Wikipedia talk:Main Page featured article protection/Archive1

Discussion

I wasn't actually suggesting that; I just wasn't paying attention. El_C 07:17, 18 September 2005 (UTC)

Hehehe [I'm amused that you didn't notice it was on the main page - never thought I'd hear that one :) ]

I didn't mean to imply that you asserted we should protect main page featured articles - I was just pointing to this because it is the clearest explanation of why we shouldn't protect them (because I have to answer that damn question so many times) →Raul654 07:21, 18 September 2005 (UTC)

Heh, I know you didn't. Don't mind my spammage, just a means for me to divert attention away(?) from my incompetence! :D El_C 07:24, 18 September 2005 (UTC)

It's not that unusual to not notice an article is a main page FA; I've done it myself (acted on a request on WP:RFPP without noticing it was on the main page). Some people jump directly into their watchlists or recent changes and do not even look at the main page (if you asked me which is today's FA, I'd have to look before answering). --cesarb 16:06, 20 October 2005 (UTC)

Semi

Any plans to update this to include semi-protection? It doesn't need to be rethought too much, but a mention of the new situation and your thoughts would be helpful. -Mysekurity^{(have you seen this?)} 20:40, 25 December 2005 (UTC)

Done. Raul654 20:51, 25 December 2005 (UTC)

Policy

Why not promote this to policy?

Why not promote the contents of this page into a Wikipedia policy or guideline? I think perhaps it belongs in the Wikipedia namespace instead of a user namespace for both authenticity and visibility reasons. As it is a summary of other established policies anyway it shouldn't be too controversial. -- stillnotelf has a talk page 03:56, 30 January 2006 (UTC)

Well, I never envisioned it as a policy; I was just really, really, really tired of typing the same response over and over again :) Raul654 04:24, 30 January 2006 (UTC)

Making this a policy is even more absurd than not protecting the main page FA in the first place.Rlevse 02:36, 7 February 2007 (UTC)

Policy

So, this is a policy? It's being treated like one, and I've been told it's one. WikiFanatic 06:36, 10 February 2006 (UTC)

It seems to be a condensation and restatement of existing policy. So, if it's not technically policy, it agrees with policy in every way. -- stillnotelf has a talk page 03:25, 11 February 2006 (UTC)

Exactly - it's not policy, as much as it is an amalgamation of policy. Raul654 03:28, 11 February 2006 (UTC)

• Pretty much. Because some new users didn't seem to understand that, I've moved this page to mainspace and flagged it as such. Comments welcome, and if you can think of a better title please do so. (Radiant) 14:53, 15 November 2006 (UTC)

I see this as a memorialization of long-held practice of the community and with significant community assention to its validity ∴ a policy :) --Trödel 20:21, 16 November 2006 (UTC)

• I agree. Because this page is an isolated and codified snippet of existing policy, this page should be policy as well. --causa sui ^talk 20:50, 10 February 2007 (UTC)

Rewording

Now this is in the policy space, I've reworded it a bit to be more policy and less essay-like. I think I've stuck to the contents pretty much, but if anyone disagrees with the rewording I'd be happy to discuss here ;) --Robdurbar 23:20, 15 November 2006 (UTC)

• Looks reasonable so far. Thanks. (Radiant) 09:40, 16 November 2006 (UTC)

Dissent

I think this was promoted to policy without enough input from the community. --evrik ^(talk) 19:35, 16 November 2006 (UTC)

• I agree that there has not been enough input from the community, yet I support the promotion none the less. KOS | talk 05:23, 17 November 2006 (UTC)

It could be argued that the fact that many were treating it as policy means that there was an implicit conesnsus. --Robdurbar 09:40, 17 November 2006 (UTC)

• Indeed. It has been existing practice for years now. I would be interested to see instances where the Main Page FA was protected, and of the likely-ensuing debate. (Radiant) 09:53, 17 November 2006 (UTC)

• Well really the only user's who could put this policy (remember it wasn't always policy) into effect were admins. I'm fairly certain that there are more editors without the extra tools than editors with them. Still as stated above, I do support this being promoted to policy. My point is that though it may have been accepted practice for years now, and that it has been treated as policy does not mean that it has implicit consensus, since only admins have been able to enforce this policy. KOS | talk 00:02, 18 November 2006 (UTC) If this comment is unclear just let me know and I'll try and clear up anything that is confusing about it. :) KOS | talk 00:02, 18 November 2006 (UTC)

One only needs to look at the plight of today's featured article, San Francisco, California, to see to folly of this policy. One of the many times that the page got replaced with a homophobic vulgarity did not get reverted for seven minutes.--DaveOinSF 19:03, 17 November 2006 (UTC)

But the policy doesn't state 'never semi-protect'. For example, look at between about 23:07 and 23:37 on 17 November - those levels of repeated vandalism from a number of users can be adequate for semi protection. But the idea, or spirit of this policy, would then encourage such protection to be in place for say, 30 minutes or an hour (as it was that example was the end of the day, but ignore that for now). But over the last few weeks, semi-protection has been left in place for 5 or 6 hours... this is what this policy wants to prevent. --Robdurbar 00:35, 18 November 2006 (UTC)

So one clueless admin unprotects San Francisco and two minutes later the vandalism recommences, with the vulgarities replacing the main page five times in the next six minutes before a separate admin steps in. Another clueless admin unprotects again, only for the vandalism to resume immediately and trash the page 13 times in 11 minutes, leaving it to another administrator to clean up the mess. If an administrator is going to unprotect a page, he or she is OBLIGATED to 1. learn why it was protected in the first place and 2. stay around to ensure that the vandalism does not immediately resume. Otherwise, they are being irresponsible.

