I am sure that this came along many times, but I was not able to search for it through the mailing list.
OK, so to start, users without account (anons) contribute a lot of content, at least as typo fixing, etc., and many established users started as anons.
However, it is also true that the vast majority of vandalism is committed by anons, and it takes a tremendous energy to police the more than 1.7 milions articles for vandalism.
Hereby I suggest that only people with account be allowed to edit, and that they also suppy an email address when registering, which is then confirmed by sending an email to the supplied address and having the user clicking on a link.
This is quite standard nowadays on the vast majority of websites which allow more than just reading things, and people are rather used to it. It takes little time too to register and confirm one's email.
I am aware that this may decrease somewhat the number of people who get hooked on Wikipedia and the amount of contributions. I'd argue however that Wikipedia is at a stage now where it has a very large amount of users, articles, and recognition. At this stage we should care a bit more than in the past about the quality than the quantity of users and articles (while of course we should hope that the community and the number of articles will increase).
In short, I believe that having people make account and confirm their email is going to bring much more gain than loss.
Comments?
Oleg Alexandov
On 11/04/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
I am sure that this came along many times, but I was not able to search for it through the mailing list.
OK, so to start, users without account (anons) contribute a lot of content, at least as typo fixing, etc., and many established users started as anons.
However, it is also true that the vast majority of vandalism is committed by anons, and it takes a tremendous energy to police the more than 1.7 milions articles for vandalism.
Hereby I suggest that only people with account be allowed to edit, and that they also suppy an email address when registering, which is then confirmed by sending an email to the supplied address and having the user clicking on a link.
This is quite standard nowadays on the vast majority of websites which allow more than just reading things, and people are rather used to it. It takes little time too to register and confirm one's email.
I am aware that this may decrease somewhat the number of people who get hooked on Wikipedia and the amount of contributions. I'd argue however that Wikipedia is at a stage now where it has a very large amount of users, articles, and recognition. At this stage we should care a bit more than in the past about the quality than the quantity of users and articles (while of course we should hope that the community and the number of articles will increase).
In short, I believe that having people make account and confirm their email is going to bring much more gain than loss.
Comments?
Oleg Alexandov
I'm sure people who wanted to vandalise would just create accounts to do so - it wouldn't make that much difference (and we'd have to start using CheckUser more). Many readers who fix typos may not bother to create accounts and we'd lose some valuable work.
Most importantly, I think we'd lose claim to being an encyclopedia that "anyone can edit". The greater the number of restrictions we introduce, the harder it is to edit. I, for one, am far less likely to contribute to a website if I'm forced to create an account to do it.
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
This is quite standard nowadays on the vast majority of websites which allow more than just reading things, and people are rather used to it. It takes little time too to register and confirm one's email.
How does us having a valid email address on record help the project? If the registration process is quick and easy, how does it discourage vandals? What problem will your proposal fix, and how will it fix it? It sounds like "vandalism is bad, so let's annoy new users".
Steve
On Wed, Apr 11, 2007 at 07:14:51PM +1000, Steve Bennett wrote:
How does us having a valid email address on record help the project? If the registration process is quick and easy, how does it discourage vandals? What problem will your proposal fix, and how will it fix it? It sounds like "vandalism is bad, so let's annoy new users".
That's the whole point I think. If you supply your email address when you make your account you're less likely to vandalize. And if you get blocked you can't start a new account with the same address while the block is on.
In short, people who want to be on the project will bother registering with a valid email, people who want to just vandalize won't.
On 11/04/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
That's the whole point I think. If you supply your email address when you make your account you're less likely to vandalize. And if you get blocked you can't start a new account with the same address while the block is on.
Q. How many email addresses do I have? A. As many as I like.
- d.
On 4/12/07, David Gerard dgerard@gmail.com wrote:
On 11/04/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
That's the whole point I think. If you supply your email address when you make your account you're less likely to vandalize. And if you get blocked you can't start a new account with the same address while the block is on.
