Talk:Global locks/Archives/2010

Points for further discussion

I'm pointing the questions on the draft here for discussion; we can move the conclusions of that discussion to the policy page when we're ready.  – Mike.lifeguard | @en.wb 16:06, 23 May 2008 (UTC)

Widespread vandalism

  • Point for discussion: how widespread? Should we define that?
Two basic issues (to me) here.
  1. Info derived from CU data where IPs have been exploited to create vandal accounts on a number of wikis
  2. Simply "badly behaved" IPs. I would often find vandalism from IPs on en wb & discovered that they had long blocks on en wp & had "migrated" elsewhere
Maybe simplistic but a starter for 10. The former probably should be considered for long blocks, the latter I think would more be a case of showing them we can do it? Certainly in both cases I would suggest that if three wikis were affected it would be worth examining in detail? --Herby talk thyme 07:30, 24 May 2008 (UTC)
10 is a fine number to me, but one should also take into account the history - if there is some strong indication they will be affecting many wiki shortly, there is no need to wait until they hit #10 before taking action.  — Mike.lifeguard | @en.wb 19:24, 21 August 2008 (UTC)

Widespread spamming

  • Point for discussion: How do we determine the subject of 'clear disregard'? Must we warn them?
    • Static IPs who change domain 'on the fly'.
    • Static IPs who have a wide range of domains.
    • Static IPs who consistently push links in inappropriate ways which are widely used across many wikis, and which do have legitimate use.
    • SpamBots can be blocked while waiting for the blacklisting to kick in.
I dont think they should be warned, because such spambots would certainly not stop the vandalism if they get a warning. --MF-W 17:08, 23 May 2008 (UTC)
Some sort of communication is necessary unless the editor's actions are clearly malicious. The IP spammer referred to here was clearly acting maliciously. Furthermore, we know that there was communication because as soon as each domain was blacklisted and it took effect, they switched to a new domain and continued. Given the number of domains for which this occurred, there is no doubt that this person/bot knew their actions were disruptive, and that a swift global block was needed to prevent further disruption. In such cases where leaving a message on a talk page somewhere is totally ineffective, I agree that doing so is not needed. But where it would potentially be effective, it should be done, and communicating through blacklisting domains is one method of (very clear) communication.  – Mike.lifeguard | @en.wb 19:52, 23 May 2008 (UTC)
warning spambots won't really help but firstly you must try to find out if its a spambot or a very funny human :p Spambots are always looking for trouble, and they hardly ever desist and are always persistent in self-promoting themselves/links. and since spam blacklist have a habit of taking minutes to 'kick in', it really is a good idea to block these ips because when and if it kicks in, the bot controllers might realise it and change the links to avoid being blocked, and its safer to know that the Ip in questions won't be used for spamming again, than only adding the spamlink to the blacklist and allowing the bot to go on a spree with some other url spamlink..--Cometstyles 21:23, 23 May 2008 (UTC)
Again I'm with Cometstyles - warning is not really an option across wikis with quite fast moving bot style edits. To me the warning arrives by way of the block. Depending on the activity the block might not need to be long (open proxies, static IPs may be different). To me this tool is to prevent disruption not punish per se. --Herby talk thyme 12:11, 24 May 2008 (UTC)
Just to be clear, that is what I mean. Blacklisting or a block is as much warning as some deserve - requiring more than that is not a good idea.  – Mike.lifeguard | @en.wb 20:58, 24 May 2008 (UTC)
I am all for warning, but we don't have such a tool (yet?). Global messaging may do the trick, but warning an IP at the moment is impossible, you never know which of the 722 (?) mediawiki wikis (if not other wikis) they hit next. If the IP keeps adding external links, a short block while blacklisting kicks in is at the moment the only possibility.
Do we need to discuss short (say, which can be measured in maximum several hours) and long-term blocks differently here? For blacklisting external links an hour block should be enough (administrators can add the links spammed, investigate if there are other domains likely to be spammed (e.g. domains on same server, see the -pics.com spammer of a couple of days ago) and add those too, and let the block expire). But for long-time vandalism or spam-bots that may not be enough. --Dirk Beetstra T C (en: U, T) 17:37, 25 May 2008 (UTC)