I do not know how many people come to Wikipedia, click on the link Today's Featured Article and were instead greeted with a vulgarity. Blind adherence to policy by clueless administrators unconcerned as to the consequences of their actions should be punished.--DaveOinSF 06:43, 18 November 2006 (UTC)

Well San Fran had an usually high level of vandalism. But, let's look at what happened, shall we? Knowledge of Self's protection at 08:04 was fully justified - it should have just been lifted more quickly. After I unprotected, there were no vandalising edits for well over half an hour, and only 4 within 2 hours - hardly the picture that you're painting. Vandalism maintained at a steady, but manageable, level till the evening when Glen S quite rightly protected it. After that, the page was twice unprotected and then quickly reprotected when the vandals were still about.

All I can say is that the policy was followed perfectly - it's intended to be used like that. Its just unfortunate that by the nature of the topic, San Fran was unusally open to vandalism. Robdurbar 11:20, 18 November 2006 (UTC)

Pretty low standard for perfectly here. THe admins who unprotected then didn't pay any attention to the page, even though the vandalism recommenced almost immediately, and allowing the vandalism to take over far longer than it would have. They acted irresponsibly. And by the time someone else came in, the damage was done - complaints on the talk page about the vulgarity and who knows how many potential visitors turned off to Wikipedia. We should strive to do better.--DaveOinSF 18:02, 18 November 2006 (UTC)

• The same thing happened to Eagle Scout, more than 250 edits in a day, most of them vandalism and reverts.. --evrik ^(talk) 16:03, 18 November 2006 (UTC)

I don't like the language forbidding full protection of the FA... it's too binding for Wikipedia. It's not too hard to think of ways a premedidated attack could compromise even a semi-protected article, and there certainly could be (and at least once, has been) a situation where even strong adherents to "Don't protect the FA" end up full protecting the FA. The language is too binding and could lead to the only admin around not dealing with a serious attack on the FA correctly, because the page says "under no circumstances"... even though there are at least some circumstances. --W.marsh 21:55, 18 November 2006 (UTC)

• People who object to the wording, or believe they can improve it, are of course free to edit the page. (Radiant) 23:05, 18 November 2006 (UTC)
  • You don't say! I just thought I'd discuss it first... --W.marsh 23:24, 18 November 2006 (UTC)

Proposed changes to this policy

A few proposed changes to the policy. Remove emphasis on "NEVER" and include administrators' responsibilities when deciding to protect the featured article or when deciding to remove protection. Removed text is stricken, added text is underlined.--DaveOinSF 20:22, 19 November 2006 (UTC)

Policy

Wikipedia's Main Page featured article is one of the most visible and heavily edited on the site. For this reason, it receives a lot of vandal edits from unregistered users visiting Wikipedia.

It has been suggested many times in the past that the featured article should thus be protected or semi-protected. However, Administrators are advised never to use extreme discretion when deciding whether to protect, semi-protect, or unprotect this page. and to only semi-protect it under certain extreme conditions.

Rationale

There are several reasons for this policy.

• Almost without exception, featured articles are improved by their time on the main page (some of them greatly improved). Check out these before-and-after diffs from September, 2005: [1][2][3]. Protecting the featured articles means that these pages may not be improved.
• A featured article is supposed to "exemplify our very best work, representing Wikipedia's unique qualities on the Internet". This includes being editable by anyone. Visitors often tend to look at our most visible articles, and having those articles editable helps attract new users to the project.
• Vandalism (especially to highly visible articles like the main page featured article) is cleaned up very quickly, often in only a matter of seconds, helped by automated bots such as Tawkerbot2.
• Although the more visible featured article of the day attracts more vandals than other articles, it also attracts more curious and good faith editors. A lot of vandalism on the day's featured article is reverted by other anonymous or recently-registered users (e.g. [4]),
• This is codified in the page protection policy: When a page is particularly high profile, either because it is linked off the main page, or because it has recently received a prominent link from offsite, it will often become a target for vandalism. It is best not to protect pages in this case. Instead, consider adding them to your watchlist, and reverting vandalism yourself. - Wikipedia:Protection policy
• The featured article receives many page views from non-editing readers as well. It is damaging to Wikipedia to encourage new readers to visit a page that is under a severe vandalism attack.

Protection

Protection prevents anyone without administrative powers from editing an article. This should almost never occur on the day's featured article, and should only be used in rare situations where semi-protection is ineffective.

Semi-Protection

Semi-protection prevents all unregistered or recently registered users from editing a page. The main page featured article should also almost never be semi-protected in only rare situations. However, it is recognised that there are some extreme circumstances in which semi-protection can be introduced for a limited amount of time. This could occur when, for example, a range of dynamic IP addresses are being used to vandalise the featured article page in quick succession; where personal information or potentially distressing content is being repeatedly placed onto the article; or where a few minutes of protection are needed to remove harmful vandalism from a page.

Semi-protection can be introduced for a limited amount of time; it is preferable to give vandalising users a brief block, rather than semi-protect the day's featured article.

Notification

Should an administrator deem that protection or semi-protection of the Main Page featured article is necessary, a notice should be placed at the Wikipedia:Administrators' noticeboard and the page's talk page as to the reason and rationale behind the decision to protect or semi-protect, and whether there are any recommended steps for the unprotection of the page.

Unprotection

Because protection or semi-protection is not a step that is taken lightly, administrators should use caution when they discover that the day's featured article has been protected or semi-protected. Such an administrator should check the page history, the talk page and the Wikipedia:Administrators' noticeboard before deciding whether to unprotect the page. Should the administrator decide to unprotect the page, she or he must continue to monitor the page for a resumption of vandal attacks for a minimum of 30 minutes following the decision to unprotect the article. is encouraged to continue to monitor the page for an immediate resumption of vandal attacks.

