more than zero hours a week to such an ill-considered and technically inept proposition, when they could be editing or writing articles? - David Gerard 22:51, 10 Dec 2004 (UTC)
You forgot that anyone can register their own domain. Some sites (i.e. dot.tk dot.tk) even let you do this for free. And personally, I think monitoring these sites and free email providers is a gigantically large waste of time for an isolated problem. So we lost a few editors. So what? People who can't take the heat shouldn't enter the fire. The internet is a naturally trollish environment. Trying to alter it will only stifle growth. The proper thing to do now would be to speed up arbitration and dispute resolution dreadfully instead of destroying the wiki process. Johnleemk | Talk 18:57, 10 Dec 2004 (UTC)
No, I didn't forget that people can get their own domain. I already answered that argument. It would be easy to spot people abusing/creating sock puppets if they are using the same domain to abuse the same articles. What good is a fast arbitrator if he cannot block the abuser? How would arbitrators block a transparent proxy IP used by the largest ISP in South Africa without blocking everone who is using that ISP? You have not understood the problem OneGuy 19:30, 10 Dec 2004 (UTC)
No, you have. If admins are expected to block based on usernames, trolls can create even more and continue their harrassment. If admins continue to block based on IP, this proposal solves nothing. Either way, this proposal affects trolls little to nothing while disenfranchising thousands of legal editors. The minor trolls go away after a while. Persistent ones don't mind spending time and money harrassing us. Johnleemk | Talk 19:48, 10 Dec 2004 (UTC)
Exactly. If admins block based on usernames that are associated with valid emails, that significantly reduces the problems. It would be a lot of work for POV pushers to find valid emails to create new accounts. That reduces the problems that I mentioned significantly. IP based block for a day or two are useless. If a POV pusher finds another e-mail, he will be careful not to continue the abuse or get banned again. There are persistent POV pushers on wiki. Go to Arab-Israel or religious pages OneGuy 19:54, 10 Dec 2004 (UTC)
There are a lot of POV pushers - but there are more NPOV editors who will actively stop vandalism and be reasonable in writing articles. Like God for instance - a religious article - has improved greatly but stayed NPOV against the hordes. --[[User:Whosyourjudas|Whosyourjudas\talk]] 20:55, 10 Dec 2004 (UTC)
Here is an opposite example. Just one POV pusher has successfully kept this article Historicity of Jesus] remained protected for a month or so now. Another example are edit wars going on at Violence against children in the Israel-Palestine conflict. Hundreds of more such examples can be cited. How many good editors would be disheartened by this and leave if these POV pushers can't even be banned because of proxies? OneGuy 21:13, 10 Dec 2004 (UTC)
Oh and see the wonderful history of Violence against children in the Israel-Palestine. These are persistent POV pushers, many of them. These problems will grow. These people must be dealt with before they take over most of these controversial topics and drive away good editors OneGuy 21:42, 10 Dec 2004 (UTC)
Yeah, right. I already told you, these people don't mind wasting their time reregistering emails. You're being extremely naive by assuming these trolls will eventually give up. Johnleemk | Talk 21:01, 10 Dec 2004 (UTC)
How are they going to keep finding these emails? Even if they do (not possible for all POV pushers to keep finding valid emails), they would keep getting banned if they continue to push POV/troll. Right now you can't ban them at all OneGuy 21:06, 10 Dec 2004 (UTC)
Most POV pushers and trolls are so determined, they will easily find ways to circumvent this. The threat of being banned doesn't faze them. As David Gerard said below, we need to toughen up policy. And as I said before, we also need to toughen up the arbcom to enforce this policy. Johnleemk | Talk 08:46, 11 Dec 2004 (UTC)
That's not true. Most POV/pushers trolls don't have access to infinite amount of e-mails. You would get rid of most of them. If a few do have access to infinite emails, they will keep getting banned anyway. Currently there is no way to ban them without blocking everyone on that ISP OneGuy 13:12, 11 Dec 2004 (UTC)
Like I told you, these people don't give a damn whether they keep getting blocked or not. The ones who do are already despatched by our current policies. And in any case, just because we can doesn't mean we will. Even if we do make registration mandatory, have you considered that we will need to implement policy that allows for blocking POV pushers without arbcom approval? If we don't, then this proposal is worthless, because it won't solve anything. The issue right now is that the only way to punish POV pushers is to get the arbcom to make a ruling on them. Once they do, then, and only then, do ISP proxies come into the question. And that is why they are blocked only for 24 hours. Maybe we'll hurt a few thousand users. Better than hurting our entire user base by denying them the chance to edit at all, and in turn denying ourselves of new blood. Johnleemk | Talk 17:31, 11 Dec 2004 (UTC)
Yes, we are talking about people who after being banned by arbcom can't really be banned because they are on AOL or other large proxies. Not all these people are expert hackers who somehow are going to find infinite amount of e-mails. They can hardly use one e-mail that they do have. They usually don't have a clue about the internet other than post insulting comments on forums. What would wikipedia do about these people when more of them find this place and start editing/pushing POVs in articles? OneGuy 18:07, 11 Dec 2004 (UTC)
You don't need to be an expert hacker to know how to use Google. It's the vandals and vanity page authors who fit your stereotype better. And what would we do with POV pushers? Continue blocking them like we do, provided the arbcom actually do their work. Better to end up blocking a thousand or two users than blocking ten or twenty thousand users. Johnleemk | Talk 19:00, 11 Dec 2004 (UTC)

