Forum:Protecting the exchange namespace

From the RuneScape Wiki, the wiki for all things RuneScape
Jump to: navigation, search
Forums: Yew Grove > Protecting the exchange namespace
Archive
This page or section is an archive.
Please do not edit the contents of this page.
This thread was archived on 31 August 2009 by Azaz129.

I've noticed that a good chuck of IP vandalism here comes from blanking exchange pages, since these are mainly updated by bots now, it may be a sensible idea to semi-protect the pages in this namespace to prevent IPs and accounts less than 4 days old from editing them. This would force them to go through the trouble to create an account and logging in every time, which will significantly cut down on the effort from sysops and rollbacks here and better put that time towards improving the wiki. Any thoughts? --CharitwoTalk 13:21, 3 August 2009 (UTC)

Go for it! There is always {{editprotected}} Rich Farmbrough, 13:25 3 August 2009 (GMT).
I am wary about protecting the pages on second thoughts, because it stops IP's who want to get an update version of the Calculator pages from fixing the prices. Some of these are sensitive to price, in terms of profit/loss. Rich Farmbrough, 13:37 3 August 2009 (GMT).

That seems like a lot to do. O_o. Maybe we can ask Azliq and her Azbot since Azbot has sysop rights? Santa hat.png Powers38 おはようヾ(´・ω・`) 13:28, 3 August 2009 (UTC)

It's very easy to do with a bot+sysop. Just a matter of whether the community wants such a change. --CharitwoTalk 13:31, 3 August 2009 (UTC)

OK here is another suggestion. Use an editfilter. This can prevent blanking in a given namespace. It could also check for non-numeric characters in the price fields and misformatted dates. Rich Farmbrough, 13:35 3 August 2009 (GMT).
Abuse Filter? That's another option. --CharitwoTalk 13:39, 3 August 2009 (UTC)
I agree that semi-protecting the entire Exchange namespace wouldn't hurt the community and would stop a lot of vandalism. ShinyUnown T | C | E 13:47, 3 August 2009 (UTC)
With the editfilter how about we prevent the edit and warn the user to not make bad edits as it is considered vandalism and show the help on how to update articles? - TehKittyCatTalk Wikian-Book 14:26, 3 August 2009 (UTC)

How about adding a new permission "exchangeedit" for editing the Exchange Namespace and giving it to the autoconfirmed users group? I think Abuse Filter would be an interesting solution though. Since that could be used to reduce vandalism overall not just the exchange pages. - TehKittyCatTalk Wikian-Book 13:51, 3 August 2009 (UTC)

If I'm not mistaken, we've had this discussion before. If I remember correctly, there was no consensus to protect the exchange pages as some believed that protecting the exchange pages wouldn't change much as the vandals would probably vandalise mainspace articles instead. I'll try and find it. C.ChiamTalk 14:02, 3 August 2009 (UTC)

Found it. Forum:GE Pages. C.ChiamTalk 14:10, 3 August 2009 (UTC)
The main issue with it last time was malfunctioning bots, the solution to that is when a bot malfunctions, ban it temporarily. - TehKittyCatTalk Wikian-Book 14:26, 3 August 2009 (UTC)
I'm not getting you. The issue that was being discussed was whether IPs should be allowed to edit GEMW pages, I don't see the relation to malfunctioning bots. C.ChiamTalk 14:46, 3 August 2009 (UTC)
The relation was that if the bots are causing the equivalent of vandalism why protect the pages and the fact the IPs could fix the bad edits by the malfunctioning bots. - TehKittyCatTalk Wikian-Book 14:54, 3 August 2009 (UTC)
Oh, I see. That's one of the reasons mentioned as to why we shouldn't protect the exchange namespace, but malfunctioning bots are only temporary issues anyway. What about the other issue about vandals moving on to vandalise mainspace articles instead? C.ChiamTalk 08:00, 4 August 2009 (UTC)

Comment - I don't know. I'm kind of neutral about this issue. Sure, it will reduce vandalism, and make it so that if vandalism goes unnoticed, it won't harm any of the pages that do use the page. The thing I am worried about is the Category:Needs price update, and the three subcategories. Some IPs might decide to correct some of these. Even if the price didn't get updated since, it might be okay just to change the date to show that no change has shown. The abusefilter idea sounds nice to me, as it would still allow users to edit these pages. ~MuzTalk 14:36, 3 August 2009 (UTC)

The cateogry Needs price update was for a while obsolete, as SmackBot did a weekly or twice weekly update of all prices, even if they hadn't changed. However this meant often updating something GEBot had just updated minutes before so I suspended that function.