Move Protection

To qualify for featured article status, the day's featured article will be at a stable and agreed-upon title. Therefore, in the event of page move vandalism, it is acceptable to protect the article from being moved. For housekeeping and process reasons, this protection should be lifted at the end of an article's stay on the front page

Other front page articles

These are covered under the semi-protection policy. Although they can be semi-protected, admins should generally be more cautious in applying protection to these pages. To qualify for semi-protection, front page linked articles should have a higher frequency of vandalism than other articles need.

Responses to proposed change

• Well the clause about having to monitor with it for half an hour is unworkable. Firstly, its unprovable and slightly bizarre - I unprotect the FA, the phone rings and what - I ignore it? I appreciate what it's trying to do, but it could bring up all sorts of accusations. For example, I know that I've tried 'patrolling' the FA after I've unprotected it to avoid such accusations; but I have a slow internet connection and have been 'beaten to it' almost every time. It's instruction keep; perhaps retain the sentence about giving it thought though.

• I like the notification process idea

--Robdurbar 21:56, 19 November 2006 (UTC)

I oppose the proposed changes - we can trust the admins to protect when there is a special situation and do not need to outline it --Trödel 22:08, 20 November 2006 (UTC)

I think the request to drop a line at ANI is a good idea. The other changes are pretty much semantics. --Robdurbar 23:13, 20 November 2006 (UTC)

True, but I think the semantics matters here in at least one case - we want to clearly favor semi-protect, not just carefully choose semi-protect or protect - --Trödel 23:20, 20 November 2006 (UTC)

Well, the intent was not to equate protection and semi-protection, but to remove emphasis on the word "NEVER", promote greater communication, and to encourage caution for admins who might want to unprotect a protected article. Obviously, I've failed in that regard, and will give it some thought as to how to better do this.--DaveOinSF 00:20, 21 November 2006 (UTC)

Sounds good thx!!!--Trödel 22:52, 21 November 2006 (UTC)

September 2005 diffs

I'm tempted to call "b---s---" on this (which is both the current and proposed policy): Check out these before-and-after diffs from September, 2005: [5][6][7]. Why bs? Because of those three edits, two were done by registered users. A semi-protect policy would have allowed those two edits to be made. No editor (I think) is arguing that an FA should be fully protected, so why offer evidence of how great the current policy is by citing two examples that would have happened even with semi-protect automatically in place?

And as for having those articles editable helps attract new users to the project, exactly what evidence exists for that? Why not argue the reverse - that permiting anonymous editors to vandalize FAs gives them their first taste of how much "fun" they can have on wikipedia, and encourages them to vandalize other articles? John Broughton | Talk 02:11, 20 November 2006 (UTC)

Because of those three edits, two were done by registered users. - wrong. Look at the timing on the diffs - they cover the entire 24 hours the articles were on the main page. So they encompass all anon and non-anon edits for the day those articles were on the main page. (Side note - perhaps this should be made more clear in the description)

Also, notice the dates on those diffs were the 3 days immediately prior to me writing that particular statement. I did this intentionally to avoid accusations I cherry-picked the examples. Raul654 02:18, 20 November 2006 (UTC)

I stand corrected, but remain convinced that these examples are being misused as a "evidence" of anything important. The diffs cover 5+ days, 2+ days, and 9+ days, respectively. What they "prove" is that a mixture of registered and anonymous users, over a two to nine day period, can improve a FA. Who would disagree? (Why not do diffs for just the 24 hours that the articles were on the front page, rather than for multiple days?)

The argument isn't - or shouldn't be - about whether an FA should be FULLY protected or not, it's about whether it should be SEMI-protected. For the purpose of that argument the three examples prove exactly nothing. There could be 100 anonymous vandal edits and no constructive anonymous edits (other than reverts/fixes), for all anyone knows, based on those diffs. What is needed (in my mind) to prove that anything would be lost by blocking anonymous diffs would be examples of specific anony edits that really added value.

So if the policy states "FAs should not be FULLY protected because their their time on the front page can result in them being improved", sure, that's almost certainly true. But who is arguing for full protection? What the policy now says I still think is BS: it says (implicitly) "FAs should be NEITHER fully nor semi-protected because - look - here are three examples". Yet the examples say NOTHING about semi-protection, because they fail to distinguish between contributions from registered users and unregistered ones. John Broughton | Talk 22:08, 21 November 2006 (UTC)

Really, what you've brought up is that the 'rationale' section doesn't state anything about semi-protection, which is a weakness. --Robdurbar 15:15, 26 November 2006 (UTC)

If you're saying that the policy wording predates the implementation of the option of semi-protection, then the wording is quite understandable, and the example would have made sense at the time. But that was then, and this is now, and arguments against "protection" are now being taken - rightly or wrongly - as arguments against BOTH semi- and full protection. The proposed change to the policy (above), in its rationale section, does not separate out arguments for full protection versus arguments for only semi-protection. I think that's a serious flaw, as discussed in the Redux section, below, which might be a good place to continue this discussion (assuming no one wants to dispute my comments on the September 2005 diffs). John Broughton | Talk 15:33, 26 November 2006 (UTC)

Semi-Protection Redux

Having looked over the edits for today's Featured Article (Duke University) - I have to wonder about the rationale behind not semi-protecting the daily FA, at least for that day. Looking over the edits so far, rather than encouraging new editors to make constructive changes to the page, all that has happened is a rapid-fire string of vandalism and reversion, the latter sometimes involving the appropriate warnings and sometimes not (I have tried to supplement the nots). I agree with John Broughton above; I feel like the time and effort of the bots and the human editors could be better spent building an encyclopedia rather than protecting a high-profile page from anonymous vandals for just one day, and that new editors would get more out of seeing a featured article in the form that made it an FA rather than getting an introduction to WP editing that shows how easily editors' hard work can be temporarily defaced by vandals. I suppose I have just makred myself as a protectionist...whoah. DukeEGR93 14:14, 26 November 2006 (UTC)