OneGuy, you keep not answering two basic questions: (1) Many, if not most, ISPs allow multiple e-mail addresses. I am allowed a certain number at any one time, but I can change any of those e-mail addresses at will. So I could create ten, have them all blocked, create another ten, blocked, another ten, and so on. Just like Hotmail addresses. And it takes only seconds to create one; actually faster than Hotmail addresses. Therefore, you would have to block my ISP, which is a large and common one, and that is precisely the problem you were trying to avoid with this proposal. I would say your proposal fails on that objection alone. (2) You haven't addressed the question of how people outside North America and Europe would be affected by your proposal. Are you saying that an editor in, say, Iraq, a war zone, would not be allowed to edit Wikipedia he can't afford an ISP-based account (or such a thing may not exist) and even though there's a working computer in the local library, he can't use it for Wikipedia because he'd only have a Hotmail account? If you are saying yes to this, your proposal ought to fail on that count too. Slim 21:15, Dec 10, 2004 (UTC)

Yes, in cases where people don't have valid emails they will have talk to admins or post in a special section then get an account after admins are satisfied. Their initial activities could be monitored to ensure no suspicious behavior. If an ISP allows infinite emails, then yes, that ISP would have to be treated like hotmail. Users from that ISP would have to ask permission and then get an account and will have to be monitored etc. OneGuy 21:28, 10 Dec 2004 (UTC)
Well, that's most of the big ISPs barred then, treated like Hotmail, in which case you may as well allow Hotmail. Also, don't you think it's unfair and anti-Wiki philosophy to require that poor people (because that's basically what you're saying) be monitored in a special section to ensure no suspicious behavior before being given an account? Slim 23:05, Dec 10, 2004 (UTC)
They can be monitored for a few days only. If they don't immediately start pushing POVs/trolling, they can be removed from monitoring. OneGuy 13:11, 11 Dec 2004 (UTC)

Yet another alternative suggestion

edit

Perhaps instead of making things harder for would-be vandals (many others have already discussed the potential problems with this), we could think of ways to make things easier for the people doing RC patrol. What I'm thinking is we could divide users up into three groups: "trusted users", "unknown users", and "suspicious users". A trusted user is any registered user who manages to make, say, 100 (200? 500? suggestions?) edits without getting banned. An unknown user is either an anon who's never been banned, or a registered user without enough edits to be trusted. A suspicious user is any IP or registered user who's ever been banned. Perhaps we would want to take people off the suspicious list if they can go x number of months without getting banned again.

The point of classifying users like this is that we could put links somewhere, say for example on the RC patrol page, for "view recent changes by unknown users" and "view recent changes by suspicious users." There could also be links to filter the new pages based on user type. Hopefully, this would allow anyone doing RC patrol to focus their efforts on the pages most likely to contain problems. This change would be completely transparent to legitimate newbies.

Of course, anyone who's a regular at Slashdot can tell you that we need to be careful with anything like this, so it doesn't turn into a game. However, we already have something somewhat like this -- the option to hide registered users from Recent Changes. I'm suggesting a refinement. I anticipate the "unknown" list would contain most newbie tests and a good percentage of the vandalism, as well as vandal/troll sockpuppets that haven't been caught yet. The "suspicious" list would contain the troublesome IPs others have discussed (high schools, internet cafes, etc.) as well as persistent vandals, whether or not they have an account. In the rare case of a vandal or troll getting onto the trusted list, that will only last until an admin bans their account.

Don't like the groups I suggested? That's OK, I'm just throwing this out as a starting point. People who have been around here longer than me would need to make sure the details are right. Dave6 21:48, 10 Dec 2004 (UTC)

This is a good suggestion for monitoring vandalism. But there are persistent POV pushers/trolls. These people do not vandalize random pages. They violate wiki policy of, say, NPOV, sometimes openly sometimes more subtly. They argue about their edits but are not open to reason. Perhaps in your suggestion only trusted users would be able to edit controversial articles that are put by admins in "level 2" (one of my suggestion above). That might solve the problem without requiring registration. Only wiki software developers can probably make these changes (if they are even convinced it's the right idea). OneGuy 22:29, 10 Dec 2004 (UTC)
I think we need to be clear about what problem this or any other proposal is intended to solve. If the problem is persistent, subtle POV pushing, we need a proposal that deals with that specifically. Required registration and other such speedbumps are usually used by websites to keep out link spammers, "lol don is gay", and similar junk. It doesn't work against someone with an ax to grind that isn't blatantly vandalizing or flamebaiting. If this POV pushing is limited to just a few articles, I could see myself supporting a proposal based on your "level 2" suggestion. Let's not discourage people from editing the 99% of articles that are noncontroversial if we don't have to. Dave6 00:57, 11 Dec 2004 (UTC)

