Wikipedia:Administrators' noticeboard/Archive2

Contents: December 9, 2004 - January 31, 2005

Tasks

WP:PUI

From WP:CU: Wikipedia:Possibly unfree images needs some serious attention from an administrator. The page lists images that are in some way problematic and has apparently not been attended to since late September. Thuresson 15:10, 9 Dec 2004 (UTC)

It has not been neglected so long, but does have a very large backlog of old entries. -- Infrogmation 00:08, 13 Dec 2004 (UTC)

Backlog of old stuff is now pretty much cleared out. Yay. -- Infrogmation 18:31, 23 Dec 2004 (UTC)

WP:CP

Help in housekeeping on Wikipedia:Copyright problems is appreciated. -- Infrogmation 08:06, 10 Dec 2004 (UTC)

So what do we do about ones like Brazil for Christ Pentecostal Church? - Ta bu shi da yu 09:27, 10 Dec 2004 (UTC)

And Marcellin Champagnat from [1]. The article was rewritten and now the copyvio is in the edit history! - Ta bu shi da yu 09:32, 10 Dec 2004 (UTC)

Also see Ödön von Horváth which needs to be removed (by a developer no less) from the edit history. [2] - Ta bu shi da yu 09:34, 10 Dec 2004 (UTC)

Help determining dificult cases is much appreciated. Work on shoveling out the obvious cases when appropriate, however, is always needed to help keep the page at a managable size. -- Infrogmation 00:08, 13 Dec 2004 (UTC)

Blocking bug

Due to a somewhat critical bug in MediaWiki 1.4, no blocks are expiring at or after their given date and time of expiry - a block is only lifted if the user/IP is manually unblocked. I did a lot of unblocking this morning, but until the bug is fixed, if anybody has a few free minutes, I suggest periodically browsing through Special:Ipblocklist and unblocking the users whose blocks have expired. Rdsmith4— Dan | Talk 17:54, 26 Dec 2004 (UTC)

Incidentally, another bug causes the expiry time for "indefinite" or "infinite" blocks to be given as the current date and time, so don't unblock those. Rdsmith4— Dan | Talk 19:51, 26 Dec 2004 (UTC)

I've done a few more this evening. I hope the bug get's fixed soon. In the meantime adims should probably try to go easy on using blocks in the first place, at least until the bug is fixed. Theresa Knott (The snott rake) 00:10, 27 Dec 2004 (UTC)

Thanks; not only do I now know about the bug, but now I know about this fantabulous noticeboard. :) --Golbez *** 09:47, 27 Dec 2004 (UTC)

This might have something to do with the fact that in the block log shows the block reason in the place of the block length (172 blocked "User:165.247.204.19" with an expiry time of (plagiarism and revert warring)). --fvw* 14:02, 2004 Dec 27 (UTC)

These bugs in unison are unfortunate, any ETA on them getting fixed? I may have accidentally unblocked some IPs that were meant to be permanent; hopefully they'll persist in their behavior so that they can be spotted easily again. I promise to be more careful. --—Ben Brockert (42) UE News *** 04:28, 1 Jan 2005 (UTC)

To whoever ends up archiving all these discussions: please leave this one on the page as a reminder until the bug gets fixed. Rdsmith4— Dan | Talk 03:59, 3 Jan 2005 (UTC)

VfD backlog

Could someone please look into this? Vacuum c *** 01:25, 2 Jan 2005 (UTC)

A reminder about reverting vandals

I'd like to ask everyone once more to please check their past edits when reverting vandals (I'll admit to forgetting to do this myself too, but it is very important). --fvw* 14:07, 2005 Jan 5 (UTC)

Found this in an incident report that was being archived. I'm sure we all remember most of the time, but a reminder can't hurt... Noel (talk) 15:08, 7 Jan 2005 (UTC)

Wikipedia:Templates for deletion help

A number of templates are cleared for deletion, and there is a bit of a backlog. Can someone please help us out ? -- Netoholic @ 23:03, 2005 Jan 6 (UTC)

This seems to be somewhat better now. However, see the next entry.... Noel (talk) 03:57, 8 Jan 2005 (UTC)

I've been trying to keep up, clearing to-be-deleted templates from pages, and other cleanup tasks. If any admin can just watch the "Holding area", which is where we'll place to-be-deleted templates. The ones ready to go will be noted as "Cleared". Thanks. -- Netoholic @ 07:58, 2005 Jan 8 (UTC)

Wikipedia:Categories for deletion help

Looks like this needs some cleaning up, too. Noel (talk) 20:00, 7 Jan 2005 (UTC)

This is really majorly backed up. I spent most of the day hacking away vaste swathes of material, and it's still totally overgrown. This page desperately needs some admin to take this on as their permanent bailiwick, and become its local beat cop. Alas, I just don't have the time. Do we have a volunteer? Noel (talk) 03:57, 8 Jan 2005 (UTC)

I whack at it every time I play with wikipedia. It's gotten to the point that it's almost the only thing I do on wikipedia. Even so, I've been slacking lately, and spending time on my other hobbies. Every little bit anyone does helps. When (the rare times anyway) I finish with CfD, I start whacking at the categories marked for deletion (in the category) but not listed on CfD. (Essentially, if they are empty, they go. If not and they are not on CfD, the notice gets removed.) --ssd 05:01, 9 Jan 2005 (UTC)

I have spent a few hours deleting only the categories most obviously needing deletion, and there are still some left. However, I have not cleaned up the residue from the CfD page; some non-admin can do that as well as I can probably. Also, there are a large number of things that have been voted on, but no action taken (like moving articles to different categories). This should probably be done by bot by large category, and either person or bot for smaller ones. (This also could be done by a non-admin.) There is a bot, but I think it is on vacation. --ssd 08:06, 9 Jan 2005 (UTC)

I think the bot is already working on a fairly large backlog of categories that need to be moved (from Wikipedia:Categories for deletion/US vs U.S.), so it's just a matter of time before it all gets cleared up. --ssd 13:57, 10 Jan 2005 (UTC)

I've spent the last day or two cleaning it up so it's length is more tolerable now but there's still a fair number of categories from December to be dealt with. I've used my bot to help move pages on category renames. RedWolf *** 23:46, 15 Jan 2005 (UTC)

Handling 3RR violations

May I request that if someone deals with a 3RR violation reported in #Incidents, can they please drop us a short note here to say that they have done so? That way, others won't have to spend time looking at the case only to find out it has already been dealt with. Thanks! Noel (talk) 14:19, 9 Jan 2005 (UTC)

Featured Picture Candidates in need of attention

Featured Picture Candidates needs someone familiar with the process to go over the nominations and do some promoting and archiving. Some of them are long overdue in getting a promotion. Mgm|^(talk) *** 10:20, 13 Jan 2005 (UTC)

I think we've caught up now. It doesn't need an admin, just a couple more editors willing to help. The instructions are at the bottom of the page. User:Solitude used to do a fair bit of this, but went on a WikiBreak just before Christmas, and User:Ed g2s was also away for a couple of days. That just left me, handling the promotions and managing the Picture of the day templates. With Wikipedia crawling, it took me nearly 3 hours to promote 3 pictures on Thursday. -- Solipsist 11:32, 15 Jan 2005 (UTC)

Wikipedia:Introduction

This has replaced Introduction 2 as the intended place for test-editing, so there's no need to revert apparent nonsense unless it's profane or offensive, though it might be a good idea to check every once in a while to see that the first two lines are intact:

{{Please leave this line alone}}
<!-- Feel free to change the text below this line. No profanity, please. -->

Thanks a bunch! Rdsmith4— Dan | Talk 03:32, 21 Jan 2005 (UTC)

Added this to Wikipedia:Cleaning department. Noel (talk) 22:27, 5 Feb 2005 (UTC)

NowCommons needs emptying

The NowCommons category has many images which have been moved to the Commons and are in need of deletion before the commons version will be used. Using NowCommons is better, IMHO, than listing all of these images on Category:Candidates for speedy deletion. BrokenSegue 17:15, 23 Jan 2005 (UTC)

Agreed. I'm on it. - UtherSRG *** 17:29, 23 Jan 2005 (UTC)

Nice job. BrokenSegue 18:51, 23 Jan 2005 (UTC)

Should have told you to watch for broken images. BrokenSegue 03:56, 24 Jan 2005 (UTC)

Thanks.... - UtherSRG *** 04:18, 24 Jan 2005 (UTC)

Added this to Wikipedia:Cleaning department. Noel (talk) 22:18, 5 Feb 2005 (UTC)

MVP

Yes, the Wikipedia WP:MVP, or Most Vandalized Pages. This page is apparently slightly out of date, and I don't know how many Administrators are checking these pages. It would be nice if everyone contributes to what pages they see frequently vandalized and check this supposely shared watchlist. -- AllyUnion (talk) 09:19, 28 Jan 2005 (UTC)

General

Double-blocking

What do you guys do to avoid blocking an already blocked IP? Do you check the log or what? I'd like to see a note on someone's user page if they're blocked, so I know I don't have to bother checking. Mgm|^(talk) *** 11:52, 10 Dec 2004 (UTC)