Well, the vandalism in FA's is making me crazy. Every minute there are vandals, and people can't revert them. ATM The FA of the day's footnotes are screwed, and it has been like that for 20 minutes. People have reverted it after, but a lot of vandalism gets through. Then the reason for not to protect is that because few good edits by IPs have been dugged, from the millions of vandalism edits. To be honest. If there are so obvious fixes, they should have been fixed before making FA, or FA of the day in my opinion. --Pudeo (Talk) 15:37, 27 November 2006 (UTC)

I'm beginning to think that some hard data (number of IP edits of front page article in (say) a 24-hour period, number of clear vandal edits, amount of time before reversion, number of constructive edits, categorization, etc.) might add some light to this discussion. Thoughts of others? John Broughton | Talk 16:51, 27 November 2006 (UTC)

Well the page could be protected - indeed should be - to remove the vandalism if its been stuck on the page and can't be reverted. Then, when it's gone, unprotect. Simple 5/10 minute protections for those sort of problems are allowed by the policy. But the levels of vandalism to the Duke page really wern't that high for main page FA.

As for John's comments - this would smack of instruction creep. Do you mean that this page should stipulate a certain level of vandalism needed to require semi-pro? I don't think that would go down well. Robdurbar 17:16, 27 November 2006 (UTC)

Apologies for any lack of clarity: I wasn't proposing to come up with a number/threshold for when a page should or should not be protected. I simply want to address the question: How much do anonymous IP edits of FA articles help, and how much do they hurt? ("Help" means improving the article; "hurt" means taking up editor time and presenting a vandalized article to other readers.) Let's suppose (for example) that of 100 anon edits, 20 were neutral, 40 helpful, and 40 vandalizing, and the total elapsed time for serious vandalism (however defined) was 1% of the viewing time in the period analyzed. Then most people might say that semi-protection of the article would have been a bad idea. On the other hand, suppose the numbers were 10/5/85, and 5% of viewers saw a significantly damaged page. That might be considered a strong argument that the policy against semiprotection (in most cases) should be revisted. John Broughton | Talk 23:04, 27 November 2006 (UTC)

I agree that some hard statistics would be very helpful. The policy now states as fact that FA articles featured on the main page benefit from edits from IP and recently-registered editors. Like John, I suspect that the number of cases where this is true is far outweighed by vandalizing edits from IP and new editors. The policy states that any semi-protection of FA articles will end up discouraging new editors, and as existing editors lose interest, eventually there will be no one left. This is stretching things quite a bit, as there are about 1,300 non-FA articles for each FA one, and the ratio for FA articles on the main page to all articles is currently about 1,500,000 to 1. There is ample opportunity here to recruit new editing talent without relying on main-page featured articles.--Paul 23:36, 27 November 2006 (UTC)

This "policy" (how on earth did it get to be a policy?) is very wrongheaded. FAs are supposed to be essentially done—not to say they shouldn't be freely editable under ordinary circumstances, because there is usually still some potential for improvement, but we should be assuming that they are at a very high level of quality where they do not greatly need whatever random contributions come through that day, and changes can still be suggested on talk or simply performed a day later. I am not at all convinced that there is a general benefit to having FAs unprotected as a matter of common practice, and to have such a strict policy that basically makes protection impossible is in my opinion absurd. Semiprotection should be applied to FAs of the day automatically. Not only will it eliminate most of the vandalism, it will give a very visible impression of control that our naysayers are always saying we lack. Everyking 11:59, 28 November 2006 (UTC)

Whaaaaaa???

From rationales:

The featured article receives many page views from non-editing readers as well. It is damaging to Wikipedia to encourage new readers to visit a page that displays vulgar or offensive vandalism.

That sounds like a strong reason not to have this as a policy. Why is it on there in support of the policy? Grandmasterka 03:25, 28 November 2006 (UTC)

It's meant to justify occaisional semi-protection. --Robdurbar 08:42, 28 November 2006 (UTC)

Occaisional semi-protection is very different from the other points which advocate no protection whatsoever; being grouped with the other points doesn't make sense. "Whaaaaa???" indeed! The other points under Rationale is for "Don't protect Main Page featured articles". This point should be under it a separate heading. Goose 16:01, 30 November 2006 (UTC)

Suggestion

Gathering some numbers

This may already be recorded somewhere, but I suggest we get before and after diffs for every single FA of the day and post them somewhere. This would enable us to systematically review the trends and would help prevent cherry-picking of evidence. Everyking 10:37, 1 December 2006 (UTC)

I'd be happy in taking part in some sort of analysis - say all the FAs in December? Or over a shorter period? --Robdurbar 10:54, 1 December 2006 (UTC)

Ok, I'm gonna try and create something at December Main Page FA analysis. All should feel free to participate. --Robdurbar 11:23, 1 December 2006 (UTC)

Thanks for doing this. A great start. Three concerns:

• I strongly recommend breaking up each day's edits into (say) four hour blocks, for the purpose of analyzing IP edits (but not assessing the change in the article). I recommend this because (a) it's pretty daunting for any individual to take on an entire day's edits (hundreds?); splitting the day into blocks encourages people to volunteer to take a smaller chunk of the work, and (b) it's a lot easier to check someone else's results if the results are only for part of a day, not a whole day.
• Have you considered doing edit COUNTS rather than edit PERCENTAGES? Counts are easier to do (one less step); where there is a block, the impact is easier to assess; differences in counts (other than for blocks) may be useful to know; and, of course, percentages can always be calculated after the fact
• The MOST important thing is to count the right things. If this doesn't happen, then it's possilbe that most of the effort will be wasted. Here are the five things that could be counted with regard to IP edits:

