Jump to content

Wikipedia talk:External links/Archive 35

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 30Archive 33Archive 34Archive 35Archive 36Archive 37Archive 40

The policy is a bit vague about how external links should be handled in stand alone lists. The example I came across is List of real estate companies of the Philippines. It includes links to the website of each company listed (as well as phone numbers). These external links come across as inappropriate to me, but I'm unsure. Any thoughts? Deli nk (talk) 13:22, 4 May 2015 (UTC)

Seems promotional and inappropriate to me. IMO that list should be limited to companies that are bluelinked and/or reliably-sourced in a manner that establishes their significance. DonIago (talk) 15:10, 4 May 2015 (UTC)
Then we need to make the guideline clearer, since such links are inappropriate. --Ronz (talk) 16:24, 4 May 2015 (UTC)
Currently, the policy says, "Stand-alone lists or embedded lists should not be composed mainly of external links. These lists are primarily intended as internal navigational aids, not a directory of sites on the web." The word "mainly" suggests that including weblinks would be acceptable as part of a fuller listing. How would you suggest making the guideline clearer? Perhaps bringing in the concept of Wikipedia is not a business directory? Deli nk (talk) 16:45, 4 May 2015 (UTC)
I went through with a machete. I didnt actually check the philstar sources - they may just be adverts or directory listings and more should be removed. -- TRPoD aka The Red Pen of Doom 17:44, 4 May 2015 (UTC)
Let's get rid of the word "mainly". Any rare exceptions would be covered by IAR.- MrX 20:37, 4 May 2015 (UTC)
@MrX: That would give "...lists should not be composed of external links" which might be confusing as worded. Is "...lists should not contain external links" consistent with your suggestion? Deli nk (talk) 20:52, 4 May 2015 (UTC)
Yes, Deli nk, that would be better.- MrX 21:03, 4 May 2015 (UTC)
I agree that the word mainly should go. --Dirk Beetstra T C 03:25, 6 May 2015 (UTC)
Since there seems to be agreement, I have made a slight change the wording. Deli nk (talk) 18:11, 11 May 2015 (UTC)

"Lists should not list external links" is going to be a problem. First, MrX's assertion aside, policing spam attracts editors who follow absolute rules absolutely. Secondly, we have a persistent problem with people interpreting "citation" as meaning "something enclosed inside ref tags". For some stand-alone lists, the 'external link' is a citation, but it's formatted exactly like an external link. We want those, both because they're appropriate at ELs in some cases (e.g., an official link to an organization that's being mentioned) and because they do verify the material (e.g., URLs to radio stations, in a list of radio stations).

What's meant by "should not be composed mainly of external links" is do not have a list whose contents are:

nor even a list whose entire contents are:

but definitely do accept a list whose contents are:

  • Radio 1 – 91.1 FM. This is the first radio station to broadcast in this area.
  • Radio 2 – 94.5 FM. The second radio station founded; closed in 1998.
  • Radio 3 – 103.3 FM. The fifth radio station for this area; the name comes from the frequency number.

or (especially if each of these could be notable radio stations)

  • Radio 1 – 91.1 FM. This is the first radio station to broadcast in this area.[1]
  • Radio 2 – 94.5 FM. The second radio station founded; closed in 1998.[2]
  • Radio 3 – 103.3 FM. The fifth radio station for this area; the name comes from the frequency number.[3]

And it is especially important to not end up with this:

List (redundant one)

  • Radio 1 – 91.1 FM. This is the first radio station to broadcast in this area.
  • Radio 2 – 94.5 FM. The second radio station founded; closed in 1998.
  • Radio 3 – 103.3 FM. The fifth radio station for this area; the name comes from the frequency number.

(or any equivalent, such as wrapping those URLs in ref tags and calling them ==References== instead).