Jwrosenzweig's comments

edit

OneGuy, you wanted comments, so I'll oblige also. I agree with everything ImSoP said, and I don't feel you have addressed his concerns adequately. Even if, however, you are right and we can monitor the free emails and restrict things appropriately, I still object, because we will be discriminating against a class of human beings I was in until recently. This is the class of users who do not have enough money to pay for an ISP -- we use libraries, friends' computers, etc. to access the Internet, and free email accounts. I don't see why we should impose hurdles for these people. How would I have convinced an admin to accept my "application"? I don't see any way that will work. And what if some admins simply refuse to respond, or disappear from the site? I may be leaving a note for an admin who will never get back to me, or who will never give me a chance to prove my identity (and again, I don't know how you think I would have done that). OneGuy, what it comes down to is this: you will drive off more good editors (myself included) with the above policy than any band of trolls ever organized has been able to do here yet. Our problems with trolls aren't the ease of their access to this site -- it is that there is no clear standard for "trolling" and therefore no way to effectively punish many of our rule-breakers. What we need is a way to distinguish what trolling is. Once someone's banned, I'd say we're generally effective in keeping them away -- most of them couldn't hide for more than a day, as their biases control their activity too completely. And if we have a good way of identifying trolls, then even the ones that return will be banished quickly enough. No, our problem isn't people's ease of access -- it's that we know exactly who they are and why they're here, but we have no policy that allows us to get rid of them. If I were you, I'd expend your energy on refining a description of trolling that includes our worst offenders without severely restricting free speech and hampering our good editors when they work on controversial topics. I grant it's not easy, but it would be better than the above. Jwrosenzweig 22:43, 10 Dec 2004 (UTC)

A more firmly enforced policy against personal abuse and personal abusiveness would do a lot more than this technically infeasible proposal - David Gerard 22:52, 10 Dec 2004 (UTC)
David, as usual, we agree. I think we are far too lax on incivility, especially if we believe that the incivility is "understandable" -- all that does is encourage people to exercise the worst aspects of their character. Jwrosenzweig 23:00, 10 Dec 2004 (UTC)
Agree completely with Jwrosenzweig and David. Poor people should not be excluded from Wikipedia, or be made to jump through hoops. That violates the entire ethic of the product. And secondly, as these users point out, we KNOW who the POV pushers and trolls are, but have no way to ban them, not because of technical difficulties, but because there is no clear definition of what unacceptable behavior is; or rather (and here I slightly disagree with Jwrosenzweig) I think we do have definitions; we just don't enforce them with any rigour or consistency. Slim 23:12, Dec 10, 2004 (UTC)
Slim, I guess I'd say that I believe you and I do agree -- I think there are people who violate this site's policies and principles who are not censured in any way, and it's because we're inconsistent. I think we also have some very cagy trolls who manage to subvert the process and interfere with this site's goals without actually violating any policies, and I think they're very hard for us to combat. I don't know if you agree, but I suspect you do, and certainly this is what I had initially intended to say. Jwrosenzweig 23:16, 10 Dec 2004 (UTC)
Yes, I agree absolutely. Slim 23:34, Dec 10, 2004 (UTC)
Slim and Jwrosenzweig, I propose a change of rules that may facilitate dealing with POV pushers i.e. we should state explicitly and more adamantly that contributors should take their information from multiple scientific, scholarly, or mainstream sources. Use of single non-mainstream, partisan, non-scholarly, non-scientific sources should be explicitly discouraged. To paraphrase Slim, anyone who favors ideology or a belief system over facts is a danger to Wikipedia. Andries 11:40, 11 Dec 2004 (UTC)
I think this is a solution looking for a problem. Our current no original research and NPOV policies are working fine. The problem with POV pushers is that they are left alone for too long. We need to streamline the arbitration process. Johnleemk | Talk 11:53, 11 Dec 2004 (UTC)
Regarding this: "Once someone's banned, I'd say we're generally effective in keeping them away." If that is true, then I don't have any problem. But I would think it would be pretty hard (if not impossible) to keep them away if they use AOL or other large ISP proxies. How about the suggestion that only these ISP or computers behind a proxy, NAT are required to register? OneGuy
I don't know who wrote the above statement, but I'm happy to reply. AOLers and proxies are tough, of course, but I think we're generally good at using auto-reverts and short-term bans to drive off vandals and trolls from these IP addresses while minimally inconveniencing genuine users. Reverting a few times usually drives off all but the toughest cases, in my opinion, and for the toughest cases, I think we can risk the 24 hour bans we impose (yes, it may drive off the occasional editor, but I doubt it's very many). Jwrosenzweig 23:18, 10 Dec 2004 (UTC)