Good question. I usually check the IP block list, but that's a bit of a pain. A warning + confirmation that the IP/user is already blocked would be pretty helpful. - Ta bu shi da yu 12:39, 10 Dec 2004 (UTC)

I leave a message on the IP's talk page so that other admins following up don't have to waste the time checking. Still, since not everyone does that, the only way to be sure is to check the IP block list before blocking. SWAdair | Talk 03:45, 12 Dec 2004 (UTC)

Perhaps a new feature for MediaWiki could be an automatic symbol for blocked users. It could show up next to their name/IP in recent changes and have a sign on their user/talk pages that isn't part of the article (and thus not removable). violet/riga (t) 17:20, 13 Dec 2004 (UTC)

I would like such a feature very much. It saves vandal hunters so much time. Mgm|^(talk) *** 09:17, 14 Dec 2004 (UTC)

That would be a nice feature, but at present, I don't worry myself over double-blocking. They were blocked for a reason. How many cases of reformed vandals do we have? --Golbez *** 10:17, 27 Dec 2004 (UTC)

IP addresses and sock puppets

CWS, Netoholic, Viriditas and TBSDY (myself) have all been in discussion about the whole sockpuppet to IP address issue. Personally, I've noticed suspicious edits on articles by anons where users have been blocked for 24 hours. Some sort of verification of the user's past IP addresses and matching them to the user might be helpful. Except that auto-blocking comes in useful here... - Ta bu shi da yu 23:04, 10 Dec 2004 (UTC)

I've been meaning to ask somebody this, and I figure you may know: is there any way for an admin to view the last active IP of a non-anon user? Many thanks. -- ClockworkSoul 23:04, 9 Dec 2004 (UTC)

No, only a system administrator (Tim Starling comes to mind) can find that out. -- Netoholic @ 00:38, 2004 Dec 10 (UTC)

Ah, okay. Thanks. -- ClockworkSoul 22:17, 10 Dec 2004 (UTC)

CWS, why do you need to know? - Ta bu shi da yu 02:25, 10 Dec 2004 (UTC)

In reading recent accusations of sockpupperty and watching recent vandals rotating usernames, it seemed to me that the ability to cross-reference IP addresses would render the former moot, and the latter somewhat easier to deal with. -- ClockworkSoul 22:17, 10 Dec 2004 (UTC)

tell me about it. I said the same to Viriditas! - Ta bu shi da yu 22:21, 10 Dec 2004 (UTC)

Do you think its worth it for us to propose that administrators get the ability to see the source IP address of an edit, in addition to the author? I can't think of any major security issues, especially if it's only admin that can see them. -- ClockworkSoul 22:25, 10 Dec 2004 (UTC)

Absolutely! That would solve many sock-puppet issues. - Ta bu shi da yu 22:26, 10 Dec 2004 (UTC)

Would you like to write survey, or shall I? :) -- ClockworkSoul 22:30, 10 Dec 2004 (UTC)

Unfortunately, such a proposal will not currently solve any sock-puppet issues if the users in question are editing from a secure browser (java disabled) and using an anon proxy. IIRC (and I could be mistaken), the only way to solve this problem is to ban open proxies. Many IRC servers have implemented this solution, but I believe there are ways around it, depending on the type of proxy. --Viriditas 22:56, 10 Dec 2004 (UTC)

No, it wouldn't entirely eliminate sockpuppetry, but it would cut it down somewhat, and would make the lives of us vandal-hunters somewhat easier. It's also a bonus that the software could support such a change with a near-trivial amount of effort. I agree that your proposal, Viriditas, would go alot farther, but is also a grander undertaking. -- ClockworkSoul 23:15, 10 Dec 2004 (UTC)

• If people with a free mail service need to contact an admin to sign up, it's likely they won't go through the effort causing us to lose valuable editors. I'd like to see some better vandal and trouble user logging, perhaps added to their user page so vandal hunters can easily check an IP's or user's past. Mandatory sign up isn't a problem, as long as we don't make verification neccesary. Mgm|^(talk) *** 14:44, 11 Dec 2004 (UTC)

Question: Why do you need to spy on users through the IP addresses? Contributions should be judged on what they are, not who wrote them. Sock puppets are annoying, yes, but there's nothing that can be done about them without massive changes to Wikipedia (for ex., requiring proof of identification before one could create an account, and blocking anon. users) that wouldn't be popular. I'm against the Big Brother type activity. If someone's contributions are so one-sided that he's an obvious sock puppet, then delete/demote them on POV grounds rather than surmised identity. EventHorizon 05:57, 12 Dec 2004 (UTC)

Welllll'... normally I would have agreed with you on this one. But lately we've been blocking for violation of the 3RR. We've been seeing many similar edits on the articles we've blocked for, and the first thing a lot of admins thinks is that those edits come from blocked editors. So it would clear up a lot of issues if we could see the IP addresses used of editors. Mind you, from what I understand its a bit of a moot point because the autoblock automatically blocks those IP addresses anyway... Ta bu shi da yu 20:39, 12 Dec 2004 (UTC)

Even though seeing an editors IP may seem trivial (considering developer people can do it no problem), I fear that the seeing a editors IP would be a power abused by admins. It is always nice to know something other people don't, eh? Considering that admins have been abusing blocking policy regarding 3RR (a policy DISASTER), not a good idea to give them more "responsibility". A temptation to much to resist. Sock puppets are not illegal per se. Abusing them to circumvent policy is.

The only way I could support admins seeing an editors IP would be if 3 lay (non-admin) editors, approved that course of action beforehand. As has been stated, admins have more responsibility but authority lies with the wikipedia community. I do not trust admins, sorry.

It does seem to be a moot point, considering the autoblock. Seeing a "rogue editors" IP, which can be changed easily, will solve very, very little. Wikipedia has not been damaged by admins inability to see an editors IP. Mrfixter 23:42, 12 Dec 2004 (UTC)

OK, fair enough these are your views. However, where have admins violated the 3RR? Also, why do you think we have to block because of repeated violations? Well, because the point is that people are reverting and not even bothering to discuss their changes. Then when we block because of it they get all pissed off. Well, I say tough. If they had been making edits in good faith they would have taken their comments to talk to work on. And they would have tried to gain consensus. Reverting in the manner they do is bullying, plain and simple. Why? Because they are trying to impose their own POV on people. Take for instance CheeseDreams. She reverted and lost at least three changes on Cultural and historical background of Jesus [3] and along with other editors has forced the protection of that page. Do you think that is fair? I don't. - Ta bu shi da yu 08:31, 14 Dec 2004 (UTC)

Where have admins violated the 3RR blocking policy? HistoryBuffErs 4 reverts in 26 hours block, Alberunis 4 reverts in 24 hours 5 minutes, Nasrallahs block after two reverts in 24 hours, Sam Spades blocking. The 3RR policy should be followed to the letter by the admins. Now, without getting off topic, I would characterise the practices of these editors as very far outside of what I would consider civil behavior. But policy should be applied evenly and justly, which the admins have failed to do. I don't think admitting that is a radical stance to take.

3RR blocking policy has also meant more sockpuppet, anon ip reverts, and has slowed down revert wars but has not ended them. Was that the point of 3RR blocking policy? Like other things in wikipedia, the long-term solution lies within the wikipedia community, not in extending admins executive power.

The wikipedia community has not given admins discretion in enforcing 3RR. 3RR must be enforced, not the individual admins interpretation of 3RR, if an admin is serious about their duty. Mrfixter 17:29, 15 Dec 2004 (UTC)

Yeah, but users who deliberately go out of their way to push their POV into an article may find themselves blocked. Those users need to learn how to play nice. Do you think I like blocking regular editors? - Ta bu shi da yu 14:39, 19 Dec 2004 (UTC)

In practice admins do have discretion and are requred to exercise it fairly. They're not robots. HistoryBuffer and Alberuni's cases, if they're as you describe them, were abuse. Whether they stuck to the letter of the 3RR or not is irrelevant, if they intentionally tested its boundaries so blatantly. No editor needs to revert an article over a disagreement with another editor more than once in a blue moon, and we always have the option of modifying instead of reverting where it is appropriate to do so. The 3RR is so absurdly generous as to encourage revert warring, in my opinion. --Tony Sidaway|Talk 15:19, 19 Dec 2004 (UTC)

Edits should be judged for what they are, not by who made them. But the concept of voting, and the 3RR, are undermined by sockpuppets. E.g., votes for deletion are a regular target of sockpuppet attacks. Take RFA into account, and they become a direct threat to the project. Recording IPs is not spying. Come on. Every website you visit has your IP, and they can do with it what they like. WP accounts are a gratuitous way to hide your IP, but I don't see any inherent users' right to this service. dab (ᛏ) 13:19, 22 Dec 2004 (UTC)

A strategy for identifying and blocking vandals

Discretion and skill are needed in interpreting the IP numbers with a view to sock identification. I'm currently suffering intermittently from an IP block that was applied to my ISP's http proxy legitimately because of a vandal who is sharing the same proxy. There's nothing much we can do about that; it doesn't render editing impossible (but this is also true for the vandal).