Some prices rarely change. Comments on this welcome. Rich Farmbrough, 22:11 5 August 2009 (GMT).

Ok I set SmackBot to update everything on Thursdays, this s good because there are circumstances where a page can become "uneditable", and this can be relatively hard to detect. Rich Farmbrough, 22:31 5 August 2009 (GMT).

We could hide the categories using __HIDDENCAT__, since no one has any reason to use it except for bots these days. - TehKittyCatTalk Wikian-Book 14:41, 3 August 2009 (UTC)

Oppose - we are a wiki, where anyone can edit. We need to assume good faith that IPs are here to help the wiki as well, which includes assuming good faith in the Exchange: namespace. Butterman62 (talk) 16:13, 3 August 2009 (UTC)

Oppose - per RS:AEAE and RS:AGF. Andrew talk 17:13, 3 August 2009 (UTC)

Oppose protecting, support AbuseFilter - Per Butterman and Soldier. I like Rich Farmbrough's suggestion of editfilter to prevent page blanking and prevent/warn on incorrect updates. AbuseFilter could also have a lot of other uses on the wiki, maybe we could prevent page blanking on main, exchange, forum, and the update namespaces. - TehKittyCatTalk Wikian-Book 18:09, 3 August 2009 (UTC)

Support AbuseFilter; oppose Exchange protection - per Soldier 1033 and Butterman.

Here's the link to the extension: AbuseFilter. And here's the example of its special page: Special:AbuseFilter. It is quite a powerful administrative tool, and some of the actions it may perform are:

  • Warning (warns the user that their edit "may not be appreciated")
  • Throttling
  • Disallowing (prevents users from doing something, i.e. blanking the page)
  • Blocking and range-blocking (blocked automatically when a specific filter is triggered)
  • etc.

Like Rich said earlier, this tool can be "programmed" to detect anything: formatting errors, common British spelling mistakes, Runescape instead of RuneScape. I think the amount of work for rollbackers and the Counter Vandalism Unit will be significantly reduced if we installed this extension. Smile   az talk   19:05, 3 August 2009 (UTC)

Support AbuseFilter - Per Az. ShinyUnown T | C | E 19:19, 3 August 2009 (UTC)

(already stated oppose to protection above), Oppose AbuseFilter - "Powerful" doesn't mean "good". The big problem I have with these types of tools is that they are automated; what we are planning now is not so bad (blocking blankings to Exchange: namespace), but I'm worried that it will set a dangerous precedent. We may end up setting a thousand rules of what is allowed and what isn't, judged at the whim of a computer, which will surely discourage both newcomers and regular users. In their full power, AbuseFilters can disallow edits, block users, and even remove user rights (for example, sysop and rollback) with only a computer making the call. Computers cannot assume good faith.

But even with this one rule, we cause problems. For example, seeing that we're on the subject of blanking exchange pages, let's say some vandal created a fake exchange article called "Exchange:Brown dye" and filled it with fictitious information about prices and such. Then, a second IP comes along and blanks it, because it realizes that such content is misleading, but doesn't know about speedy-deletion tags and the like. A good-faith-assuming user would recognize that the IP was trying to help by removing the offending content, and said user would take notice and tag the article with a speedy deletion template. An AbuseFilter would, most likely, disallow the edit, post an automated vandalism notice to the IP, and block said IP, all with no human intervention. Do we see a problem here?

We must assume good faith that, unless such a user or IP has a precedent of vandalism and is given reason to assume bad faith of, any edit to the wiki is made for the purposes of helping us out. We have many intelligent users with varying levels of technical tools to handle edits; while some edits are blatantly unconstructive, others might have the tiniest bit of merit. Our users are real people; they are capable of, instead of casting off an edit as crap, trying to understand why a user made a certain edit or find nuggets of useful changes, and accomodate those changes into an article, or even recognize that a huge section blanking or other might have been necessary.