Yes, this was my original reason for this RfC. However, I am not talking about vandals. If, say, a user using AOL proxy is banned by the arbitrators, and this user is a persistent POV pusher/personl insults (involved in wide variety of political/religious topics) before he got banned, how would you keep him away? There is a problem here, but perhaps Dave6 suggestion above of "trusted users" being able to edit controversial topics is the answer OneGuy 23:38, 10 Dec 2004 (UTC)
I think a key point is that blocking based on e-mail would be no more effective than blocking based on IP address. In the case of a persistent abuser signing on from a large ISP [e.g. AOL], they are likely to have access to multiple IPs, which are indistiguishable from those of honest users - and also multiple e-mail accounts, which are indistinguishable from those of honest users. So while you have identified a real problem, your proposal doesn't solve this problem. [Or, to put it simply: how would you keep such a user away?] - IMSoP 00:40, 11 Dec 2004 (UTC)
They might have access to multiple e-mails but they won't have access to infinite e-mails. Gosh, I have been banned from email lists many times. I could have signed back using a different free email but didn't bother because of the hassle of creating a new account (even creating a free one like yahoo is a hassle) and I knew I would get banned right away if I post, so I didn't see the point. That's the point. The e-mail based banning (instead of IP based) will reduce the problem significantly; the persistence abusers would keep getting their new accounts banned as soon as they abuse again. This would definitely reduce the problem to a very large extent. You might have other objections to my proposal but to claim this would do nothing to reduce the problem is not right OneGuy 02:29, 11 Dec 2004 (UTC)
You underestimate how determined the serious trolls and POV pushers are. Irismeister even filed complaints with the FTC about several arbitrators and administrators here. It's a hassle, but they don't have anything better to do. Johnleemk | Talk 08:42, 11 Dec 2004 (UTC)
Most POV pushers don't have access to infinite e-mails. Anyone who has used e-mail lists knows that kicking/banning emails works in most cases. You might have other objections to the proposal but I will not accept that this solution will not reduce the problem significantly. OneGuy 13:20, 11 Dec 2004 (UTC)
Yes, they do. I already told you that they have the advantage of a lot of time to waste. With POV pushers, they have an ideology as well. While banning works for conventional trolls, Wikipedia is not a conventional mailing list. And this sort of thing rarely, if ever, works on POV pushers. For example, one message board I know was faced with repeat visits from a troll despite being IP banned. This troll was a radical male chauvinist who was also a fundamentalist Muslim and kept shoving his viewpoints in other posters' faces. Despite concerted effort from moderators, he has promised to return once every few months under a new account. POV pushers don't mind the extra effort they have to waste. And they do have an unlimited access to free emails — this is the bloody internet, after all. For every free email provider we block, there's another one still open to trolls. They're very easy to find. Your average troll is smarter than the average non-techie. Average non-techies may vandalise Wikipedia for a while but eventually leave. The long term ones don't care. Johnleemk | Talk 13:32, 11 Dec 2004 (UTC)
Okay, so there would still be a few persistent POV pushers left who somehow would keep finding e-mails accounts and keep getting banned. Doesn't matter. Vast amount of POV pushers can't do that, nor would they have the time and patience. Currently you cannot ban any troll/POV pusher (not even a guy who doesn't have a clue about the internet other than logging in and typing). The problem would be reduced by getting rid of these people, which we can't right now OneGuy 13:45, 11 Dec 2004 (UTC)
I'm not yet 100% convinced that there is a large enough group of people who:
  • can't be banned by IP [for longer than it would take to get a new e-mail address] due to the "collateral damage" of blocking dynamic IPs
  • and aren't put off by being told "you're banned"
  • but would be put off by the extra effort involved in finding a new e-mail address
I say "large enough", because what's proposed is a pretty radical change of policy, with serious consequences for the "feel" of the site (in terms of inclusiveness etc), so it's not worth doing unless it will give a significant reduction in a real and present threat. I don't think it's true that we can't ban these people at all right now - we can and do put temporary blocks on IP ranges (which annoys people, but certainly no more than making them beg for acceptance if they didn't have a "real" e-mail address would); the abusers come back, and we have to ban them again but we would have to with e-mail verification anyway. Perhaps they won't have "infinite" addresses, but even with an absolute limit of 20, a troll can stage one comeback a fortnight for a whole year, or one every day or so for most of a month. The benefit really doesn't seem to outweigh the cost, in my view. - IMSoP 15:06, 11 Dec 2004 (UTC)
Yes, the problem is not large enough yet (except in cases like AOL and other large ISP proxies where banning just one IP bans everyone on AOL). Wikipedia currently has around 35,000 registered users. It's possible that in future, as this place grows, the problem would become very large. Would you support the change if that happens? I still believe that e-mail as a criterion to ban would be more effective solution (i.e. won't effect other users from that ISP -- not infinite e-mails available to everyone etc). Does the software currently even has the feature of allowing edits by only registered users whose account has been verified using automatic e-mail confirmation? If not, then at least the software should have this feature, even if the feature remains disabled. What do you think? OneGuy 15:54, 11 Dec 2004 (UTC)
My personal suspicion is that if and when the problem with persistently reoffending AOL users becomes large enough that the inability to do long-term IP blocking is a severe handicap, we will have to come up with some new policy, but that this probably won't be it. My reasoning is that if people are willing to take advantage of unblockable IPs, they will equally take advantage of unblockable e-mails (AOL, last I checked, allowed multiple addresses; do we have any evidence that you can't just cancel one and activate another in its place? If so, we are back with no more effective a filter than the IP address). E-mail may be a slightly more reliable criterion than IP, but I strongly suspect that it is not enough of an improvement to justify the high social cost of switching to it. So, although I can't predict the future, I suspect I would never support this policy in its current form.
And adding a feature to the software that no-one actually wants to use is a waste of the already overstretched developers' time. - IMSoP 16:32, 11 Dec 2004 (UTC)
Even if a few of them continue to create new e-mails (and most people can't nor have the patience for it), you solve the problem of blocking everyone who is using that ISP (public library, university, internet cafe), in order to block one person. The problem would be significantly reduced. Anyway, we are just repeating ourselves now. The discussion is exhausted, at least on this point. If someone asks the question again why unregistered users are allowed to edit, they can be referred to this RfC (much better discussion than on the talk page of meta.wiki where I didn't find these arguments) OneGuy 16:56, 11 Dec 2004 (UTC)
Only "trusted users" being able to edit controversial topics is a terrible idea. It's a cliched phrase, but seriously that goes against the principles of Wikipedia. Jonpin 01:18, Dec 11, 2004 (UTC)