1. Edits that are vandalizing
2. Edits that revert vandalizing by another IP
3. Edits that revert vandalizing by a registered user
4. Edits that are beneficial, other than reverts
5. Edits that are in good faith, but are not reverts and are not considered beneficial

You've combined 1,2, and 3 into one column; 2, 3 and 4 into another column; and 2 through 5 into a third column. Let's run a scenario: suppose that there 100 IP vandalizing edits, 50 reverts by IPs to such vandalization, 5 reverts by IPs to registered user vandalization, 10 beneficial edits other than reverts, and 40 good faith edits that are neither reverts nor beneficial. Then your columns would show: 160 vandal-related edits, 65 beneficial edits, and 105 good faith edits. What would one conclude from those three numbers, NOT knowing the detail?

By contrast, in this scenario, knowing the detail (all five categories) would allow this conclusion: if all IP edits were blocked, there would 100 fewer vandalizing edits, 5 more reverts that would need to be done by registered users to fix vandalization by other registered users, and 10 fewer beneficial (content-related) edits that improved the contents of the article. The second set of conclusions makes clearer, I think, what the real impact of semi-protection would be (given these five hypothetical underlying numbers), which is, I think, the entire purpose of spending the time for this analysis. So I really think you should reevaluate what people should count. [And if that is not convincing, try this: suppose the numbers were 100, 50, 0, 0, and 40. Then IP editing would clearly be totally without value'. But the existing three columns would show: 150 vandal-related edits, 50 beneficial edits, and 90 good-faith edits -- and that could certainly be interpreted as it being useful for IP edits to continue, even though - with these admittedly made-up numbers - such a conclusion would be absolutely false.] John Broughton | Talk 18:17, 1 December 2006 (UTC)

Some reasonable points. As you note, its important to keep this doable and cut down the amount of work, in order for people to take part. I would propose, then, that 2,3,4 remain combined as all supporting ip editing. If we take 2 away from 1 then we have a clear figure for how much direct 'damage' ips are causing; we can then also isolate 5 and 2. I do see your point that an IP edit reverting an IP vandal could be worth counting seperatly from other beneficial ip edits as basically being IP-neutral; but, again, I was trying to keep the number of measurers down so that we can encourage people to take part. As a test, though, I'll try doing 1st December FA now and see how long it takes and I might get back to you! Robdurbar 00:36, 2 December 2006 (UTC)

This is a great analysis you're doing. The latest FA that's been analyzed, Battle of Austerlitz, apparently spent a total of an hour and 21 minutes in a vandalized state during the 24 period. To me, this really contradicts the claims about vandal fighting being so instantly effective that the risk of readers seeing vandalized versions is negligible. Furthermore there was no significant editing aside from some rewording. It'll be interesting to see what the comparative results are for the upcoming FAs. Everyking 01:02, 4 December 2006 (UTC)

That's more time than I'd have thought... it would be interesting to know how many people viewed the FA that day, then we could know how many people saw a vandalized article (probably in the thousands). --W.marsh 01:44, 4 December 2006 (UTC)

It is proving interesting. Let's remember when reading this that some of the logic behind this policy is due to the ideology of Wikipedia, as well as the practicalities of the article. Looking at next two FA's, we've got Down Syndrome and "Weird Al" Yankovic; I would expect to see a lot more vandalism on these... --Robdurbar 10:39, 4 December 2006 (UTC)

Pros and cons

Ideologically speaking, I always felt that the goal of Wikipedia was to create a great encyclopedia, not just to allow anons to edit. Historically, anon editting has been more or less helpful, but we should always remember what we're actually here to do... it seems quite misguided when people act like the purpose of Wikipedia is simply to allow anonymous editting and fight the resulting vandals. --W.marsh 16:40, 4 December 2006 (UTC)

Anon editing is undoubtedly a good thing for the project as a whole and I think it would be very bad to see any result from this analysis used to condemn anon editing in general. My concern is specifically with the circumstances of FAs of the day—the existing high quality of the page, reducing the need for further edits, and the visibility of the page, which has the duel effect of attracting very high numbers of vandals as well as very high numbers of ordinary readers who will be turned off to some degree by the vandals. I don't feel that the quality issue or the visibility issue would be sufficient to justify automatic semiprotection on its own, but being together in these instances, I think there's a strong case for it. Everyking 07:04, 5 December 2006 (UTC)

Well I'm undecided about semi-protection of the main page and semi-protection in general, but I believe we should be openminded and not assume anything is totally off limits, like assuming we have to forever allow anon edits everywhere no matter what... that's just dogmatic. Of course I do understand that we're just talking about the main page FA here. It occurs to me that whenever the main page FA is getting heavilly vandalized, we usually see anons/new users begging for admins to protect the article... but when it's protected, we never see them begging for unprotection (except ones that seem like obvious vandals). I wonder how valid the logic is anymore that readers are more upset when they see the main page FA protected than when they see it horrifically vandalized. And how much damage we do by operating on an assumption that seems to be more in the minds of a few admins than readers at large. --W.marsh 16:01, 5 December 2006 (UTC)

More on the statistics being gathered, and other analyses (data)

The column titled "IP Edits Beneficial", as I understand it, includes vandalism reverts by IP editors. Since these edits would not have been necessary were the article not vandalized, it's not really a measure of a substantive edit improving the quality of the article. Furthermore, as I understand it, semi-protection blocks not just IP editors but also newly-registered users, such as the ones which attacked San Francisco two weeks ago. The value of semi-protection would thust not be counted by the current methodology. I'll try to contribute to this project by adding in some statistics that take this into account later tonight.--DaveOinSF 19:49, 4 December 2006 (UTC)

Yeah, the new user one has come up. The main reason for not including new users in the survey was just a time thing - it would (in theory) mean checking the history of every single registered editor. If there is a substantial amount of new user vandalism, that can be mentioned in the summary.

