Talk:Spam blacklist/Archives/2014-09

URL shortener. MER-C (talk) 08:09, 21 September 2014 (UTC)

Added Added -- — billinghurst sDrewth 10:30, 21 September 2014 (UTC)

URL shortener. MER-C (talk) 11:10, 30 September 2014 (UTC)

Added Added --Glaisher (talk) 11:12, 30 September 2014 (UTC)

Aerobatic Teams website is a place for every airshow fan and specially for aerobatic teams from the World. I don't understand why this site is blocked for linking, as the site is free for use and had many interesting and unique articles. —The preceding unsigned comment was added by 87.121.213.195 (talk) 16:54, 16 August 2014‎

Removed Removed After 6 years, I believe that we can again attempt to have it available, after reading the previous discussions about the abuse of the url at that time. — billinghurst sDrewth 09:30, 6 September 2014 (UTC)

Please remove g.co. It is used for internal google links, and it is used on thousands of articles on English Wikipedia. There is no acceptable reason to block such a heavily used external link without wide discussion. - Floydian (talk) 14:50, 8 September 2014 (UTC)

Removed Removed as this link is heavily used on hundreds of pages xwiki and will prevent many valid edits. Moreover, it's only a shortener for Google's websites. --Glaisher (talk) 14:59, 8 September 2014 (UTC)

Thank you :) Now I just have to revert a few thousand edits by cyberbot![1] - Floydian (talk) 15:10, 8 September 2014 (UTC)

I commented earlier on this link - although it is specific for google services, it is still a redirect (meaning that you don't always know where you go, and also that one could use that inappropriately). Although I agree that this is one of the less sensitive redirect services, I'd still advice that Wikipedia should get rid of them (I would consider an expansion bot on them). Then re-consider; on en.wikipedia, en:WP:ELNEVER prohibits their use - they cannot be (or are not) used to shorten the google.com/url?-'search-engine-internal-statistics'-redirect, right? --Dirk Beetstra ^{T C} (en: U, T) 09:24, 9 September 2014 (UTC)

I am busy slowly restarting COIBot and LiWa3 again - both will operate from fresh tables (LiWa3 started yesterday, 29/12/2013; COIBot started today, 30/12/2013). As I am revamping some of the tables, and they need to be regenerated (e.g. the user auto-whitelist-tables need to be filled, blacklist-data for all the monitored wikis), expect data to be off, and some functionality may not be operational yet. LiWa3 starts from an empty table, which also means that autodetection based on statistics will be skewed. I am unfortunately not able to resurrect the old data, that will need to be done by hand). Hopefully things will be normal again in a couple of days. --Dirk Beetstra ^{T C} (en: U, T) 17:26, 30 December 2013 (UTC)

Due to issues with determining the content of parsed pages ahead of time (see bugzilla:15582 for some examples), the way the spam blacklist works should probably be changed. Per bugzilla:16326, I plan to submit a patch for the spam blacklist extension that causes it to either delink or remove blacklisted links upon parsing, or replace them with a link to a special page explaining the blacklisting. This could be done either in addition to or instead of the current functionality. Are there any comments or suggestions on such a new implementation? Jackmcbarn (talk) 20:45, 3 March 2014 (UTC)

Hi!

I suggest, not to replace the current functionality, and will give an example for this:

In local wikis like w:de, we sometimes have the situation that we want to prevent people from using certain a domain like "seth-enterprises.example.org" everywhere in article namespace with exception of just one article (the one about the institution, e.g. "seth enterprises"). So in this case we remove all links to that domain from w:de, but we place a link to the domain in one article. Afterwards we blacklist the domain, such that nobody can add the link somewhere. In the certain article the link should still work.

Could we cope with this scenario, if the SBL functionality was changed? -- seth (talk) 15:25, 15 June 2014 (UTC)

@Jackmcbarn: I think that would break legitimate links on a wiki (sometimes a site is used minimally in a good way, e.g. in references, but massively spammed and abused further. It then gets blacklisted.

@Lustiger Seth: such links are better of specifically whitelisted. On en.wikipedia, we would whitelist the landing page ('seth-enterprises.example.org/index.htm') or the about-page (often the index.htm is 'invisible', forcing us to, in principle, whitelist the domain only, and that would open up the abuse possibility again if the problem was the linking of the domain only). In rare cases, we would whitelist the domain only. De-blacklisting, linking, and re-blacklisting is not a real solution - there are edit-scenarios where the only solution for repair is to de-blacklist again, repair, and re-blacklist. For an uninterupted edit-experience, it is better that for all blacklisted links a whitelisting solution is found. --Dirk Beetstra ^{T C} (en: U, T) 03:28, 19 June 2014 (UTC)

Hi!

White listing does not help in many of the mentioned cases, because the url of the spammers can be the same as the url that is needed in an article. If there is a better soulution, plese tell me. The edit filter could of course be used for a combination of a link-block with a specific article exception. But we try to not use the edit filter for performance reasons (if we would not do this, the edit filter would not work properly). -- seth (talk) 09:54, 19 June 2014 (UTC)

whitelisting of the type of 'http://seth-enterprises.example.org/index.htm' has on en.wiki never resulted in problems, and 'http://seth-enterprises.example.org/about.htm' neither. In fact, heavily abused websites have their index.htm's and/or about.htm's whitelisted, and are still not abused. --Dirk Beetstra ^{T C} (en: U, T) 10:51, 19 June 2014 (UTC)

We would of course not whitelist 'http://seth-enterprises.example.org' - that would open up everything, and have an end-of-string delimiter also does not help, as the main-domain is generally what is abused. --Dirk Beetstra ^{T C} (en: U, T) 11:14, 19 June 2014 (UTC)

• Talk archives