An automated filter cannot operate on the same level. Therefore, I oppose. Butterman62 (talk) 01:29, 4 August 2009 (UTC)

For those interested, see here for what these edit filters can do. Some of these capabilities are downright shocking. Butterman62 (talk) 01:32, 4 August 2009 (UTC)
It is only as not assuming good faith as we make it. I do not think we would automatically ban. But, we could prevent or at least warn when a Exchange page is blanked(and maybe auto ban bots that do it since they would be malfunctioning?). When incorrect changes are made to the exchange we could "warn" them and tell them how to correctly update the exchange since they didn't follow it correctly. A warning can be assuming good or not just by the wording of it. During the days of mass vandal bots we could have stopped it all with this. - TehKittyCatTalk Wikian-Book 17:03, 4 August 2009 (UTC)

Oppose all - Per Butterman. Also I seem to remember we are the fansite that anyone can edit. (Freely).--Joe Click Here for Awesomeness 08:12, 4 August 2009 (UTC)

Oppose all - As said on the Protection page,we shouldn't protect pages pre-emptively,or we shouldn't potect pages if they don't have a bad history of vandalising. besides,there must be hundreds of GEMW pages,we simply cannot protect them all....And I don't trust bots...What? You can't blame me..... >.>--

Water Wave icon.png
Captain Sciz
TalkEditsHiscores File:Runecrafter hat.png|link=

17:17, 4 August 2009 (UTC)

Oppose all - I don't trust bots. They may be useful or just plain silly (like the IRC bot I made) but for whatever purpose they serve, there's always going to be an error. And this bot detects bad edits and auto-blocks? What if someone made a good faith edit and didn't know what the heck they were doing? They need to be taught how to edit properly, not recieve a block. --— Enigma 17:34, 4 August 2009 (UTC)

It is not a bot, its a mediawiki extension. It only bans if you set up that action to ban. You could set warnings to give a helpful message rather than threaten. - TehKittyCatTalk Wikian-Book 17:44, 4 August 2009 (UTC)

Reply @opposes: Catcrewser is right. The ability to block/ban/whatever depends on the settings we set. We will decide what filters to use, when to use it, and whom it applies to. The filters are not permanent, as they can switched on and off depending on the situation. For example, if we had an "automated vandal" replacing pages with "w00t", we could set AbuseFilter to filter "w00t" edits until the vandal has been banned. This extension is quite flexible, and not an automated zombie bot. If banning is not desirable, we can always switch it off. If we want, we can set AbuseFilter only to warn users when they have triggered a filter. By the way, all of AbuseFilter's actions are logged, and anyone can access these logs to see what is the filter is, who had triggered the filter, what edit triggered the filter, and the action taken by AbuseFilter.

For an example, see AbuseLog.

As mentioned in the Wikipedia page, "When the extension is initially installed, the available actions will not include blocking or removing from privileged groups. This restricted usage has been (should be) determined by community consensus, and if the extension is successful, the community may decide to enable the block, rangeblock or degroup actions for use on this wiki."

IMO, this extension is an useful tool that sysops can use to deal with vandalism. I say we should at least try it first to see how it works before dismissing the idea outright. If it doesn't work for this wiki, we can just uninstall it.   az talk   06:02, 5 August 2009 (UTC)

One worry I have about the warning system though; we are all experienced users who will probably know about the AbuseFilter if/when it comes up. If we see it come up while we're making a good-faith edit, we'll just think, "ignore it". IPs/very new users will not know that. They may go make an edit, get a warning, and think, "oh, the wiki hates me! They hate this edit!", and then we've scared off another newcomer.
While I would support the temporary filters (for example, if we're being attacked by many different IPs who are all blanking pages, prohibit non-autoconfirmed users from blanking pages for an hour), I would not like to see permanent filters on the wiki, for the reasons above. Butterman62 (talk) 14:36, 5 August 2009 (UTC)
It depends on how you use it. For example the filter can merely log. So two scenarios
A. Someone sets the price of an item to blank. The filter leaves a message which says ZOmG One!!111 vandle U! Elven!11Wun "You set the price of an item as blank, can you check you meant to do this?"
B. someone moves the price of an item over a power-of-ten boundary, the filter silently logs the fact for checking.
Neither of these are likely to scare off a user.
Rich Farmbrough, 22:03 5 August 2009 (GMT).