I edited first time on wikipedia as unregistered

edit

This should have been answered in one of the "common objections" in the original proposal :)) I guess I missed it. Yes, you might have edited first time on wikipedia as unregistered (like most other people including me), that doesn't mean you would have never joined wikipedia if you had to register to edit first time. In any case, another alternative idea is that unregistered users may be allowed to edit some (even most) articles but not all articles, especially not the articles that have gone through (or are still going through) extensive edit wars by POV pushers. This would allow unregistered users to "try it" first before they register while solving the problem of POV pushers OneGuy 02:49, 11 Dec 2004 (UTC)

I think "try before you buy" is essential. I would have quickly lost interest if anonymous edits weren't allowed. Aerion 04:07, 11 Dec 2004 (UTC)
As would I. I would have simply never bothered registering if I had never been given the chance to see what it would be like. Johnleemk | Talk 08:41, 11 Dec 2004 (UTC)
Likewise, and apparently I'm now one of the top 2000 contributors by number (mostly because I have a tendency to fix spelling and punctuation and sentence structure in otherwise correct articles). Blocking edits for the unregistered would greatly reduce the influx of new Wikipediusers. DS 16:26, 11 Dec 2004 (UTC)

Consensus statements

edit

It appears there is general consensus for the following statement:

There is laxness in the enforcement of Wikipedia policies.

There may be consensus for the following statement:

This laxness is exacerbating trollish behaviour, the development of parties/cabals, and POViors.

There is some agreement for the following statement:

Election polls may weaken the impartiality of admins, arbcom members, mediators, etc.

- Amgine 05:52, 11 Dec 2004 (UTC)


I saw the first two statements, but don't recall seeing the third. Slim 06:15, Dec 11, 2004 (UTC)
Heh... I based that on *your* unrefuted comment! (it is implicit in your argument "The creation of a non-elected committee with these powers might seem harsh...") - Amgine
Oh yes, I see what you mean. I didn't really intend that to be an implication, but I can see that it is. I was arguing more for a meritocracy than a democracy, at least for this committee, because there would be a need to judge quality, rather than just have enough friends to elect you; but yes, come election time, they'd have difficulty making harsh decisions. So yes, point taken. I take my own point!  :-) Slim

Bayesian Filtering Proposal

edit

I think it's easier if we develop a Bayesian Flitering Edit system. The RC patrol can mark edits for a Bayesian filter to show what kind of changes are considered as "spamish" or "trollish"... in our case, "vandalism." Then in the RC changes, have a percentage listing how much of a chance that the article was vandalised by the edit. Factors would include large amounts of missing text, large amounts of added links, and so on, and so forth. -- AllyUnion (talk) 09:13, 11 Dec 2004 (UTC)

GREAT IDEA!!!! The bellman 10:52, 2004 Dec 11 (UTC)
Some changes will come up as 2% likely when in fact they are extremem vandalism, which RC patrol would ignore. Its only statistic. You have to actually look to see what has happened.Cheesedreams 19:34, 11 Dec 2004 (UTC)
This is certainly true, but at least the big obvious vandalism will be blaringly obvious. -- AllyUnion (talk) 06:17, 12 Dec 2004 (UTC)

Copied from my post on the Village pump:


Basically the idea is: we take the approach of vandalism the same approach we take spam. The more users agreeing the specific edit was vandalism, the more our filtering system will understand what edits were vandalism and what edits are not vandalism.

After all, if we can develop Bayesian filters for spam which learn what is wanted and what is not wanted, why not have the same ability for our edits?

Granted, there are some complex technical issues that will need to be addressed, and whether or not we will consider developing a trust network... but I think it is safe to say if we did a rating system, we could base it on a person's access control and whether they are an Admin, Board member, etc.

Or, the system would automatically mark you down if your vote of the edit is against the majority.

Of course, to prevent abuse, we can exclude certain known vandals from preventing from editing, and prevent anonymous people from voting as well.

Additionally, the filtering system would work to our advantage: Both bad and good edits would be marked by the system. This way, our system would have more and more accurate reflection of what is a good edit and what is a bad edit.