As for the lumping together of beneficial with vandal-correcting edits; I do see the point in that too, but again we've been trying to keep down the number of things that are counted seperatly, to encourage people to participate. --Robdurbar 20:15, 4 December 2006 (UTC)

Case study... Down syndrome is getting heavy vandalism and has been semi-protected twice, and it hasn't even been on the main page 2 hours yet. --W.marsh 01:47, 5 December 2006 (UTC)

After the San Francisco nonsense, I started a short analysis of this same subject here using a few randomly selected articles in the last six months. My original intention was to concentrate more on who added substantive content rather than how much vandalism was done and by whom, figuring that there is some balancing point at which the heightened level of vandalism is tolerable if substantive improvements are routinely made by anons. I haven't followed this through much, but I had the thought that using something analogous to Voice of All's old RfA edit analysis reports would be useful, as it seemed from what I recall to be pretty good at identifying major edits, manual reverts, etc. but would be less time-intensive to compile. Opabinia regalis 05:01, 5 December 2006 (UTC)

I note that your review of six FAs found only one case where an anon editor seemed to make any significant contribution.

Re VoA's edit analysis, could you point to a specific page where one appears? My sense is that only a tool that sorted and listed edits (for subsequent editor review) would be userful; any kind of summary of edits without listed details (links) is going to mix registered user edits with IP edits, and the focus here really is IP edits (because, I think, no one is seriously proposing full protection of FAs).

That's correct, over six FAs I found one anon edit, adding one sentence, that added to the article. To be fair, I didn't count things like typo/punctuation fixes and interwiki links that are productive but minor changes easily delayed for a day, or easily done by a registered editor.

Here is a random RfA I found containing VoA's output. Obviously it would need to be changed to scan edits over a time period rather than a contributions list; sorting IPs vs accounts shouldn't be so bad (since account names aren't allowed to mimic IPs) but erroneous edit classifications could be a problem, and the lack of granularity compared to the manual analyses you guys are doing. Opabinia regalis 04:11, 6 December 2006 (UTC)

Down syndrome may provide a point of examining another issue - while the vandalism is usually fairly obvious and fixable, how much of the text is deteriorating away from well-referenced medical consensus, and how much of that is from anons vs. registered users ? Are editing changes, for example, introducing problems in accuracy, POV or OR?. Sandy (Talk) 17:05, 5 December 2006 (UTC)

It's really difficult to categorize specific edits as other than beneficial/not (and even that is somewhat subjective); futher categorization (OR, inaccuracy, etc.) are really difficult, requiring an expert, I think, which most of us are not. I think we want to do some sort of analysis that is going to stand up to scrutiny by everyone; subjectivity becomes a real problem under such circumstances. Also, in general, it's best to do an initial (really good) analysis and then, if that provides interesting results, do further study of points that weren't covered or might lead to further useful conclusions. John Broughton | Talk 17:47, 5 December 2006 (UTC)

It will be very interested to see the stats for Down syndrome. The 24hr diffs show very little for all the effort put into reverting the hundreds of vandalisms (largely reverted by dedicated people like Sandy, not bots) Over 530 edits has achieved a little bit of copy-editing, which as far as I can see, was done by a few established registered users. Of the anon/newbie edits, I haven't spotted any that weren't vandalism - apart from a few attempts at reverting which almost always were incomplete. Surely more harm than good was done by providing a prominent graffiti wall to display the worst of human nature. Colin°^Talk 00:19, 6 December 2006 (UTC)

I'm more concerned about how far it has moved from an article reflecting peer-reviewed, medical consensus (also, most of the substantive improvements were done by us MEDMOS regulars, as several of us weren't around when it went through FAC - I was traveling, or some of those references wouldn't have gotten by me :-) We need a new review of the article now. It was fairly absurd for so many of us to have sit there and revert several vandals per minute, and vandalism of the most offensive type. Sandy (Talk) 00:26, 6 December 2006 (UTC)

Agreed; users should have been blocked or the page could have been semi-ed sooner. Though its early days at the moment, when 'counting the numbers' I think a fairly common sense conclusion is coming: when pages that are obvious vandal targets such as Down Syndrome or Wierd Al are on the main page, there'll be lots of rude vandalism all day. When its 'Extratropical Cyclones' then there really won't be that much at all and most of it won't be so bad. --Robdurbar 19:14, 6 December 2006 (UTC)

VoA has kindly posted a breakdown of the edits to Down syndrome over the 24-hour period in which was on the main page; see the results here. The formatting could use some work, but I think it's a useful first look - would be even more useful to separately produce the percentages for anon and registered edits. Opabinia regalis 00:31, 7 December 2006 (UTC)

More discussion on pros and cons

Part 1 (arbitratry subheading)

Anon users are important to wikipedia- and it is important that anons know that wikipedia is freely editable, that they will be attracted to the project. People who say we have enough contributors haven't been working outside WP's pop culture and computer science articles- in the humanities, social sciences, most sciences, and on articles about non-English-speaking countries WP is still very deficient. In this way, test edits are a good thing, because they acquaint people to the concept of wikipedia that they may contribute and share their knowledge. Secondly, persistent vandals will find an article to vandalize- it is better that they do it in an article that is receiving a lot of regular-editor attention, rather than a back-of-the-woods article that isn't heavily scrutinized, where misinformation can remain for weeks or even more. This policy should stand.Borisblue 04:10, 6 December 2006 (UTC)