One way to move forward may be to permit editing only by logged-in users. This would mean that a vandal can be blocked by username. Vandals attempt would get around the block by storing up usernames prior to vandalism and switching to an unblocked username when required. However this strategy could be defeated by recording the IP number under which each username is registered and then raising an auto listing (on some Special page) of users performing recent edits using the same IP number as banned users. This would provide a watchlist of candidates to watch for further vandalism.

It would also mean that crude and often ineffective bans on IP numbers, such as the one I'm suffering, would not be necessary. --Tony Sidaway|Talk 14:56, 11 Dec 2004 (UTC)

Banning all anons is anti-wiki and frankly extremely infuriating to editors who don't want to register and subsequently log in all the time. But, I think that admins ought to have the developer power of viewing IPs. Unlike what Event Horizon said, this isn't a "big brother" tactic - Wikipedia is not a government or a public place, it's run by Jimmy Wales and the Foundation and so on, and the owners of websites can view the IPs of their visitors anyway. In Wikipedia's case, developers already have that power as well. The only difference here is giving that status to appointed admins. Andre (talk) *** 06:20, 12 Dec 2004 (UTC)

Thanks. I have an account but I accept cookies from Wikipedia and have my preferences set to login automatically. I would accept a policy that might not appeal to wiki-purists if it meant that I did not occasionally find myself locked out of Wikipedia because of something somebody else did using the same web proxy I use. --Tony Sidaway|Talk 21:14, 12 Dec 2004 (UTC)

• I've been thinking about this for some while now, and though I personally don't see the harm I do think publishing everyone's IP (even just to admins) would be a bad idea. I propose that instead we publish a hash of the both each editor's entire IP, and just the first 20 bits of their IP. To prevent brute-forcing, it would have to be the hash of those values concatenated with some secret key. That way, anyone can verifiy whether two users are using the same IP and whether they're using the same ISP, but noone can retrieve their actual IP, and their ISP can only be revealed by someone using the same ISP. It seems like a reasonable compromise to me, the only drawback would be that we couldn't distinguish an IP from a named user with that IP. This could be resolved by not using the User:IP name anymore but User:<hash of ip>; That would lose us the ability to see the address of anon's though, which I rather wouldn't. --fvw* 19:15, 2005 Jan 5 (UTC)

good idea, although I imagine this would take some time to implement? dab (ᛏ) 20:01, 5 Jan 2005 (UTC)

I think the larger part of the time between it being accepted and actually happening would be waiting for the next update of mediawiki to be install on Enwiki. This is yet another of my I'll-code-it-so-it-actually-happens-some-time-soon-I-wanted-to-play-with-mediawiki-anyway idea's, the only thing is I'd like to make sure it will actually be used up front. Would this require a policy vote or something or should it just sneakily be coded and should the devels then be bribed to turn it on? --fvw* 20:06, 2005 Jan 5 (UTC)

Re acceptance of IP address on all edits visible to admins as a datapoint LiveJournal and its clones can all display the IP address of commenters on a post, visible to either the original poster or the maintainer of that community and I've never seen any complaint about that (NTSIDEOC) --Vamp:Willow 15:08, 7 Jan 2005 (UTC)

Changing username

Could one or more admins have a look at Wikipedia:Changing username. As far as I can see, no request has been fulfilled since 14 October so there's quite a backlog. Thanks. jguk 13:26, 11 Dec 2004 (UTC)

Unfortunately you need a developer to do that. Andre (talk) *** 19:45, 11 Dec 2004 (UTC)

Any ideas of how to get hold of one (or more)? jguk 20:13, 11 Dec 2004 (UTC)

m:Developer has a listing, but fair warning: they're overworked, and chorese like changing usernames are (except in unusual cases) low-priority compared to working on Mediawiki. Isomorphic 20:59, 11 Dec 2004 (UTC)

Links to Wiktionary

Hi, all, there's a proposal (and a template) at Wikipedia:Soft redirect for handling pages that ought to be links to Wiktionary. I'd like to start using this, which looks good to me, but I'm not sure it's really official yet; can we get some more comment there so we can make this formal? (And my apologies if it already is, and I was just clueless - the page doesn't sound like it, and I hadn't heard about it until very recently.) Noel (talk) 17:45, 13 Dec 2004 (UTC)

I think this is a good policy for topics with no encyclopedic potential. -- Jmabel | Talk *** 23:00, 13 Dec 2004 (UTC)

Three revert rule

I ask here because I think it may be the best way of getting a quick feel for the consensus on a recent change. Recently the three revert rule was made enforceable. Now the case of a simple revert is easy to identify and most edits which are termed "reverts" fall into this category: a diff between two different versions of the text of a page or section shows that they are identical, with all intermediate changes reverted.

More recently I have noticed users deleting the added text of other users selectively. This is slightly more difficult to recognise because a diff between two versions isn't identical. What shows up however is that if two versions are compared, added text in intermediate changes is selectively deleted. So for instance in one example the first user made a cosmetic edit, correcting the spelling of a single word, and a second user made a more substantial edit in which text was added. A third user then came along and performed an edit to delete all of the added text of the second user while retaining the cosmetic edit of the first, and also tweaked the heading of a section. The effect was to remove the added text of the second user.

This becomes a question because the third user had very shortly before performed two reverts on the same page, and the third edit could be seen as an attempt to evade the three revert rule.

I'd like to open the question up. Is this kind of edit covered by existing policy or practise? Does it qualify in this case as a third revert? --Tony Sidaway|Talk 02:41, 14 Dec 2004 (UTC)

My opinion is that a revert which includes minor changes (such as spelling changes and capitalisation) is still a revert. The purpose of the 3RR is to prevent revert wars, and clearly spelling changes don't make much difference to most such wars. I think a warning might be a sufficient rebuff for a first offence.-gadfium (talk) 03:08, 14 Dec 2004 (UTC)

I would agree with this comment. - Ta bu shi da yu 08:32, 14 Dec 2004 (UTC)

Me too. The advantage of having people decide rather than machines is we can see through attempts to works the system. Theresa Knott (The snott rake) 09:01, 14 Dec 2004 (UTC)

Agreed. Mgm|^(talk) *** 11:36, 14 Dec 2004 (UTC)

I thought it was pretty clear in the discussion at Wikipedia:Three revert rule enforcement that reverts that also make other changes should count for the purposes of the three-revert rule. For example, see my vote on that page. It shouldn't matter whether the other changes mixed in with a revert are major or minor, it's still a revert. However, I have just checked back at Wikipedia talk:Three revert rule enforcement#Spirit vs. letter of the 3RR and Wikipedia talk:Three revert rule enforcement#Mixing reverts and significant edit, and things no longer seem as clear as I had remembered. I would support a clarification to the 3RR saying that a revert mixed with other changes is still a revert. —AlanBarrett 17:39, 15 Dec 2004 (UTC)

I think a revert mixed with changes is still a revert; if someone wants to make unrelated changes, they can certainly do so in a separate edit. And indeed, if someone were to do the revert change, and then make other unrelated changes, they would definitely be blocked under the 3RR. Most admins I have seen have been interpreting revert+edits as a revert. Jayjg 20:33, 15 Dec 2004 (UTC)

I think the key word here is "unrelated" changes, because those are just cosmetic and an attempt to get around the rule. Substantive and/or responsive changes whether contiguous with the disputed text or not, as is appropriate to the article and issues involved, should not count as a revert. I propose also, that in the interest of community, mere deletion reverts be treated more critically, since unless the additions were vandalism, the assumption should be they are good faith contributions.--Silverback 04:43, 3 Jan 2005 (UTC)

I strongly disagree. Users have used this exact excuse to get around the rule before; on the fourth revert, re-inserting a disputed paragraph, while making substantial edits to other sections. If they want to make their substantial edits to the other sections in good faith, they can certainly make them in a separate edit. Re-insertion of disputed text along with other major edits is bad faith. Jayjg | (Talk) 04:56, 3 Jan 2005 (UTC)

If it is a related and substantive change, as in the case I have in mind, it wouldn't make sense to make the other major edit separately from the insertion, they would be somehow supportive of each other or make a connection that is responsive to stated objections, the opposite of bad faith.--Silverback 05:08, 3 Jan 2005 (UTC)

It's crystal clear to me that if we don't consider a "reversion+edits" as a reversion, to count toward the 3RR, we might as well ditch the 3RR rule - because otherwise everyone will make an edit as well, every time they do a revert, and will thereby avoid ever triggering the 3RR. Noel (talk) 04:23, 1 Jan 2005 (UTC)

I have raised this issue (that unless a revert+edit counts as a revert, the 3RR is a dead letter) at Wikipedia talk:Three revert rule#Edits and reversions; comment there would be welcome. Noel (talk) 23:17, 1 Jan 2005 (UTC)

It will take judgement. Substantive and/or responsive changes (responsive to edit summary or talk page points). Minor changes, such as adding a wikilink that addresses issues raised can be substantive and responsive.--Silverback 04:43, 3 Jan 2005 (UTC)

In most cases I've seen, it's just an attempt to get around the 3RR. I suppose it's possible in some cases that it's not, and admins should always attempt to exercise good judgement in all situations, of course. Jayjg | (Talk) 04:56, 3 Jan 2005 (UTC)

Missing Redirects project

For those who haven't run across this before, Nickj has created a really cool tool that runs across the database and collects information from piped links, using it to suggest additional redirects. (More about this project here, with prior discussion here.)