This is why "mainly" is important language: we don't want a list that is entirely URLs, but we don't want to prohibit sensible formatting if we have a list that contains significant other information. WhatamIdoing (talk) 14:55, 3 June 2015 (UTC)

Unsurprisingly, I disagree. Citations should always be wrapped in <ref></ref> tags because that's how the software is designed to work. I agree that if an editor added bare external links as a good faith attempt to provide citations, then those links should not be removed; they should be converted to citations by wrapping them in ref tags. The question is, what is the bigger problem: legitimate references being removed by overzealous editors, or spam links masquerading as encyclopedic content? A recent example is List of TED speakers, which had citations for the list as a whole and hundreds of links to the individual presentations.- MrX 15:31, 3 June 2015 (UTC)
Agreed. Links that are not references don't belong. References should be formatted as such. Links to examples, official sites, etc are linkspam and don't belong. --Ronz (talk) 15:49, 3 June 2015 (UTC)

MrX and User:Ronz, do you mean, citations should be formatted the way that WP:Citing sources recommends, which just happens to include what's shown here? Ref tags are popular but strictly optional. There are half a dozen styles of WP:Inline citation. Some people would say that WP:ECITE, which is what's used here, is deprecated (and I personally dislike it, because it's usually misused), but it's not actually prohibited.

But what I'd like to know is why you think this would be preferable:

It's preferable because it allows readers to easily distinguish between sources and external links that provided further information about the subject. It's also a widely accepted convention across Wikipedia. Citations with URLs can easily be followed by enabling Reference Tooltips in user preferences.- MrX 15:08, 4 June 2015 (UTC)
But these particular links, in practice, are both "sources" (that verify the fact that this item belongs in this list) and "external links" (that provide further information about the list entry), so your answer amounts to "I prefer having separate, redundant lists because it misleads people about the nature of the link". Is that actually such a good idea? WhatamIdoing (talk) 16:06, 4 June 2015 (UTC)

List (redundant)

  • Radio 1 – 91.1 FM. This is the first radio station to broadcast in this area.[1]
  • Radio 2 – 94.5 FM. The second radio station founded; closed in 1998.[2]
  • Radio 3 – 103.3 FM. The fifth radio station for this area; the name comes from the frequency number.[3]

References

to this:

List (compact with sources)

  • Radio 1 – 91.1 FM. This is the first radio station to broadcast in this area.[4]
  • Radio 2 – 94.5 FM. The second radio station founded; closed in 1998.[5]
  • Radio 3 – 103.3 FM. The fifth radio station for this area; the name comes from the frequency number.[6]

or even this:

Station name Frequency Description Source
Radio 1 91.1 FM. This is the first radio station to broadcast in this area. [7]
Radio 2 94.5 FM. The second radio station founded; closed in 1998. [8]
Radio 3 103.3 FM. The fifth radio station for this area; the name comes from the frequency number. [9]

While the redundant system uses a popular formatting style for citations (and thereby shoos away editors who can't tell the difference between a inline citation to a primary source and a list of external links), the effect is that the entire list gets repeated on the page. Imagine a list with 100 entries. Is that page actually improved by having 100 list entries followed by repeating the entire list of 100 items, only this time with a URL added? Why not just put it all together, on one line? WhatamIdoing (talk) 04:00, 4 June 2015 (UTC)

More

The examples would benefit with numbering. The only one that I think is acceptable is the "or (especially if each of these could be notable radio stations)". --Ronz (talk) 16:36, 3 June 2015 (UTC)
I also, unsurprisingly, disagree with the assertion that external links should be allowed with list entries. If the list entries meet notability standards, they can have their own Wikipedia article with their own Official External Link on that article. There is no reason for any other list than this:
List
  • Radio 1 – 91.1 FM. This is the first radio station to broadcast in this area.
  • Radio 2 – 94.5 FM. The second radio station founded; closed in 1998.
  • Radio 3 – 103.3 FM. The fifth radio station for this area; the name comes from the frequency number.
And, as those all should be pointing to existing Wikipedia pages with the Official websites linked there, there shouldn't be an external link in the EL section at the bottom of the article for that entry. And, any refs should be pointing not to the official site, but to a reliable, third-party published source that shows justification for inclusion on the list. Stesmo (talk) 18:29, 3 June 2015 (UTC)
Stesmo, do you actually mean that an ==External links== section should be prohibited in stand-alone lists? WhatamIdoing (talk) 04:02, 4 June 2015 (UTC)
@WhatamIdoing I think having an EL section with a minimal number of links that meet WP:EL is good for a few stand-alone lists (the TED Talks list mentioned above, for example). I'm against banning an EL section from all stand-alone lists. However, the links should be furthering the encyclopedia knowledge of the subject of the article. And, as the subject of the article is "List of Radio Stations" (used primarily for navigation to Wikipedia articles about particular Radio Stations), external links to individual radio stations wouldn't fit the bill on the List article. Especially as the list entries are probably notable and would have wikilinks to their very own articles with an External Link there. So, ban EL sections on stand-alone Lists? No. Should most stand-alone lists have an EL section? Probably not. Stesmo (talk) 16:52, 4 June 2015 (UTC)