-- AllyUnion (talk) 09:24, 11 Dec 2004 (UTC)

By the way, as suggested by Khendon in IRC, a feed could be done on a separate server of the recent changes. Basically, what would need to be logged is the differences (diff) of the edits. Then all it is marking which "differences" are bad and good. -- AllyUnion (talk) 09:47, 11 Dec 2004 (UTC)

Second idea: Allowing everyone to register with any e-mail

edit

Anyone would be able to register with any e-mail, including free hotmail/yahoo emails -- but email must be valid, not fake. Now, let's say arbcom blocks a user for six months. If, during this six months, the blocked user creates a new account using a different email, the software would automatically compare all the IPs used by the new user to the list of IPs that were used by the blocked user. If the IPs match, this would send a notice to admins (even automatically banning the new account in cases when the IP is not a proxy or a dynamic IP of large ISPs). If the user is found editing/abusing the same articles in similar way, this new ID would be blocked again immediately. This solution removes the objections against the first proposal of not allowing hotmail/yahoo emails but still enforces the decision made by arbcom of 'long term' bans (not a day or two bans). This technique can also be used to immediately catch and ban sockpuppet accounts.

The reason for this proposal is specifically the current inability to effectively ban abusers See: What problems are you trying to solve?

OneGuy 22:12, 11 Dec 2004 (UTC)

Support

edit
  1. :)) OneGuy 19:22, 11 Dec 2004 (UTC)

Oppose

edit
  1. New users won't bother to register even if it's less restrictive. Johnleemk | Talk 19:27, 11 Dec 2004 (UTC)
  2. Cheesedreams 19:32, 11 Dec 2004 (UTC)
  3. I fear even this is unlikely to fly. The current situation where anon edits are deprecated compared to logged-in edits is fine - David Gerard 01:46, 12 Dec 2004 (UTC)
  4. Still no. But maybe having voting like this so soon after the proposal isn't a good idea. Aerion 01:59, 12 Dec 2004 (UTC)
  5. No. [[User:The Epopt|➥the Epopt of the Cabal]] 16:17, 12 Dec 2004 (UTC)
  6. Just makes it harder to track. -- AllyUnion (talk) 23:45, 15 Dec 2004 (UTC)
  1. --Jirate 19:41, 2005 Apr 5 (UTC)

Comments

edit

I gave up on the first proposal above. I concede that the objections raised against ISP based e-mails were valid. Most people don't have e-mails other than free e-mails. OneGuy 19:36, 11 Dec 2004 (UTC)

I think there's a better way to resolve this, but it would also be a radical change to the site. Anonymous edits and edits by "untrusted" recently created accounts should be proofread before they are incorporated into the encyclopedia. They would be checked that the conform to policy, and in particular that the information comes from a verifiable source (and it generally should be verified). Keep a score for each account, giving points for accepted edits. Once it reaches a certain level, edits are accepted without further checking. A queue of submissions would need to be kept and a link available from each article to its unchecked submissions. - Anon.

That's too Nupedia-ish to succeed. Johnleemk | Talk 20:02, 11 Dec 2004 (UTC)
I thought the problem with Nupedia was the entire article wouldn't be accepted without peer review? Under my system, it's still possible to make small contributions, but there will be a delay (of length depending how hard the edit is to check) before it's applied. - Anon
That idea should be carried on a different Encyclopedia that works along this one, making two Encyclopedias, perhaps one having more reliable facts and the other just being larger. Having editors who have to check facts would dramatically decrease the growth of this one. OneGuy 20:18, 11 Dec 2004 (UTC)
It's not just that. We will have to devote manpower to checking these edits and manually approving them. And in the first place, this proposal is intended to handle POV pushers and trolls. What you suggest is a solution looking for a problem, since the POV pushers and trolls this proposal is intended to address seem to have fitted into the system just fine. Johnleemk | Talk 20:09, 11 Dec 2004 (UTC)
Yes, but now there's no way to get rid of them. They can just come back with a IP address or username. At least they could be busted back to untrusted acounts. You are right about the extra work required, but there may actually be people who prefer to do this than write original contributions. See Project Gutenberg Distributed Proofreaders for example, which by coincidence does have a scoring system based on a simple page count. BTW I gave up contributing to Wikipedia because of the reliability issue, so I am no doubt biased against the current system. - Anon.
Well we'll see about the workload. Edit checking is being brought in with 1.4. My prediction is that in fact it will dramatically increase the capacity of the community to deal with "trivial" crappy edits, thus giving us more capacity to deal with more persistent "rogue elements". Pcb21| Pete 22:07, 11 Dec 2004 (UTC)

What problems are you trying to solve?

edit

Rather that continuing to spell out unpopular proposals, why don't you start by clearly spelling out the problem you're trying to solve (more specifically than "stop trolling"), preferably with examples, so that we're all on the same wavelength and can discuss solutions in a more organized fashion? RadicalSubversiv E 22:37, 11 Dec 2004 (UTC)

O well. I thought I did that many times. Let me try again with two examples.

(1) Let's say I post some abusive comments about you on your talk page. I have been an abusive wiki user for many months. Arbcom decides that I must be blocked from wiki for 4 months. They block my ID (OneGuy) for 4 months. The ban is not going to work because I can come back using a different user ID (or without any ID). To ban me, they must ban a range of IPs that I have used (IPs are assigned by ISP -- Internet Service Provider, in my case South Western Bell). However, when you block a range of IPs, you block everyone who might have those IPs assigned (IPs are dynamic in my case). As a more extreme example, when you block AOL web proxy, you block everyone who is using AOL. That means admins cannot block AOL IPs for more than 24 hours because it blocks all users from that ISP.