It has created a lengthy list of suggestions. To avoid making bogus links, the process is human-mediated. Nick has set it up so it only takes two clicks to create each one, so there's no work involved. You just sit and click if you think it's a good link. The lists are broken up into blocks of 5 suggestions, so that they are not overwhelming, and so that people can do as much or as little as they like. Now we need helpers. Where is the best place to publicize this project? Noel (talk) 13:33, 14 Dec 2004 (UTC)

To me the community portal or note on RC seems like good places to start. Mgm|^(talk) *** 13:56, 14 Dec 2004 (UTC)

Asksql

does anybody know what happened to Special:Asksql? There used to be a db query form, but now there is "No such special page"... dab (ᛏ) 08:34, 16 Dec 2004 (UTC)

I think it was temporarily locked to speed up the database, but I might be wrong. Mgm|^(talk) *** 08:55, 16 Dec 2004 (UTC)

I could be wrong since I never used it anyway, but I think it's been "temporarilly" locked for a very long time. I wouldn't hold your breath. Isomorphic 21:05, 16 Dec 2004 (UTC)

It's unlikely to ever come back. One poorly written query could bring the entire site to its knees. -- Cyrius|✎ 01:44, 18 Dec 2004 (UTC)

It's gone indefinitely, not temporarily. The problem I have with it is not performance (we can always run it on a small slave server) but rather the fact that you need thousands of GRANT statements to set up the appropriate permissions. -- Tim Starling *** 10:16, 19 Dec 2004 (UTC)

pity. although it was quite esoteric anyway. I suppose the permission commands could be generated by a perlscript? and if I am correct, there is a 30s timeout on queries, so that the system would not go down even from the stupidest commands? but it's ok, I have hardly ever used it myself, if was nice to have, and not in any way essential. dab (ᛏ) 10:32, 19 Dec 2004 (UTC)

Undelete help

I tried to undelete AMOK! South Africa but now I can't get to the page history. Can someone please advise on the best course of action? - Ta bu shi da yu 08:41, 19 Dec 2004 (UTC)

It seems to be there now. Page history is:

    * (cur) (last)  08:45, 19 Dec 2004 Ta bu shi da yu
    * (cur) (last) 08:44, 19 Dec 2004 Ta bu shi da yu (blanking copyvio)
    * (cur) (last) 09:02, 5 Dec 2004 Vague Rant m (Update link.)
    * (cur) (last) 01:47, 5 Dec 2004 Vague Rant (Copyvio.)
    * (cur) (last) 02:18, 29 Nov 2004 168.209.97.34 (Weblinks)
    * (cur) (last) 02:17, 29 Nov 2004 168.209.97.34 (AMOK! South Africa)
    * (cur) (last) 16:08, 27 Nov 2004 Mudthang m (Overview)
    * (cur) (last) 15:53, 27 Nov 2004 Mudthang m (Created 27 Nov 2004 by mudthang)

HTH. -- ALoan (Talk) 08:53, 19 Dec 2004 (UTC)

Undeletes sometimes take some time before they show up. Cool Hand Luke 10:25, 19 Dec 2004 (UTC)

FWIW, I discovered soon after that if I did an edit then it all would come back :) Seems to fix things. - Ta bu shi da yu 14:33, 19 Dec 2004 (UTC)

Another possibility that comes to mind is that your browser had a cached version of the history page from before the undelete and it was showing you that rather than retrieving the updated version. I've had that happen to me while doing deletion-related stuff before, and it freaked me out because I thought I'd irretrievably lost the Cassini-Huygens probe article while trying to move it. On many browsers holding down the shift key (or control key Pedant) while clicking the page reload button will bypass the cache, that could fix things if you run into this problem again. Bryan 19:18, 19 Dec 2004 (UTC)

Nope, thanks for the advise but Wikipedia definitely brought up a "history not found in database" error. - Ta bu shi da yu 19:53, 19 Dec 2004 (UTC)

List subpages

Hi, can anyone tell me how to list the sub-pages of a given page (i.e. a command or page which, once given/looked at, provides a list). I've looked in Wikipedia:Subpages and Meta:Link#Subpage_feature, but neither says anything of how to do this. Sub-pages always include a link to the parent - is there anything analagous in the other direction? Thanks in advance for the answer (if any :-). Noel (talk) 18:23, 19 Dec 2004 (UTC)

You could bring up "what links here" and them do a search within that... -- Jmabel | Talk *** 00:14, 20 Dec 2004 (UTC)

Nope. Those links back to the parent don't get included in "what links here". Noel (talk) 00:42, 20 Dec 2004 (UTC)

One method is to do a Wikipedia search for the page; see, e.g.: this search; that brings up some of the subpages for my User pages under "Article title matches". — Matt Crypto 00:53, 20 Dec 2004 (UTC)

If, as Matt's specific example might suggest, it's a matter of finding your subpage that you lost track of, and if you've always kept Watch this page checked since starting to create them, the complete list'll appear at yourM watchlist-maintenance page, neatly grouped together under the Us of the alphabet. The red links do not reflect deletions, but rather User talk: (sub-)pages that lack corrsponding User: (sub-)pages -- typically your talk archives. --Jerzy(t) 21:28, 2004 Dec 20 (UTC)

Where to report three revert rule violations

On another note, we need a specific place to report 3RR violations (e.g. not on the talk pages of large numbers of individual sysops). If there were a particular page (Wikipedia:Excessive reversions in progress? Wikipedia:3RR violations?) then people could discuss (quickly) which users did or didn't violate the 3RR. Pakaran (ark a pan) 20:45, 23 Dec 2004 (UTC)

What is the best way to alert administrators to 3 revert rule violations? Now that blocking is policy, should they be brought here so admins can deal with them? Jayjg 01:14, 24 Dec 2004 (UTC)

I think so. Offenders often feel that the opposing side cherry-picks admins sympathetic to their cause. Violations are typically obvious enough that any admin should be able to block infringing parties, and posting them here would unify any revert-count disputes in one place without having them sprawl across user pages. Cool Hand Luke 01:28, 24 Dec 2004 (UTC)

Sounds good for now. If it picks up to be too much traffic here, we can move it to a separate page. Noel (talk) 07:18, 24 Dec 2004 (UTC)

Is this where we report 3RR violations now? I tried to find a proper place for it (not wanting to block somebody over a dispute I am myself involved in). I ended up putting a note on WP:RFP (please see there). dab (ᛏ) 12:25, 26 Dec 2004 (UTC)

I believe so; this place seems more appropriate IMO. Johnleemk | Talk 14:19, 26 Dec 2004 (UTC)

WP:RFP probably isn't a good place, for several reasons. First, if a user violates the 3RR, the solution is to block them temporarily (repeated violations should be taken to ArbComm, who may ban them). Second, my sense is that policy was not to protect pages because of edit wars unless there's a really heated edit war among many parties, and two people warring doesn't meet that standard (and see the first point). Noel (talk) 18:06, 26 Dec 2004 (UTC)

no, you are quite right. I think we will need a special page where people can request such blocks (this page doesn't seem an obvious choice) dab (ᛏ) 21:19, 26 Dec 2004 (UTC)

In line with User:Charm's point at WT:AN#Are people finding this useful? (which also resulted in the recent addition to header of this page), this sounds like a reasonable suggestion. Please propose a location for 3RR violation reports! Noel (talk) 16:13, 27 Dec 2004 (UTC)

I added a note to Wikipedia:Three-revert rule suggesting that violations be reported here. —AlanBarrett 17:00, 8 Jan 2005 (UTC)

I suggested the Project Page, itself, by used to report violators. That idea was reverted without discussion in less than 2 minutes. My thoughts on the talk page of the 3 revert page. -Husnock 2Feb05

TBSDY has been bold - see below. Noel (talk) 07:36, 2 Feb 2005 (UTC)

New undeletion ability

crossposted to several places
Sysops now have the ability to undelete only selected revisions of a page. Please read over the explanation on Wikipedia:Viewing and restoring deleted pages by sysops and join in the discussion on Wikipedia talk:Undeletion policy. Thank you. —Charles P. ^(Mirv) 13:44, 24 Dec 2004 (UTC)

Special pages

Is there any way to get special pages like Special:Log/block on my watchlist? Back when it was called Wikipedia:Block log, I found it useful to be able to see updates on my watchlist. Gamaliel 21:12, 24 Dec 2004 (UTC)

Unfortunately not. I miss that too. BLANKFAZE | (что??) 21:26, 24 Dec 2004 (UTC)

I haven't tried this to see if it works, but you could try putting a link to Special:Log/block (and any similar things you want to keep an eye on) on a sub-page of your user page, and then go there and hit "Related changes" and see if that works. (That's what people used to do before watchlists, IIRC...) Noel (talk) 12:43, 25 Dec 2004 (UTC)