For the record, AbuseFilter is already installed on Wikia, just not enabled by default. It can be requested with consensus. --CharitwoTalk 05:30, 6 August 2009 (UTC)

Comment - Whether it's a bot, system extension, I don't care what you want to call it. Programmed things can always go wrong and are bound to eventually. — Enigma 07:12, 6 August 2009 (UTC)

Like using the "save page" button on this wiki! And the routing and DNS software that gets you here - and your keyboard is programmed... and and and... Rich Farmbrough, 23:29 6 August 2009 (GMT).

Support AbuseFilter - Per Az. ~ Fire Surge icon.png Sentry Telos Talk  09:13, 6 August 2009 (UTC)

Support AbuseFilter - Per Az. Should it start messing up, then we do have plenty of admins to disable it if needs be. Quest.png Gaz Lloyd 7:^]Events!99s 11:41, 6 August 2009 (UTC)

Comment - While I can see the convenience of this filter, I must ask is this really necessary ? We dealt with vandalism just fine for years without the need for such filters and the way I see people discussing vandalism on the exchange (or any namespace even) here is if we have just been hit with a vandalism rate of vandal bot proportions. Therefore I must wonder again, is this filter really necessary or do we just think it is?--

Helm of neitiznot (charged).png Azaz129 Crystal shield.png Talk Edits Contribs

23:48, 6 August 2009 (UTC)

A necessity no, but very useful and will help reduce our work and vandalism. Depending on how it is used it can help educate users. One reason for it is to stop those vandal bots without giving them what they may want: to stop the wiki from being edited, which we do with protectsite when they attack. - TehKittyCatTalk Wikian-Book 01:57, 7 August 2009 (UTC)

Support Abuse (Edit) Filter - I have read over what that filter can do. Some of it is pretty intense (de-grouping) but I do have enough faith in teh community that we would atleast set it to "log only mode" for a long enough time to see if it would work or cause more problems than it solves. I would think that giving a 1 or 2 month test period we would see how many actions taken by the filter would have been counterproductive, and asses teh severity of any mistakes. If the filter cannot be tweked to prevent the issue, and the issue(s) are sever, then obviously we won't be able to use it. I don't feel that fear of what the system can do should prevent a log test (as in no action taken) to see what teh system actualy will do. There is no reason that it should ever be set to ban for blanking. Disalowing should be enough, and if thats all the thing will ever do no matter where applied, that could still be an improvement. I say give teh system its test run, program teh filters and improve them as far as we can. IF teh end result does what we wan't without collateral damage (or what we deem unacceptable collateral damage), great. If it causes problems, then we scrap it and no harm done because the system was in log only mode. I see a no-lose situation here with a testing, so why not try it. We can all come back here after the test, see teh results and descide then. RandoxTalk 20:17, 8 August 2009 (UTC)


Apparently, AbuseFilter cannot be installed at the moment. This was the comment made by Angela when I contacted Wikia:

Until Wikia moves to a later version of MediaWiki, the AbuseFilter will not work too well as we need to use an older version. It is currently in testing on a couple of wikis and may be available more widely later on.

For protection of an entire namespace though, there are other options. A recent feature we added was that anonymous users must complete a captcha when they create a mainspace talk page. Do you want something similar to be applied to the exchange namespace?

Kind regards,
Angela

Captcha doesn't sound like a good idea to me since it slows down the process of updating Exchange pages. What do you think?   az talk   07:51, 14 August 2009 (UTC)

Wouldn't that only apply to IPs and newer users less than 4 days old?--Degenret01 07:58, 14 August 2009 (UTC)
Yes. AbuseFilter was meant to warn users if a particular filter was triggered. CAPTCHA on the other hand will only apply to edits made by IPs and new users, regardless of whether the user was updating correctly or not. The only advantage of using CAPTCHA (that I can think of) is that it prevents vandalbots from, erm.., vandalising.   az talk   08:32, 14 August 2009 (UTC)