Q. How many email addresses do I have? A. As many as I like.
This is really the crux of the issue. To end IP editing would:
* not impact complex/serious vandalism, the type that's really dangerous - someone editing maliciously (as opposed to recklessly) is not going to be stopped by such a low bar as registering an account, even with the email requirement;
* probably reduce drive-by vandalism (the simple, obvious to catch stuff like page blanking), but, and here's the kicker:
* reduce drive-by improvements.
Aaron Swartz analysed a small sample of articles to see how much of the content in each edit persisted into the current version. He found that for each article, of the top 10 contributors by content (not contributors by edits), only around a fifth of them are registered.[1]
Bottom line: registered users make the vast majority of edits (and, as of 2005, the top 1400 make 73% of them) but it's IP editors (and small-time registered users) who contribute most of the content.
-- [1] http://www.aaronsw.com/weblog/whowriteswikipedia
Bottom line: registered users make the vast majority of edits (and, as of 2005, the top 1400 make 73% of them) but it's IP editors (and small-time registered users) who contribute most of the content.
These (and others) are all good points. I've been also wondering if banning anons would be throwing the baby with the bathwater.
However, from my own experience of watching 800 articles daily, most anon edits are bad, and the time I spend now reverting vandalism could actually be used writing things. This is just a sample case of course.
Anyway, I am very sure at this point that the Citizendium people are making a mistake narrowing so much their pool of editors and with extra regulations. Whether that means that the Wikipedia model does not need more editing restrictions will remain to be seen.
Oleg Alexandrov
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
Bottom line: registered users make the vast majority of edits (and, as of 2005, the top 1400 make 73% of them) but it's IP editors (and small-time registered users) who contribute most of the content.
However, from my own experience of watching 800 articles daily, most anon edits are bad, and the time I spend now reverting vandalism could actually be used writing things. This is just a sample case of course.
Your sample is biased, so don't be surprised when the results don't match. The articles people watchlist tend to be more popular and well-established, so they attract more vandals. Try checking the edit histories of 800 randomly-selected articles.
On 4/12/07, Mark Wagner carnildo@gmail.com wrote:
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
Bottom line: registered users make the vast majority of edits (and, as of 2005, the top 1400 make 73% of them) but it's IP editors (and small-time registered users) who contribute most of the content.
However, from my own experience of watching 800 articles daily, most anon edits are bad, and the time I spend now reverting vandalism could actually be used writing things. This is just a sample case of course.
Your sample is biased, so don't be surprised when the results don't match. The articles people watchlist tend to be more popular and well-established, so they attract more vandals. Try checking the edit histories of 800 randomly-selected articles. -- Mark [[User:Carnildo]]
I can only reply to this thread as a whole by repeating something that I've said before, which is that if anonymous editing weren't possible, I for one wouldn't be on Wikipedia today. I'm one of the many who came across an article that contained something that needed fixing, discovered that I could fix it myself, and got involved from there. If I'd had to register an account to fix it, it's highly likely I would have just wandered off.
Newyorkbrad
On 12/04/07, Newyorkbrad (Wikipedia) newyorkbrad@gmail.com wrote:
I can only reply to this thread as a whole by repeating something that I've said before, which is that if anonymous editing weren't possible, I for one wouldn't be on Wikipedia today. I'm one of the many who came across an article that contained something that needed fixing, discovered that I could fix it myself, and got involved from there. If I'd had to register an account to fix it, it's highly likely I would have just wandered off.
I look at it like this:
If registration is required, Wikipedia is no longer "the encyclopaedia that anyone can edit".
James Farrar wrote:
I look at it like this:
If registration is required, Wikipedia is no longer "the encyclopaedia that anyone can edit".
Anyone can edit, as long as they sign up.
Registration is more in line with our principles than the way we ban people, if that's your concern.
Myself, I'm of two minds - I don't consider most anonymous editing to be useful, but I also have a larger number of young adult novels on my watchlist that I'm constantly battling with. At the same time, there's not any real difference between editing with 4 numbers and editing with a completely unrelated nickname, so what's the big deal, I suppose.