Proxy list

I'm confused about something. Should we be blocking all open proxies currently being used by people like the GNAA? I have found an updated list at http://www.q-cat.com/goodproxies.txt but I don't particularly want to start blocking these until I work out whether we should actually do so. Apparently some of our Chinese users use these IP addresses. - Ta bu shi da yu 23:04, 29 Dec 2004 (UTC)

My concern is that Chinese users using these to get to WP may be putting themselves in a potentially rough situation (whether by doing so, or by saying things on WP that might be found objectionable). Of course, that's their choice. I do think if the GNAA, our favorite psychic, and friends are actively using a proxy, a brief block as per AOL proxies to get rid of them is in order (or wait and rollback). Just my opinion. Pakaran (ark a pan) 04:42, 1 Jan 2005 (UTC)

Block the lot of them. See Wikipedia:Blocking policy#Anonymous_and_open_proxies. --fvw* 05:04, 2005 Jan 1 (UTC)

Policy decision on VfD

I just wanted to bring to everyone's attention a bit of policy that is being decided at Wikipedia:Votes for deletion/User:Amgine/Maureen's RfC. The issue is whether or not it is appropriate for someone to keep a copy of an uncertified RfC in their user space when policy calls for the RfC to be deleted. Is this similar to copying a deleted article to BJAODN or is it circumventing Wikipedia policy? Your input is welcome. SWAdair | Talk 02:58, 30 Dec 2004 (UTC)

Subpages of personal pages were always material for speedies, and I think that Amgine's attempts at an RfC on "his" subpage deserve deletion without further trouble. JFW | T@lk 03:11, 30 Dec 2004 (UTC)

Cat bug

There's an annoying bug in Category:League of Nations which lists the same subcat, Category:League of Nations Mandates, in triplicate. Any idea how to fix this? I've tried deleting the subcat and then recreating it but that hasn't helped. AndyL 19:33, 30 Dec 2004 (UTC)

Seems fine now. Noel (talk) 03:04, 31 Dec 2004 (UTC)

Sockpuppet template

To be used in cases beyond reasonable doubt: Template:sockpuppet

Example usage: {{sockpuppet|user=Alberuni}} - David Gerard 20:37, 31 Dec 2004 (UTC)

I deleted it. If it is a sockpuppet, then block the person up the wazoo. On the other hand, the potential exists for trolls to misuse this template to harrass people--and they have. Danny 14:46, 15 Jan 2005 (UTC)

Autoblocker

The autoblocker appears to be going wild right now, does anyone know what's going on? Jayjg | (Talk) 00:22, 31 Dec 2004 (UTC)

Without knowing the IP address involved, it's difficult to say for sure what's happening, but it appears that either blocked user User:Deeceevoice was sharing a proxy with a large number of people (possible, but unlikely), or that user had a bunch of aliases they tried to use (more likely - note the times on the autoblocks - all within a few minutes of each others, about 3/4 hour after the original block was put in place). Seems to have stopped now. Noel (talk) 02:59, 31 Dec 2004 (UTC)

AOL - David Gerard 23:36, 1 Jan 2005 (UTC)

Speedy deletes marked as VFD

I've run across a few cases on RC patrol where a speedy delete has had the vfd tag added to it (but no VFD enetry yet created). Currently I'm leaving them but unless solves it's problems soon this may not be a good idea in the long run (it also gives a way to get an article on wikipedia for onee week). What do other people do? Geni 21:43, 2 Jan 2005 (UTC)

I'd notify the person adding the tags that some articles do not require a vote before their deletion, and point them to WP:CSD. Rdsmith4— Dan | Talk 22:31, 2 Jan 2005 (UTC)

Histories have gone haywire

Something appears to have gone seriously wrong with histories. Versions are not showing up in the history lists, and "diff" appears to be really screwed up too. I just archived a bunch of incidents from here in separate moves (as is my habit), but the changes don't show up in the history here. Anyone know what's going on? Are we damaging the database? Noel (talk) 14:09, 4 Jan 2005 (UTC)

I noticed that an edit I made on WP:FARC at 9:46 today is visible on the page but does not appear in the history. Very odd. Filiocht *** 14:42, 4 Jan 2005 (UTC)

As far as I understand the mediawiki setup for wikipedia: No, we're not damaging the database (apart from by repeatedly hitting submit while creating articles, but the corruption this causes isn't serious and occurs without what's currently going on too). When you change an article, the change gets fed to the master database, which pushes changes to the slaves which serve database reads. Currently the slaves are lagging in processing the changes, so it takes a while for changes to show up and in the meantime things may be displayed inconsistently. (This all comes with the disclaimer that I might not have a clue what I'm on about ofourse). --fvw* 17:11, 2005 Jan 4 (UTC)

Wikipedia Disclaimers

There are a couple of holes in the legal disclaimers which might need addressing.

1. From Wikipedia:Content disclaimer, last element in the list, link "Wikipedia does not give legal advice" is linked to [[Wikipedia:Legal advice]] which is an unprotected redirect to <nowkiki>Wikipedia:Legal disclaimer</nowiki>.
2. From Wikipedia:Content disclaimer, last element in the list, link "medical advice" is linked to [[Wikipedia:Wikipedia medicine standards]], but should probably be linked to [[Wikipedia:Medical disclaimer]]

Not sure where else to bring this to the attention of admins. - Amgine 18:24, 7 Jan 2005 (UTC)

I have fixed the first one - fixed all article pages that referred to the disclaimer to go straight there. I don't think we need to protec the redir now. I did the second one too (now goes to Medical disclaimer, as suggested). Noel (talk) 04:40, 8 Jan 2005 (UTC)

Sock puppets

What do we do about controversial users who use sock puppets? I ask this because Cheese Dreams has confirmed she has two: User:Cheesedreams and User talk:Cheese Dreams. She setup redirects to her main account User:CheeseDreams. In order to make enforcement of arbcom orders more easy, I would like to request that these two accounts be either removed or blocked (without blocking her main account). It also makes it difficult to correspond with her. - Ta bu shi da yu 05:05, 9 Jan 2005 (UTC)

I wouldn't call them "sock puppets", since those are usually intended to hide the fact that there is only one person there. That is clearly not the case here - it's the same person (and also it's not someone impersonating them). I think the term "alternate username" is more appropriate in this case. Noel (talk) 13:38, 9 Jan 2005 (UTC)

I think that makes sense, but I also think the current rule is that sockpuppets are allowed except when used to evade policy. Jayjg | (Talk) 05:16, 9 Jan 2005 (UTC)

which these are, used for editing while under a block Pedant 19:59, 2005 Jan 12 (UTC)

Test wiki broken

The test wiki at http://test.wikipedia.org/ appears to be in its default, unconfigured state. It makes it hard to do tests that are inappropriate for WP or WP:SB. :-)  — Saxifrage | ☎ *** 09:26, 10 Jan 2005 (UTC)

See Brion's wikitech-l post: The test wiki has been taken down. Lupo 13:19, 11 Jan 2005 (UTC)

open proxies

would it be possible to consider an automated mechanism for blocking open proxies (i.e. an 'admin-script' that auto-blocks IPs after establishing that they are open proxies)? They are a bane, but I cannot be bothered to go after Cantus' list, because there are just too many of them. I suppose I could hack together such a script myself, but I imagine an 'admin-bot' would be controversial, and should be operated by the developers. dab (ᛏ) 11:19, 11 Jan 2005 (UTC)

According to Wikipedia:Blocking policy#Anonymous and open proxies:

User:Proxy blocker was once used to automatically block open proxies, but was turned off as it "spooked" some people's ISPs.

so I guess the answer is "no"! :-) Noel (talk) 16:31, 12 Jan 2005 (UTC)

Great minds think alike… I've currently got a basic script that parses the RC-dump feed off kohl (I wanted to write that for other purposes too, email change notification and such) and was going to hook basic port 8080 and 3128 checking in it (with heavy caching ofcourse, not more than one probe every few months). No auto-blocking of course, but just a "warning, user is posting through proxy" message to my inbox. --fvw* 16:48, 2005 Jan 12 (UTC)

Auto unblocking

Is this still broken? I've been unblocking expired blocks each day since shortly after the move to Wikimedia 1.4, but today it looks like someone has done a very thorough job just before I got there, or the blocks are being auto removed on expiry again. There's nothing in the block log to indicate someone's done it manually.-gadfium 01:03, 13 Jan 2005 (UTC)

Silsor removed the item from this page with the edit summary stating that the bug was fixed, so I assume it is (and there was much rejoicing). --fvw* 01:10, 2005 Jan 13 (UTC)

It appears have been fixed.Geni 01:11, 13 Jan 2005 (UTC)

Yes, that bug and the one which caused the expiry time not to appear have both been fixed. Rdsmith4— Dan | Talk 12:22, 15 Jan 2005 (UTC)

Image copyright problems

There seems to be a lot of images slowly being listed for copyright issues, with no prior warning on articles' talk pages which they pertain to. As such, it seems several articles have missing images and it's very difficult to track why they were removed. See Star Trek and Outpost (computer game) as examples. -- AllyUnion (talk) 08:48, 14 Jan 2005 (UTC)