There are sensible-sounding but ungrounded arguments to be made in both directions; I'd rather have some empirical data first before deciding which way we want the evidence to point. It's just as sensible to say that expert editors in the traditional academic disciplines will be turned off by finding what is supposedly an example of Wikipedia's best work in a vandalized state, particularly if it's the usual juvenile penis pictures/'John Smith is gay'/replacing the text with 1000 copies of the word 'fuck'/etc. I sometimes poke in on Wikipedia at work but I don't look at the main page FA to avoid exactly that. Opabinia regalis 04:57, 6 December 2006 (UTC)

A lot of the costs/benefits are unexaminable by statistics. How do you quantify Wikipedia losing reputation from users viewing vandalised articles? Conversely, how do you quantify te number of dedicated users who got introduced to wikipedia by making an edit on the FA as an anon? Borisblue 05:16, 6 December 2006 (UTC)

The statistics can help examine the extent of the problem, but the costs and benefits of front-page FA editing will likely be gray enough that people will form quite different conclusions based on the stats. Based on their... existing philosophy. So, I suggest that this is a philosophical issue. As I wrote on the Down Syndrome talk page, 'It's almost a form of wiki-[adoration] when editors promulgate the "anyone can edit" party line over the ostensible mission of creating an encyclopedic user experience. Which is the higher value?'. I find it farcical to present our highest-quality articles on the front page—to illustrate to visitors the benefit of Wikipedia—and risk having the "average reader" see a horribly vandalized article. (We're talking about what is now a very highly visited web page.) The average reader can't be expected to know what's going on, hit History, and view the previous revision. A good dose of end-user orientation and some concern with the marketing of this endeavor lead to the sound conclusion that the front-page FA is an exceptional circumstance that needs special treatment. I'm not that interested that the article improves by being on the front page—it has, after all, gone through a significant process to prove that it is already "excellent". Finally, on a lighter note, isn't it appropriate for an article that has risen to such heights to "take a day off", and let visitors just take it in, as they would a painting? On the front-page day, wiki rested. –Outriggr § 05:01, 6 December 2006 (UTC)

Why can't users visit Britannica or Encarta if they are that put off by this kind of thing? What makes wikipedia unique is that users can participate- we have to encourage anon, new users participating in the project. It was this philosophy that has gotten us where we are, and we still need new participation if we are to improve on our flaws in coverage. The hidden cost of lost anon participation is something that we have to take into account when restricting anon access to the main page. Borisblue 05:16, 6 December 2006 (UTC)

Well, I don't really understand the question... those aren't free encyclopedias (to the same extent, at least). You also seem to be assuming an "omniscient" user who knows all his options. My point and my concern are that users are just users. They don't start out with grand ideas of contributing to an open-source encyclopedia, or knowing what vandalism is. To start with (and by definition, the front page is a staging area), they want to read an article of interest. If it's good, they'll come back. Gradually, they become interested in contributing. Vandalism of the front-page FA will ruin this experience for some statistical portion of users. So I see it just the opposite: not protecting the article is discouraging users, who are the future's potential contributors. –Outriggr § 05:26, 6 December 2006 (UTC)

Which would discourage user participation more? Coming across vandalism, or being outright prevented from participating? Borisblue 05:39, 6 December 2006 (UTC)

You've successfully framed the question in such a way that only the answer you previously advocated is possible. Obviously "being prevented from participating" discourages participation. That's not quite the right question. "What's more likely to discourage readers from becoming productive editors?" unfortunately does not have such a simple answer. Opabinia regalis 06:27, 6 December 2006 (UTC)

I am perhaps relying on anecdotal experience, but I have had anon vandals whom I warned for vandalism become involved enough to become administrators. Seeing a vandalized page will discourage a user from relying on wikipedia as a reference (in many contexts this is a good thing) but I find it difficult to see how this will prevent a user from eventually becoming a contributor. Giving anons an opportunity to fix FA vandalism means they are participating meaningfully in wikipedia. Even going on the talk page noting vandalism is a meaningful contribution to wikipedia, and means that an editor is participating in the process, making it far more likely that (s)he will contribute in the long run. The first step for being a productive editor is editing, not being impressed with a fancy article. Rather, if an anon user only sees polished, unblemished parts of wikipedia, it is not unreasonable for him or her to conclude that he has nothing worth contributing.

Furthermore, I don't see any benefit in whitewashing wikipedia's vandalism problem. This is something that will always be with us, I don't see why it is beneficial to present an illusion that vandalism doesn't happen. Borisblue 06:49, 6 December 2006 (UTC)

Further thoughts (pre 06:49)... I'm not sure that question needs an answer. To me, it's a straw man, and the degree of "encouragement/discouragement" isn't at the heart of the matter. I see a lot of people focusing on editing. Before one is an editor (chronologically), one is a user, and the user's initial (passive) experiences with the project are surely important. When WP says "this is what we're featuring", make sure it leaves a good impression. Nevertheless, the proposal is to limit editing for one article out of 6,905,938. Can anyone honestly say that's "discouraging"? –Outriggr § 06:54, 6 December 2006 (UTC)

Well, a featured article receives, say 100 test edits. Either the user then goes on to make a test edit/vandalism in another page, (which would then mean it takes longer for it to be discovered and reverted) or the user would not make a test edit at all, which would thus tragically deprive him of the opportunity to experiment with, and be introduced to the project. In terms of "putting up a good impression"- isn't it better to put up an accurate impression of what wikipedia is about? The message I want a curious new user to get it "this project is awesome, but has many holes that you can help fix- so please do!".Borisblue 07:04, 6 December 2006 (UTC)