-Jeff
On 12/04/07, Jeff Raymond jeff.raymond@internationalhouseofbacon.com wrote:
James Farrar wrote:
I look at it like this:
If registration is required, Wikipedia is no longer "the encyclopaedia that anyone can edit".
Anyone can edit, as long as they sign up.
That is an enormous exception.
On 4/12/07, James Farrar james.farrar@gmail.com wrote:
On 12/04/07, Jeff Raymond jeff.raymond@internationalhouseofbacon.com wrote:
James Farrar wrote:
I look at it like this:
If registration is required, Wikipedia is no longer "the encyclopaedia that anyone can edit".
Anyone can edit, as long as they sign up.
That is an enormous exception.
But *anyone* can sign up.
However, I'm against the proposal as a whole, because it will make vandalism a _lot_ harder to spot.
~~~~
I hope this [1] will defuse any account creation mandate proposals. (For a faster-loading version, see [2])
[1] http://preview.tinyurl.com/22omv3
[2] http://preview.tinyurl.com/2eshda
~~~~
On 4/13/07, Jeff Raymond jeff.raymond@internationalhouseofbacon.com wrote:
Registration is more in line with our principles than the way we ban people, if that's your concern.
I don't see how anything can be more in line with our principle of allowing anyone to edit articles without registering, than allowing anyone to edit articles without registering.
[[m:foundation issues]]
On 4/12/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
That's the whole point I think. If you supply your email address when you make your account you're less likely to vandalize. And if you get blocked you can't start a new account with the same address while the block is on.
Well it *might* work that way. But more likely, I think that by forcing everyone to be registered, we will simply be depriving ourselves of this simple measure of "likelihood of being a vandal". If you notice a trend that drug dealers wear green clothing, then the last thing you want to do is ban green clothing.
Steve
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
However, it is also true that the vast majority of vandalism is committed by anons, and it takes a tremendous energy to police the more than 1.7 milions articles for vandalism.
Most of the *obvious* vandalism is done by IP editors, yes.
But: * nearly 26% of vandalism reversion is also done by IP editors, [1] * vandalism accounts for only 3% of edits (plus or minus 1.7%). [1]
I fear the problem looks much bigger to those people who watch Recentchanges all day.
It also can't be forgotten that the idea that Wikipedia is the encyclopaedia anyone can edit is a fundamental philosophical touchstone of the project, and that's why it's one of the [[m:foundation issues]].
-- [1] http://en.wikipedia.org/wiki/Wikipedia:WikiProject_Vandalism_studies/Study1
On 4/11/07, Stephen Bain stephen.bain@gmail.com wrote:
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
However, it is also true that the vast majority of vandalism is committed by anons, and it takes a tremendous energy to police the more than 1.7 milions articles for vandalism.
Most of the *obvious* vandalism is done by IP editors, yes.
But:
- nearly 26% of vandalism reversion is also done by IP editors, [1]
- vandalism accounts for only 3% of edits (plus or minus 1.7%). [1]
I fear the problem looks much bigger to those people who watch Recentchanges all day.
It also can't be forgotten that the idea that Wikipedia is the encyclopaedia anyone can edit is a fundamental philosophical touchstone of the project, and that's why it's one of the [[m:foundation issues]].
-- [1] http://en.wikipedia.org/wiki/Wikipedia:WikiProject_Vandalism_studies/Study1
-- Stephen Bain stephen.bain@gmail.com
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: http://lists.wikimedia.org/mailman/listinfo/wikien-l
Stephen makes some important observations but what I'm worried about is that recognizing vandalism from registered accounts is a lot harder. We don't have vandal tools that filter out anon edits for nothing.
Mgm
On 4/11/07, MacGyverMagic/Mgm macgyvermagic@gmail.com wrote:
Stephen makes some important observations but what I'm worried about is that recognizing vandalism from registered accounts is a lot harder. We don't have vandal tools that filter out anon edits for nothing.
That's a good point too. Think also of the advantage of having IP information for drive-by IP vandals. We can easily find out if the vandal is coming from a school, or a large proxy, or so on, and modify our strategy accordingly (eg, imposition of range blocks where appropriate, or shorter durations to minimise collateral damage).