Part of this, at least, may be related to the continued glitch of image pages too often not displaying what links there. Perhaps until the glitch is resolved, known pages an image appears in be listed on the image page when adding a Wikipedia:Copyright problems or Wikipedia:Possibly unfree images notice. -- Infrogmation 20:56, 15 Jan 2005 (UTC)

Deletion of those images via either copyright problems or possibly unfree images should be unlikely - they are obviously fair use and I'd hope that nobody would delete them on the basis of simply an unsupported listing as possible problems - anyone deleting from the copyright issues areas should at least be able to recognise listings of images as obvious as these as without merit. Perhaps someone with an excess of zeal and lack of understanding of fair use? The previous possible copyvio requirement of removing the image from the article using it (and thereby letting those who watch the article know there is a problem) now seems missing. That's good (because fair use requires knowing where the image is and it used to be hard to find those uses after the image was removed) but bad because it doesn't notify those who are best placed to both know the status of the image and find a replacement. Adding a requirement of a notice on the talk page of the article seems a suitable replacement. While there's ongoing work on improving the reliability of the links to images, it's always necessary to use an image search to verify lack of use before deleting. Jamesday 22:24, 21 Jan 2005 (UTC)

Deletion bugs

If you encounter an error - "Can't delete this article because it contains block-compressed revisions. This is a temporary situation which the developers are well aware of, and should be fixed within a month or two. Please mark the article for deletion and wait for a developer to fix our buggy software." - when trying to delete a page, replace its content with {{pending deletion}}, and protect it until the bug is fixed. Rdsmith4— Dan | Talk 01:25, 18 Jan 2005 (UTC)

What are "block-compressed revisions"? Has anybody reported this bug? RickK *** 05:47, 18 Jan 2005 (UTC)

It's not a bug. It's a glitch :-) Old revisions (I think all those prior to December 2004) have been compressed in the database to save space. As a side effect, this makes deletion of revisions before Dec 2004—and hence articles created before that—impossible for the time being. A future software release will contain the fix that allows deletion of compressed revisions, and then such articles can be deleted. Lupo 07:44, 18 Jan 2005 (UTC)

James Day posted something about this a while back on the OpenFacts Wikipedia status board:

The compression of article old revisions is in progress. [...] The most visible sign after this has happened is likely to be the temporary inability to delete articles created earlier than 1 December 2004, done to ensure that there won't be any data loss with selective undeletion before full support for that is present in the software. The temporary workaround is to blank and protect the article. Developers will deal with any cases where actual legal action makes that insufficient. JamesDay 09:03 Jan 6, 2005 (CET)

So I guess we'll have to live without delete for a while. Think of it as seeing how life is for ordinary users! :-) Noel (talk) 14:42, 18 Jan 2005 (UTC)

It's not a bug. It's Tim Starling's sense of humour in the notice. Old article versions for en are about 80GB. Compressing one article at a time takes that to about 40GB. Combining many article revisionss into one record and compresing the combination is expected to reduce that to about 15GB. The new feature to undelete only some revisions doesn't understand the block compressed format so deletion of articles with block-compressed history was turned off to prevent selective undeletion causing data loss (and missing body text in the deleted article revision history or the restored history). Both features were desired so limiting the deletability of some articles for a while seemed like the best approach, compared to delay for the selective undeletion feature. It'll probably be taken care of in the next software revision, if not sooner. Jamesday 14:42, 1 Feb 2005 (UTC)

These have been popping up in WP:CP and elsewhere. Perhaps we should start something like Wikipedia:Compressed history articles for immediate deletion when such becomes possible for our articles where deletion has already been decided on but have this glitch? - Infrogmation 18:06, 9 Feb 2005 (UTC)

Ah, that's what Template:Pending deletion (see above) does - it adds them to Category:Pending deletions. Noel (talk) 18:26, 9 Feb 2005 (UTC)

Cool; thanks. -- Infrogmation 17:57, 11 Feb 2005 (UTC)

Regarding these deletion bugs, the VFD page has a log list as well, at Wikipedia:Votes for deletion/Old/Block-compress errors. If you happen to be deleting VFD articles, please add any deletion bugs to that page. It is a quick reference for pages that gone through VFD. -- AllyUnion (talk) 15:24, 26 Feb 2005 (UTC)

Which Disruptions of VfD Merit Re-Listing?

In the process of closing discussion on Wikipedia:Votes for deletion/Tom Sutter, i noticed that the VfD-Cat tag (added as part of the VfD template) had been removed from Tom Sutter by an anon ed (Summary: "reinstated 'irrelevant stuff' as it is not irrelevant it is his entire hitting career!") in the process of an otherwise verbatim-et-literatim restoration of old material.

• Arguably, consulting the Cat page is a viable approach to seeing what is on VfD, avoiding the voluminous VfD-page mechanisms, and this action reduced exposure via the Cat to 18 hours instead of 120 or so.
• That alternate means of accessing VfD debates may never have been used, but we could only know that by polling everyone who ever votes on VfD.
• This particular vote was close (6 D, 4 K), so two more votes could have shifted the result.
• I prefer not to be the one to make the decision to give it another 5 days (especially without consultation) since i voted on what is otherwise the losing side.
• Where i've left it is with my closing rescinded, pending consultation (here).

--Jerzy(t) 06:50, 2005 Jan 20 (UTC)

Hmmm. Although I'm something of a deletionist, I'm always in favour of taking enough time to make sure we have rough consensus. Generally, it doesn't matter if we wait a few extra days to delete something, if there's no harm involved (e.g. libel or other problematic material). So, in this case, I'd say there's no harm in waiting, after adding it back to the category. Noel (talk) 13:30, 20 Jan 2005 (UTC)

On reflection, i find myself more concerned about whether this is part of any pattern than with handling the individual instance correctly. Have others noticed removals of the VfD-Cat tag? Do others think they would have noticed them if they've been occurring (or is my noticing the result of my meticulous attention to detail? Is this worth adding something like the following to Wikipedia:Deletion process?:

Before marking the VfD subpage closed, please note whether the Category:Pages on votes for deletion tag is intact; if not, please either skip closing that debate, or note the fact at Wikipedia:Administrators' noticeboard#Which Disruptions of VfD Merit Re-Listing? during your closing process.

--Jerzy(t) 18:08, 2005 Jan 20 (UTC)

I wouldn't refer to this specific header (#Which Disruptions of VfD Merit Re-Listing?) because it's likely to be gone (archived) in a few days. Noel (talk) 12:29, 21 Jan 2005 (UTC)

Vfd tags are often removed from pages, but generally within a few hours of tagging (which itsself is usually within a few minutes of page creation). I think in almost all cases they get restored pretty quickly (most people have "add edited pages to watchlist" turned on, so the removal turns up in their watchlist), but there may be some that slip through the net. --fvw* 18:17, 2005 Jan 20 (UTC)

I routinely spot and revert instances of anon vandalism to VfD subpages or tags when in RC patrol. If we were to re-list or even extend the voting time in every case it would be a maintenance nightmare. I'd say it is enough if the entry remains listed in the main VfD page and the subpage stays readable most of the time; in cases of persistent vandalism that seriously hinders voting, allow it remain listed for extra few days after blocking/reverting. Your suggested wording in policy would just encourage trolls to stealthly remove the cat. tag and then pester everyone with their claims for extra voting round, admin abuse, or trolling in VfU. jni 12:18, 21 Jan 2005 (UTC)

Wikipedia:Tutorial (Editing) and the Wikipedia:Sandbox

Is there any reason why the Wikipedia:Tutorial (Editing) has a subpage Wikipedia:Tutorial (Editing)/sandbox and why it is not pointing to Wikipedia:Sandbox? -- AllyUnion (talk) 08:32, 22 Jan 2005 (UTC)

I wondered about that myself. I don't see the subpage as doing any harm, but I wouldn't protest if it was redirected. --Slowking Man *** 03:11, 24 Jan 2005 (UTC)

The first five pages of the tutorial (not including the front page) have had their own sandboxes since shortly after they were created last Spring. They don't seem to do any harm, and probably reduce the chances of edit conflicts. I don't remember if there was any other, more specific reason it was done--that was a long time ago in Wikipedia/Internet time ;). Niteowlneils 21:53, 29 Jan 2005 (UTC)

I think those subpages should include the same content as the main Sandbox itself... maybe. It's just... who's cleaning these subpage sandboxes then? -- AllyUnion (talk) 13:30, 30 Jan 2005 (UTC)

One reason was to avoid conflicts, since the main sandbox could potentially get pretty crowded. Also, I thought it was useful for newbies to be able to see other people's experiments; each sandbox has a theme of sorts. Since it's been brought up, I'll try to remember to clean them out occasionally. Isomorphic 19:34, 4 Feb 2005 (UTC)

I added the sandbox template to that sandbox. -- AllyUnion (talk) 02:00, 5 Feb 2005 (UTC)

I'm not sure what happened (maybe a newbie deleted it?), but it doesn't seem to be there anymore. Anyway, I added these all to Wikipedia:Cleaning department. Noel (talk) 22:27, 5 Feb 2005 (UTC)