(2) Another example. I posted some abuse by the user 168.209.97.34 (he also has a registered ID). Wikipedia:Requests for arbitration/168.209.97.34/Evidence. Someone else posted more examples of abuse by 168.209.97.34 on the same page. However, 168.209.97.34 is a proxy IP for the largest ISP in South Africa. There is no way to tell if the evidence posted by the second person against 168.209.97.34 was not someone else in South Africa who uses the same ISP. That's a problem. Moreover, if arbcom decides to ban 168.209.97.34, they would end up banning everyone who is using that ISP in South Africa.

Now, the above proposal solves both these problems to a great degree. First, we won't have confusion about which user from 168.209.97.34 was doing what. Everyone would have a separate ID. Second, admins would be banning people using their ID, associated with an email. When they ban an ID for six months, they don't have to worry about other people who are using that ISP (as in South Africa) getting blocked for six months too.

Some people have objected to this solution by pointing out that some abusers will keep finding new e-mails and registering new accounts anyway. I have two answers to that objection. First, not all these abusers have time or patience to keep finding/registering new e-mails. Some of them hardly know anything about the internet other than how to turn on the computer and post abusively. Second, if a new account is created by a banned abusers using a different e-mail, the software would automatically compare the IP used by the new account with a list of IPs that were used by banned account. If the IPs match, software would notify the admins (or ban it automatically if it's not proxy for large ISP). Notice this second ban again will ban using e-mail as a criterion, solving the problem of banning IPs that bans everyone who is using that ISP. If he comes back third time with a third e-mail, the software would ban him again third time automatically using e-mail as a criterion, etc, etc., eliminating the problem.

I hope it's clear now. If not, I don't know what to say OneGuy 23:52, 11 Dec 2004 (UTC)

I think the problem was already clear, and at this point you're beating a dead horse into the ground. The issue isn't that the problem isn't clear, it's that a lot of people don't like your proposed solution. Many, including me, disagree with anything that will make the process of editing significantly more difficult for newcomers. Fortunately you've eliminated the part of the proposal that involved talking to an admin - that was an excessive burden on both the user and the administrator. Aerion 01:56, 12 Dec 2004 (UTC)
I was asked to "spell out the problem." Apparently some people still don't understand it. In any case, that was last time for repeating the problem, unless someone posts some new objection OneGuy 03:39, 12 Dec 2004 (UTC)
In response to the examples you provide, wouldn't it be easier for a Wikipedia representative to deal with the ISP where the abuser's account is located? After all, if we can't ban an IP address because it is a proxy for a major ISP, then the problem should be passed to the ISP to handle -- or face the consequences. I remember that this was suggested in the past with a certain troublesome AOL customer, but I never heard if it was actually implemented or what its success was.
(And, if an ISP will only listen to someone with the correct title, the ArbCom should be empowered to appoint a volunteer with the necessary title to deal with the ISP.)
I'm suggesting this only because to adopt this proposal, we need to first consider all of the possible alternatives & determine that this is the best solution. I haven't seen this done yet. -- llywrch 19:37, 13 Dec 2004 (UTC)
That's impractical. It's not ISP's responsibility to check whether their customers are pushing POVs or are being offensive to other members on wikipedia. ISP would never harass their customers because of such stuff. The only time ISP might act is when a customer is posting child porn, or sending thousands spam email using their e-mail server, or doing something illegal like that. Other than that, this is not their problem. They neither have time (who will pay them for extra work?) nor a reason to get involved with Arbcom decisions OneGuy 19:57, 14 Dec 2004 (UTC)

A suggestion: edit throttling

edit

Often, vandals edit extremely quickly, attempting to overwhelm admins' efforts to tidy up after them. Recently, we've even seen coordinated groups of vandals at work. When these vandals share a proxy with legitimate users on an ISP that uses a proxy that hides the originating IPs of its customers, merely blocking these vandals has the effect of also blocking legitimate users. It might be a good idea to do the following: apply edit throttling to these netblocks, limiting them to only x edits per minute. x can be set at a level that would not inconvenience legitimate users at normal levels of use, but will effectively slow down attempts at rapid-fire vandalism, allowing admins to easily clean up at a rate faster than the vandals can create problems. -- Karada 14:22, 12 Dec 2004 (UTC)

I would support this as long as it replaces blocking for such cases (instead of blocking shared IPs, we apply edit throttling). Johnleemk | Talk 16:59, 12 Dec 2004 (UTC)


Anonymous comment

edit

You say above

The problem is not vandals. Vandals don't stay that long . The problem is persistent POV pusher and trolls who are here for months/years without ever leaving.

I am ok with trolls, but a POV pusher, if intelligent, keeps easily editing on at the WP. I see it (and read it a lot). 80.58.23.44 16:15, 12 Dec 2004 (UTC)


Another anonymous comment. I have always edited anonymously and it gratifies me to visit some of the pages I've put some key facts on and see that my contributions are intact and enriched.