My own worry is the not-so-obvious vandalism, and that's why I emphasised "obvious" in my preceding post. I'm not aware of any stats on this, but I would wager that the more complex the vandalism, the more likely it is to be performed by a registered account.
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
I am sure that this came along many times, but I was not able to search for it through the mailing list.
OK, so to start, users without account (anons) contribute a lot of content, at least as typo fixing, etc., and many established users started as anons.
However, it is also true that the vast majority of vandalism is committed by anons, and it takes a tremendous energy to police the more than 1.7 milions articles for vandalism.
Hereby I suggest that only people with account be allowed to edit, and that they also suppy an email address when registering, which is then confirmed by sending an email to the supplied address and having the user clicking on a link.
Before we restrict editing by IP addresses, we should perhaps try to figure out what was accomplished by restricting page creation to users with accounts. As far as I know, it is unknown whether this had any net positive effect, and no other major Wikipedia disallows article creation by IP addresses (which are easier to trace (for non-checkusers) and less anonymous than pseudonymous new accounts). Are there any studies about this that I am not aware of?
Kusma
On 4/11/07, Oleg Alexandrov mathbot@hemlock.knams.wikimedia.org wrote:
Hereby I suggest that only people with account be allowed to edit, and that they also suppy an email address when registering, which is then confirmed by sending an email to the supplied address and having the user clicking on a link.
When we have Single User Login, we should disable unregistered user editing on the higher traffic wikis. I'm not convinced about email registration. In terms of threshold against vandalism, it's trivially breakable (mailinator.net or your own mailserver with an infinite pool of addresses). I prefer captchas for the purpose of avoiding mass registration; spammers & script kiddies actually have to do some real work to break them, and when they do, you can make trivial modifications to make their life harder again. Asymmetrical resource usage=good.
Without SUL, forcing registration is going to cause lots of pain for people coming from other projects for purposes such as interwiki linking or image maintenance. But if one name & password are enough to write-access the entire Wikimedia universe, the incentive for making an account is also higher.
I do believe captcha-driven registration will deter a significant amount of vandalism & spam, but the main reason I consider it a Good Thing is that it builds community & makes people contactable in a persistent way. And, if we target OpenID as a next evolutionary step after SUL, people will be able to use their existing logins from other sites, so it's really a trivial step that could even be incorporated into the edit page.
On 4/12/07, Erik Moeller erik@wikimedia.org wrote:
When we have Single User Login, we should disable unregistered user editing on the higher traffic wikis.
Has this been discussed before? I didn't know it was a possible side-effect of SUL. We should think long and hard before doing this, as it will generate a huge amount of discussion in the media etc.
Are there not more avenues for reducing/repairing vandalism that could be pursued first? Our vandalism detection is still fairly rudimentary.
Steve
On 12/04/07, Steve Bennett stevagewp@gmail.com wrote:
On 4/12/07, Erik Moeller erik@wikimedia.org wrote:
When we have Single User Login, we should disable unregistered user editing on the higher traffic wikis.
Has this been discussed before? I didn't know it was a possible side-effect of SUL. We should think long and hard before doing this, as it will generate a huge amount of discussion in the media etc.
Yes. This is the first I've heard of this being a consequence, let alone a desirable one.
Are there not more avenues for reducing/repairing vandalism that could be pursued first? Our vandalism detection is still fairly rudimentary.
Yep. Unless you want to give every vandalhunter checkuser, 'cos if you get every IP registering they'll need it, much more than they do now.
- d.
On 4/12/07, Steve Bennett stevagewp@gmail.com wrote:
On 4/12/07, Erik Moeller erik@wikimedia.org wrote:
When we have Single User Login, we should disable unregistered user editing on the higher traffic wikis.
Has this been discussed before? I didn't know it was a possible side-effect of SUL. We should think long and hard before doing this, as it will generate a huge amount of discussion in the media etc.
Sure, there's no automatism. I just think it's a good idea--but not, as I said, primarily to deal with vandalism. There are other reasons to ask people to choose a name before editing.