Sockpuppet detection

Does anyone know if a feaure request has been filed to allow admins access to IP addresses? I took a look, but this one was the closest I could find. (I think the rough consensus in our previous discussion of this was that access to IP's - perhaps suitably hashed, so people could check for equivalence without seeing the actual location - was a good idea, and something that was done on other sites - e.g. Live Journal - with no problem.) If not, one should be filed (I'll do it if there isn't one already). Noel (talk) 13:49, 25 Jan 2005 (UTC)

IPs are not commonly available for a very good reason: having them readily available would facilitate harassment of users one disagrees with. Even to admins is too large a pool. The devs keep IPs under their hats for this reason, and only give "yes, no, probably but not certain", etc. responses to questions on sock puppetry to protect the users in question. User:Alberuni's IP got out and was discovered to likely be a workplace IP; I know more than one person spoke (idly) of collecting his posts and sending them to the workplace in question. That would be a BAD, BAD practice to facilitate - David Gerard 14:47, 25 Jan 2005 (UTC)

This problem exists now, for people who edit not-logged-in. (Yes, yes, I know, they can get a user-id to hide the IP.) Also, my note wasn't about making them "commonly" available anyway (but I see you consider admins too large a pool too). Noel (talk) 15:56, 25 Jan 2005 (UTC)

That problem could be solved by hashing IP adresses, as Jnc proposed above - all you could see would be some sort of unique identifier, the real IP adresses would always remain hidden. That should work to eliminate your legitimate concerns, and the ability to spot sockpuppets would be of great help indeed -- Ferkelparade π 14:58, 25 Jan 2005 (UTC)

Err, that wasn't my idea - I was merely summarizing something someone else said in the earlier discussion. Noel (talk) 15:56, 25 Jan 2005 (UTC)

If the hash function is known and it produces a unique number for each IP number (enabling sock puppets to be spotted) it follows that there is a one-to-one correspondence between an IP number and a hash. This would be easy to decrypt (especially with MediaWiki being open source code). --Tony Sidaway|Talk 15:14, 25 Jan 2005 (UTC)

There are technical ways around this (use a non-invertible hash function, and also add a "salt" to prevent people compiling a dictionary - you make the salt space large enough that people can't find it by an exhaustive search). Noel (talk) 15:56, 25 Jan 2005 (UTC)

Precisely, which is why I advocate that only bureaucrats be able to access IP addresses. If people have a problem with existing bureaucrats being able to view IPs, then make the feature only available to new bureaucrats. This can easily be done with the new permissions system. Johnleemk | Talk 15:47, 25 Jan 2005 (UTC)

I see I mis-understood the sense of the earlier discussion. There is no consensus for making them available. So we'll have to continue to rely on developers, if and when they get the time to do this, I guess. Noel (talk) 15:56, 25 Jan 2005 (UTC)

I accept that there is no consensus to publish IPs. But I am not convinced by David's argument above at all: Alberuni of all people is a perfect example why making public IPs could be a good thing: It could encourage people to behave. Many people edit under their real names. You can find out my home address in two minutes, if you like. As I see it, the editors we are interested in are the ones who take pride in their work. The trolls and sockpuppeteers could still post from public IPs, but I don't see the need to go out of our way to protect their anonymity. If I visit a website from work, well, I am aware I give away my IP, and fully expect it to be listed on a server log somewhere. dab (ᛏ) 16:10, 25 Jan 2005 (UTC)

You could try making a non-invertible hash function but this woudn't help much because the search space is necessarily extremely small (the set of registered dotted quads, which collapses to a tiny set of candidate' dotted quads in a real life situation) and candidate guesses can be entered. The salt could be deduced, the algorithm is known and an IP number always corresponds to the same hash value. --Tony Sidaway|Talk 16:15, 25 Jan 2005 (UTC)

The "salt could be deduced" - ah, no.

First, the fact that "the algorithm is known" is irrelevant. These algorithms are often published openly (see, e.g., RFC 3174 "US Secure Hash Algorithm 1 (SHA1)"). (It's called a "one-way .. function" for a reason!) So you can't deduce the salt from the resulting hash - and having part of the input (e.g. by looking to see what your own IP turns into) doesn't help you either - you cannot deduce the input from the output, even knowing the algorithm. (The transformation is not reversible.)

Second, if you make the salt long enough, you can't figure it out via a dictionary attack (i.e. try every possible value of the salt, to see which one works - which is why I said "make the salt space large enough that people can't find it by an exhaustive search" above). E.g. if the salt is 992 bits long, to which the IP address is appended, producing a 1024 bit result which is then hashed, then it is still the case that knowing one 32-bit-input/output pair won't help you. Noel (talk) 17:17, 25 Jan 2005 (UTC)

I think some of the misunderstanding here is due to your use of the word salt. In most cases, salts are public, which they are explicitly not in this case. I think a better term is key, even though that doesn't fit entirely either as under normal operation the key is never used for decryption or verification. Anyway, since the last discussion I've realised that the only way I'm going to get this on wikipedia is by coding it and then starting the discussion on whether we should turn it on, so that's what I'll do. My Wikipedia-plate is a bit full at the moment though, so I haven't started yet. --fvw* 21:17, 2005 Jan 25 (UTC)

Well, generally, a salt is a value which is added to the process of computing one-way hashes in order to hinder dictionary searches and/or precomputation of an output_value -> input_value reverse-mapping table (by means of precomputation on all possible input values). So it's pretty well on-target here, I think. It's true that in the most well-known case (Unix passwords) the salt is public, but in our specific application I don't quickly see a way to do what we want (allow comparing hashed versions of IP addresses, without being vulnerable to pre-computation reverse lookups) with a public salt. However, maybe I'll think one up in the morning when my brain is fresh! :-) Noel (talk) 23:20, 25 Jan 2005 (UTC)

If IPs are encrypted then how will we be able to tell if they're using an open proxy? -Frazzydee|✍ 18:36, 25 Jan 2005 (UTC)

We won't. Keep in mind this isn't intended to replace IPs as identifiers for not-logged-in users, this is just an extra piece of information about logged-in users, so you can tell if POVpusher01 and POVpusher02 are really just editing from the same IP. Incidentally, my original proposal didn't include just a non-reversible hash of the IP of the editor, but also a non-reversible hash of the editor's /20 (the first 20 bits of their IP), to prevent people just dialing in again to prevent detection. It's not perfect, but I think it'll be an improvement over what we have now. --fvw* 21:38, 2005 Jan 25 (UTC)

Well, you'd want to make the hash of the IP's for NLI users available too, so you could see if anon A.B.C.D and user X were the same person. Noel (talk) 23:20, 25 Jan 2005 (UTC)

FWIW, the developers are actively working on how to identify matching IPs in a way that doesn't reveal the IP. And on a solution to the AOL problem (where any two successive accesses from the same user might go through different proxies) - David Gerard 08:39, 26 Jan 2005 (UTC)

Ehm, the AOL problem cannot be solved without cooperation from AOL. Yes, you can do some tricks with cookies and such, but there is nothing that will solve the AOL problem against someone who has half a clue. (Yes, I know, AOL. But still). --fvw* 16:14, 2005 Jan 26 (UTC)

I don't think I made myself particularly clear. I know that security algorithms are often public, and that a salt value, which can be kept secret, can be very large indeed and will indeed be proof against a naive dictionary attack. In this particular class of security algorithm, however, the dictionary size is tiny and the utility of the salt is constrained by the requirement that the hash must always be the same value for a given input, to enable duplicate IPs to be spotted without access to the salt. --Tony Sidaway|Talk 16:00, 26 Jan 2005 (UTC)

Using SHA-1 padding of say 128 bits random secret data concatenated with 32 bits of IP, even if you knew the hashes of 2³²-2 IPs, you wouldn't be able to determine which of the remaining two IPs was witch, given their hashes. If you think this would be possible, please suggest a procedure that would determine which is which. --fvw* 16:14, 2005 Jan 26 (UTC)

This is secure for hashing IPs. Basically, we hash something like "Some really long phrase you'll never guess in a zillion years fhwuefhihqi:10.2.3.4" when we hash the IP 10.2.3.4. They know this hashes to "e4bdf4208373ad7b". The attacker knows the "10.2.3.4" part of the plain text. They don't know the "Some really long phrase you'll never guess in a zillion years fhwuefhihqi" of the plain text. Until they do, they can't figure out what "10.2.3.5" or anything else will hash to. This is secure: The SHA1 hash of the combination of a passphrase that you don't know and my IP is "9c5120896f045ad9accf0c4fc96d740b327949e3". The SHA1 hash of the same passphrase and the ip 10.2.3.4 is "c537137cdb0e178e0564b26298f7f503ef974121". What is my IP? You don't know until you know the passphrase, and do a large number of hashes. Samboy 09:47, 11 Feb 2005 (UTC)

Why couldn't a feature just be created that allows a person to type in two usernames and have it tell you if the ip's are the same? They can't find the person's ip unless they type in many many combinations. If that is an issues you could limit it to 1 query a day or somthing. You can tell I don't know much about this programming stuff, can't you. BrokenSegue 01:55, 28 Jan 2005 (UTC)