Blatantly disrupting multiple projects

  • Point for discussion: How do we define this to effectively manage disruption but ensure that each project is still sovereign?
I'm not sure I understand what cases this section is supposed to cover - spamming and vandalism are already covered above, and open proxy abuse is covered below.  — Mike.lifeguard | @en.wb 19:32, 21 August 2008 (UTC)
Personal attacks? Disrupting the community in more projects? This would not be quite new... -jkb- (cs.source) 19:58, 21 August 2008 (UTC)
I think that can be controlled adequately on a project-by-project basis. Perhaps there have been examples where a global block would have been necessary though...?  — Mike.lifeguard | @en.wb 22:45, 21 August 2008 (UTC)
I don't think making global blocks for personal attacks is a good idea. That is something the local community should deal with IMO. —Giggy 00:02, 22 August 2008 (UTC)
  • Following up: global blocks have recently been suggested for thekohser and tyciol. Both have been locked for part of this year (due to the lack of a technical option to carry out a global block). So there are cases where this is requested; we should reconsider whether it is appropriate. See below for a specific discussion about reasons for requesting a global lock. SJ · talk | translate 04:04, 11 September 2010 (UTC)

Open proxies

  • Point for discussion: How do we deal with Chinese editors trying to edit through proxies from inside the Great Firewall? I'm not really happy with this line just yet
    • Also whitelist Chinese proxies on chinese wiki(s)?? System for account creation, as these blocks are supposed to block only the IP-editing?
Maybe on chinese wikis, the blocked IP adresses should still be able to create accounts, since the crosswiki vandalbots usually dont create such; but this could cause that they create accounts what would disrupt the global blocking idea. --MF-W 17:08, 23 May 2008 (UTC)
It not really a good idea to Block any proxies unless it disrupts wikimedia and if the OP is being used for vandalism by humans, then it will be blocked for a short period of time but if its done by bots, which can usually be differentiated easily, then longer bans may be in order...--Cometstyles 21:11, 23 May 2008 (UTC)
Agree completely with Cometstyles. I ceased to be comfortable with OP blocks just because that is what they are a long time ago. Quite a lot of constructive work comes from OP across the project. Truly exploited ones maybe but that is covered by the "vandalism" one above. --Herby talk thyme 07:32, 24 May 2008 (UTC)
Perhaps have all global proxy blocks as softblocks (anon only) and put up a well publicised list of emails people can send to to request account creation over that block? giggy (:O) 01:59, 27 May 2008 (UTC)

It's not difficult for an editor in one country to use a proxy in another country. How do we know if it is really a Chinese editor using a Chinese proxy?--Cato 23:31, 21 June 2008 (UTC)

Blocking open proxies because they are open proxies is really not an acceptable use of this tool, I think. That said, there will certainly be open proxies blocked because they are being abused.  – Mike.lifeguard | @en.wb 01:11, 22 June 2008 (UTC)

Global blocks on mobile IPs

  • Discussion on adding a new rule to IP blocks: "Global blocks shall not be placed on reusable IP addresses in the range(s) assigned to mobile ISPs. Instead, an abuse complaint may be filed with the provider."

Policy does not seem to match practice

Why is this policy being used to block named accounts? It dosen't appear that blocking named accounts is in the jurisdiction of this global blocking policy as written. Hipocrite 14:29, 23 March 2010 (UTC)

Policy isn't beeing used to block named accounts, since it is technically impossible with the global blocking. Accounts can be locked, or blocked locally on each wiki the same way local administrators block users. Laaknor 15:06, 23 March 2010 (UTC)
While this is technically true, locking is being used as a substitute for blocking until the technical ability for global blocks is carried out. We really, really need to fix the technical issue to resolve this confusion. SJ · talk | translate 04:06, 11 September 2010 (UTC)

Automatic global blocking

How about to block a (SUL) user globally if he/she was blocked in home wiki? -- ChongDae 02:26, 19 August 2010 (UTC)

For registered users there is locking: see SH#lock, where you'll discover why it would be impossible and undesirable. --Nemo 07:19, 19 August 2010 (UTC)
Return to "Global locks/Archives/2010" page.