Let's keep the "allow anons to edit" discussion for a different time/place please. I think we all agree that there aren't enough editors and that there is good evidence that anons can become productive editors (though the anon-vandal to admin story is getting a little tired - and not statistically significant I'm sure). Of the other arguments, I disagree with most:

• Yesterday's DS article was not an "accurate impression of what wikipedia is about" by any stretch of the imagination. If it is then I for one would surely give up in disgust, regarding it as a project fit only for bored teenage boys.
• It is not better for vandals to "do it in an article that is receiving a lot of regular-editor attention". The worse effects of vandalism are on the visitors who view it. Hundreds, if not thousands, of people visiting the DS article yesterday would be presented with some truly vile text and pictures.
• You ask how many editors started off editing a front-page FA? I'm sure someone could generate stats or do a questionnaire on "where did you start". Personally, I started on the unloved pages since editing a featured article is scary.
• "Why can't users visit Britannica or Encarta [to avoid vandalism]" - that is so, so far from Wikipedia's aims. We are here to be an encyclopedia, not just build one.
• Fixing vandalism is not "participating meaningfully in wikipedia" any more than wiping dog mess off your shoe is participating meaningfully in a walk in a park.
• Nobody is asking for vandalism to be whitewashed - you'll find it soon enough if you click on a few links. We should, however, make sure we're not just setting ourselves up to ridicule.
• Nobody (in this discussion) is a asking for anons to be "being outright prevented from participating".
• I (and many other editors I'm sure) was put off doing any copy-editing or fact checking on that article yesterday. If I had, I'd surely get an edit-conflict. It would also be hard for other editors to review my changes amongst all the noise.

Colin°^Talk 09:35, 6 December 2006 (UTC)

Part 2 (arbitratry subheading)

Conversely, how do you quantify the number of dedicated users who got introduced to wikipedia by making an edit on the FA as an anon?

I suppose one could do a massive database dump/checkuser matching, but the likelihood of (a) getting approval to get the data generated and available; (b) doing the analysis, and (c) dealing with data difficulties like dial-up IPs, school and work IPs, etc, seem to argue that one shouldn't go there. So let me rephrase the question: How many anon users who start off by vandalizing the Main Page article go on to become dedicated contributors? Were you one? Do you know a lot of other wikipedians who posted obscenities on front page articles and then, realizing the error of their ways, became productive contributors?

I ask such questions because having done an analysis of a large number of anon edits, in the past two days, I can say that virtually all of the anon edits are deliberately damaging - blanking pages, inserting obscenities into images, dumping a large chunk of text about a local high school into the article, etc. These aren't "Hi mom" test messages, or tiny tweaks in the wording of a sentence, or anything that evidences curiosity with a mix of respect for others.

It may not be the right time to start discussing solutions, if there is no general consensus that there is a problem, but I want to note that there are technical alternatives to simply blocking anon edits - allowing anon edits but displaying only "stable" versions; allowing anon edits but storing the diffs for review by regular editors before they are posted; creating a wizard (along the lines of Wikipedia:Articles for creation/Wizard-Introduction, but simpler) that generates a comment on the article's talk page for another editor to review, and so on. John Broughton | Talk 13:58, 6 December 2006 (UTC)

Agree with Borisblue. We are here to build an encyclopedia, not just be one. My first anon edits weren't malicious vandalism, but they were "Hi Mom" style test edits, that weren't productive. When I want to tell someone about the Wikipedia, I tell them to go to the front page, and many stop there, so if that page isn't editable, the "anyone can edit" loses much of its force. AnonEMouse ^(squeak) 14:39, 6 December 2006 (UTC)

So we should unprotect the main page? You gonna revert the 5+ instances of vandalism it would get every minute? --W.marsh 15:59, 6 December 2006 (UTC)

AnonEMouse - if most of the anon IP edits were "Hi mom" types, I wouldn't care. They're NOT. You might want to pull some vandal patrol of the Main Page article and see for yourself.

Wikipedia protects lots of articles, for a variety of reasons - see Wikipedia:List of protected pages; it's simply a balance between benefits and costs. In addition to the Main Page, you might want to suggest people look at Wikipedia:Quick index, where - like a library - they can browse the shelves as they want. Or that they also go to Wikipedia:Featured articles, and check out one of more than 1000 articles, in an area they're particularly interested in. John Broughton | Talk 16:27, 6 December 2006 (UTC)

Not to be rude AnonEMouse, but I haven't seen you edit a main page FA in at least several weeks. I'm not sure you're really aware that the vandalism we're talking about is people inserting huge 3000x2000 pictures of... the kinds of stuff vandals find amusing... or adding GWB as the lead picture for Downs Syndrom (which got us mocked on Wonkette, where that kind of stuff was presented as what Wikipedia was about). Sometimes this stuff has stayed up for 10% of the time the FA was linked to from the main page. I just want to make sure you realize what you're defending here. It's not "Hi mom" style test edits. --W.marsh 16:40, 6 December 2006 (UTC)

Final comments:

• The idea that semi-protecting one article is going to discourage some imagined "potential contributors" and cause harm to the project is a red herring. Indeed, why not unprotect the main page? Think of all the editors we're losing by having the main page protected. Gosh!
• Something like this is going to happen, if only out of the sheer technical necessity of stopping an unmanageable number of crap edits to the TFA, as Wikipedia traffic keeps growing (Alexa #12 and rising). If a proactive approach isn't taken, there will be more and more protect/deprotect wars among admins, as TFAs increasingly go to tatters—with some admins apparently preferring them in tatters based on a stubborn, contextless, and over-applied idealism. –Outriggr § 00:38, 7 December 2006 (UTC)

disputedpolicy

I removed this header as although things are being discussed, it is misleading to people who are referred here to explain current policy. --Trödel 02:13, 7 December 2006 (UTC)