People need to keep a perspective on this. Trolling represents a very small fraction of the total activity that goes on here. The number of pages that are beleaguered with chronic and persistent disputes such as the religious and national pride type pages should be made into moderated pages. There is only so much true facts that can go into such pages and the contributions from those who feel they have something to say on it, are surely up to the task of being moderated if they are genuine?

So make the required pages moderated if necessary.

As a tor user I also feel that Wikipedia's stance on that issue is misguided. I will always contribute as long as anonymous editing is possible.

Readers or contributors?

edit

I think this general discussion could be divided into two parts, depending on who it is intended to serve.

If it is intended to server readers primarily, by making Wikipedia more reliable, that might be improved by some type of Wikipedia:Approval mechanism or tiers for articles. Another tier idea was suggested relatively recently. No one voiced disagreement, but the discussion died out.

If we're talking about making life easier for contributors, because of POV pushers, that might best be considered as how to improve Wikipedia:Dispute resolution, possibly including banned users who work around the ban. Maurreen 19:21, 12 Dec 2004 (UTC)

All of this discussion is important, but it's not going to go anywhere unless each proposal is considered individually. I personally disagree with most of them, though I believe WP would benefit from limiting editing in certain circumstances (i.e., having the software enforce the three revert rule). Additionally, I believe that most of OneGuy's proposals need to go through the developers, and should be at metawiki. It would be beneficial if someone split up this very broad proposal into a bunch of smaller, more concrete ones.
Maurreen, articles already are approved by the community. That's what a wiki is and what it does. If you wanted to create a new Britannica, perhaps you should apply for a job with them.
As for POV pushers, the dispute resolution mechanism cannot deal with them because it relies on the presumption that users are interested in working consensually together. POV pushers by definition are not. Besides, for some POVs, those pushing them outnumber those opposing them and know how to work the system.
The same sort of problem occurs with banned users. They have to agree to stay banned to stay banned! Here's why. I edit using my home PC at the moment. I'm highly unlikely to get banned, but say I did, I could simply use an internet cafe. If I get that IP banned, I can move to another. Then I could go to a friend's... Do you see? And having to have a User ID is no bar to that.
Barring hotmail and similar accounts would work to some extent but it's very problematic. Many users either don't have ISP email accounts or do not want to expose them but are not engaged in anything nefarious. It goes without saying that opening new email accounts with new ISPs is child's play. This doesn't begin to address the serious failing of such a policy that one of the first principles of Wikipedia is that anyone can edit an article.Dr Zen 05:05, 14 Dec 2004 (UTC)
How convenient would it be for you to post on wikipedia from an internet cafe and a friend house? How often would you be able to do that compared to the current situation when you could do it anytime from your home computer? How many of the banned users would go to the extreme of registering e-mails and going to internet cafes to post something on wiki that would be immediately reverted and the new account get banned? Not many. As for "the first principles of Wikipedia is that anyone can edit," registration doesn't mean that anyone would not be able to edit. I don't see that objection having any relevancy. OneGuy 20:12, 14 Dec 2004 (UTC)

"How convenient would it be for you to post on Wikipedia from an internet cafe". I'm not a vandal, OneGuy. I have no idea what they would or would not find convenient. Banned users would of course go to those lengths. They often return in a new guise and the process of banning is nothing like as quick as you seem to think. I know you don't see that objection as having any relevancy, OneGuy. That's why you've proposed this policy. Dr Zen 01:42, 15 Dec 2004 (UTC)


This is from the enwiki mailing list:

I consider such decisions to remain firmly within my range of discretion, so any such vote would necessarily just be a poll of community opinion, which is of course always a valid thing to do.

As it stands, it seems that community opinion is, as always, firmly in favor of openness, a position which I wholeheartedly endorse.

I *do* think there is a serious problem which needs to be addressed, and I actually agree with Adam Carr's comment that "Unless Wikipedia takes some policy initiative such as this, it will not only be unable to achieve its objectives, it will begin to deteriorate in quality as serious editors are driven away."

But I think this is the wrong approach -- editing by anonymous ip numbers is barely a problem at all. Under current blocking rules, anons are not given the same "due process" protections as logged in users, and so it's easy enough (other than some very annoying technical limitations which could eventually cause me to change my mind) to deal with.

Our biggest problems within the community are not anons trolling and vandalizing, but rather egregious trolls and pov pushers who log in and take advantage of our boundless good will. Making it harder to sign up will not help with this at all, and can actually hurt it if the overall rate of participation by the (good willed) general public declines in the face of some extra burdens of signing up.

POV pushers will jump through the hoops of signing up. Making it harder to sign up does nothing to discourage them. So, yes, we need to do something about how long it takes us to get rid of difficult people -- and this is a difficult problem to reconcile with the demands of NPOV and openness and quality.

--Jimbo

Mrfixter 18:57, 15 Dec 2004 (UTC)


Possible restrictions for anons on:

  • featured articles and controversial articles
  • all non-stubs
  • new articles

Possible implementations of restriction

  • word limits (per minute/hour)
  • edits invisible to the public until OK'd by a registered user.
  • blanket restriction

As long as registration is easy (not needing email), this type of restriction on anons - aimed at reducing vandalism - is not dramatic. Even requiring an email (free being OK) for registration would be fine as it's easy to get free addresses. (Real addresses are costly - requiring those would undermine wiki principles.)