That is a possible solution too (in combination with a "They're in the same IP range" indicator), however the system where a hash of the IP is published you'll be able to find all sockpuppeting users in a group of n POV-pushers in n operations instead of ½(n²-n) operations (whether this is a good thing or a bad thing is a matter of taste ofcourse. I just happen to think that my taste in this is the correct one). An added problem of having a form working on users is that an IP is not associated with the user, but with an edit. Each edit can have a different hash-of-ip. --fvw* 17:11, 2005 Jan 28 (UTC)

Is finding sockpuppets a big problem for Admins? At least from my view on the outside, most of the sockpuppets look pretty obvious, as if they want to be caught. For the much smaller number that are questionable, don't most of the questions concerning sockpuppets revolve around enforcement of Arbitration Committee decisions? If that is true, then would it be better to have one (or perhaps two) of the Abritrators given access to developer priviledges? Just curious. BlankVerse 10:18, 4 Feb 2005 (UTC)

Sockpuppets are often pretty obvious. Unfortunately, what is obvious to someone familiar with the user is often not obvious to others, so it would be nice to have an easy way to get solid evidence. Isomorphic 19:38, 4 Feb 2005 (UTC)

Actually, I think this is a misinterpretation of the evidence. The sockpuppets we catch are rather obvious. This is why we catch them. Other sockpuppets go undetected for ages (the Gzornenplatz/Wik thing comes to mind, though it's not quite sock puppeting since he wasn't editing with both) or for ever. --fvw* 19:57, 2005 Feb 4 (UTC)

Hmmm.... I wonder how hard it would be to include a GPG key system into the MediaWiki software... -- AllyUnion (talk) 20:19, 11 Feb 2005 (UTC)

Images protected while on the main page

After going through the protection log, I've run into a fair amount of images such as Image:Carson.JPG, Image:Caligula bust.jpg, Image:Rtm witold pilecki-aus.jpg, Image:Washingtonia filifera.jpg, Image:Basketball game.jpg, Image:A380.emirates.736pix.jpg, Image:Burma flag(1948) large.png and Image:Zhao.jpg that are still protected after falling off the main page. Would anyone object if I created a template and category so that we can keep track of these more easily? I don't think I'm missing any other reason why these images are protected, but let me know if I am. - RedWordSmith *** 21:12, 25 Jan 2005 (UTC)

Sounds fine by me. BrokenSegue 21:41, 25 Jan 2005 (UTC)

Yeah, good idea. --fvw* 21:44, 2005 Jan 25 (UTC)

Go for it. -- Jmabel | Talk *** 22:20, 25 Jan 2005 (UTC)

That sounds like resounding support, so the template is template:mpimgprotected and the category is category:Protected main page images. - RedWordSmith *** 23:59, 25 Jan 2005 (UTC)

Hm. I just checked the main page, and the second image I clicked wasn't protected. Also, I see that the sister project logos at the bottom aren't protected. They wouldn't be much of a liability, as they're only 35 pixels, but in Template:WikipediaSister, the image sizes aren't included in the source. So if one of these is replaced with something huge ... dbenbenn | talk 00:22, 11 Feb 2005 (UTC)

Autoblock

Autoblocks confuse me. Why is it that some username blocks (see e.g. "Mother Larry" on Special:Ipblocklist) have no corresponding autoblock, while others (e.g. "OneGoy") do? Lupo 12:48, 27 Jan 2005 (UTC)

Ok, answering my own question: the autoblock becomes effective the next time the blocked user tries to edit. Hence, if the user doesn't edit, only the username block shows up in the log. Lupo 13:19, 27 Jan 2005 (UTC)

Blocking shared ips

What's the policy for blocking shared ips? There are a few which have the shared template on their talk page, but all of their edits seem to be vandalism. There are several that would qualify for a 24 hour block (repeated vandalism, ignoring test3 messages, etc), but I'm hesitant to do so because I don't know what the policy is. Can someone clarify? --DropDeadGorgias (talk) *** 20:34, 28 Jan 2005 (UTC)

I Don't know about policy but my appraoch is that as a general rule the more people who are likely to be using the IP the shorter the block. For example I will block school IPs for an hour or less.Geni 20:49, 28 Jan 2005 (UTC)

Similar to Geni, if I see what I think is a shared IP vandalising, I typically block for a couple of hours, just long enough that they get bored and (hopefully) give up. Jayjg ^(talk) 20:52, 28 Jan 2005 (UTC)

24 hours is what *I* would consider the max for school-type shared-IP vandalism... unless, of course, it becomes a regularly recurring problem, in which case we may want to consider permanently blocking the relevant IP/range. --Dante Alighieri | Talk *** 21:35, 28 Jan 2005 (UTC)

It isn't just schools, of course, that share IPs. The many users within a company may share one or a few external IP addresses as well. And within a home, multiple computers may share a single public IP address, so if, for example, a child is banned from Wikipedia, an IP ban may also extend to the hapless parents and siblings of that child.

Public access terminals (such as at libraries) also present the situation where many Wikipedia users share a single IP address (and even a single browser and cookie set!).

Atlant 16:44, 31 Jan 2005 (UTC)

See (205.222.240.2 | talk | contributions), school ip, one month of nonsense. Seems to be just one person. The first 24 hour ban probably just gave him a little thrill. A longer ban might help modify this young person's habits. Duk 00:51, 3 Feb 2005 (UTC)

I don't think there's any formal policy on shared IPs, or if there is I don't know about it. If a public terminal shows a long-term pattern of being used primarily for vandalism, I would support a lengthy block. Isomorphic 20:32, 4 Feb 2005 (UTC)

Protection warning

I created my own protection warning at User:AllyUnion/Protection_Warning. I would appreciate it if several administrators could remark on it or whether it needs any improvement by leaving a note on my talk page. -- AllyUnion (talk) 13:32, 30 Jan 2005 (UTC)

re-dialling to dodge 3RR

see User talk:83.109.185.152 (norwegian ISP): 3RR-aware anons now just dial in again after their third revert. How to address this? We cannot block the entire IP range. I imagine short (30 minute) blocks whenever the anon makes another edit will bore them away. I left it at a warning for now, though. dab (ᛏ) 17:28, 30 Jan 2005 (UTC)

Yeah, the same problem occurs on AOL and networks with large DHCP pools. I usually don't even bother blocking, it's just going to annoy innocent bystanders. Blanket-reverting is a more appropriate way of dealing with these things I think. Perhaps we need a "list of blocked but technically unblockable users anyone can revert on sight." --fvw* 00:26, 2005 Jan 31 (UTC)

What if they start making good edits? Everyking 00:30, 31 Jan 2005 (UTC)

I meant as a replacement for a block, so say blanket-reverting for 24 hours. Even if they start making good edits, though those reverting are always free to ignore good edits if they want ofcourse. --fvw* 00:49, 2005 Jan 31 (UTC)

We could require people who edit from DHCP pools that are the source of lots of problem to acquire user-id's and log in to do edits; i.e. disallow edits from non-logged-in users for some address ranges. I know this goes against previous practise, but you can be just anonymous with a user-id as you can with an IP address, and if I have to trade off not being able to control anon's from DHCP pools versus this very incremental loss of anonymity, that's an easy call to make. Noel (talk) 05:49, 2 Feb 2005 (UTC)

It's a solution, and not one I'd object to (but then again I'd also like to block the AOL proxies); I doubt you'll get anything close to consensus for it though. My preferred solution is to make all edits of blocked users freely revertable by anyone (as is currently the case with hard banned users. This would probably be more effective too, as there's no way to block people entirely even when disabling IP pools (dialup accounts are pretty much free these days). --fvw* 05:53, 2005 Feb 2 (UTC)

Just an FYI -- I saw a Feature Request (#550) at Bugzilla about this. — Catherine\^talk 02:10, 21 Feb 2005 (UTC)

Another solution is not to count the revert of an edit by someone who is not logged in for 3RR purposes. An anon editing a controversial article might then have an incentive to register and log-in in order to enjoy the "protection" of the 3RR. Indeed it seems a bit unfair to count reverts of anonymous edits for 3RR purposes, since this means that anonymous editors actually enjoy a considerable advantage over registered editors in an editing dispute. A registered editor is constrained by the 3RR, but the anon's are not: an anon can either simply overstep the 3RR knowing that he can't effectively be "banned"; or he can simply dial-up again and get a different IP in order to do the next revert. --BM 02:17, 11 Feb 2005 (UTC)

That's a good point, and I personally would agree with this. However, I suspect you're going to get push-back from people who don't want to treat anon editors as second-class citizens. Also note that an editor with a user-id can pull the same stunt; come back as an anon for 3RR warring purposes, and use multiple dialups to exceed their quota. So it's probably not worth the trouble to change the policy. Noel (talk) 13:49, 20 Feb 2005 (UTC)

Can anon editors get their IP encrypted into their cookie? If they redial to avoid a block mediawiki can identify their last used IP and autoblock. This would be easy to circumvent, by deleting cookies, but it might catch the really dumb vandals (probably a majority of vandals). EyeBall 18:35, 20 Feb 2005 (UTC)

That sounds like a good idea to me. It would be easy to circumvent if you knew about it, but of course we wouldn't tell them "You're blocked because of a cookie you have. You might want to delete it." :) dbenbenn | talk 19:08, 20 Feb 2005 (UTC)