I also think that the "stand-alone lists or embedded lists should not list external links" wording is problematic. It can easily be interpreted as saying that stand-alone lists shouldn't have external links sections, which isn't true. There are numerous examples of stand-alone lists with desirable links in an external links section, and we shouldn't discourage that practice. - Eureka Lott 17:43, 3 June 2015 (UTC)

Yes, I agree that this wording will be interpreted as banning the entire ==External links== section from all stand-alone lists. I don't know if any of its previous supporters actually intended that, but it will be read that way. WhatamIdoing (talk) 04:02, 4 June 2015 (UTC)
I don't think that lists by definition should never have an ==External links== section. However, I do think that:
Station name Frequency Description Source
Radio 1 91.1 FM. This is the first radio station to broadcast in this area. [10]
Radio 2 94.5 FM. The second radio station founded; closed in 1998. [11]
Radio 3 103.3 FM. The fifth radio station for this area; the name comes from the frequency number. [12]
as a list item is redundant in having the external link (in any format). That link is supposed to be in the article Radio 1 (as the official site per WP:ELOFFICIAL), and hence is superfluous on the list. It is not needed as a primary reference either, we have the Wikiarticle (which should be having secondary sources showing notability) and hence 'it exists' - there is no need for the primary 'reference'. If Radio 32143298457 does however not have an own entry in Wikipedia, then it can only be taken up in a list if it is somewhat notable, and hence it needs a secondary reference (note: its notability does not have to rise to the level of suitability for an own article, but not be of the level of 'I also run a radiostation from my garden shed and my neighbour is enjoying it every day!'). Note that certain columns in a table could be primary sourced:
Station name Frequency Description Streaming
Radio 1 91.1 FM. This is the first radio station to broadcast in this area. Yes
Radio 2 94.5 FM. The second radio station founded; closed in 1998. Yes[1]
Radio 3 103.3 FM. The fifth radio station for this area; the name comes from the frequency number. Yes
-- refs --
  1. ^ http://radio2.com/streaming
  2. It should be either as for Radio 1, or (likely preferred) Radio 2, and not like Radio 3 (that one should be converted to be in line with Radio 2).
    I still think that external links in lists are not supposed to be there - either the items are to be 'primary referenced' to their own Wiki-article (hence showing notability because the article exists), or need independent proper references showing existence ánd notability (and if that can not be proven, the item is just failing our inclusion standards, same goes for redlinked items). Certain columns in the article can be primary referenced, but not the whole row. The external links for the individual items should also not be in (duplicated in) the external links section. --Dirk Beetstra T C 13:29, 4 June 2015 (UTC)
    Here are the problems:
    1. List entries are not required to contain only notable items, or even "somewhat notable" ones (see WP:LSC). Therefore, there might not be any page that contains other citations.
    2. Even if such a page exists, relying on the citations at the linked page is not acceptable (see many archives at WT:V).
    I agree that the external links for the individual items should also not be in duplicated in a section that says ==External links== at the top of it. I still don't understand why it would be such a good idea to duplicate the external links for the individual item in a section that says ==References== at the top of it. Either we need a completely separate and duplicate list to store the URLs for every item, or we don't, and IMO the label on the list doesn't determine whether we need to list every item twice.
    Also, given your statement, "I don't think that lists by definition should never have an ==External links== section", then you should change the current wording. "Stand-alone lists or embedded lists should not list external links" sounds a lot like "Lists should never have an external links section". WhatamIdoing (talk) 16:20, 4 June 2015 (UTC)
    @WhatamIdoing: Items in the list can be primary sourced to their existence, but that items that rely on only that can be questioned to our inclusion standards per WP:IINFO. I see by the way that WP:LSC is under discussion for something that is really close to this point - what should be in a list.
    I don't think that that says that, I think that it reads that the list itself should not have external links listed in the list - but one could add a clarification sentence that list-articles can include an external links section with links that satisfy our WP:EL guideline, noting that external links to the individual items in the list are indirect to the subject of the page, and hence fail WP:ELNO #13 (the external links talks about 'an item in the list', not about 'the list'). --Dirk Beetstra T C 07:42, 7 June 2015 (UTC)

    Regarding ==External links== sections in lists: They are rarely appropriate, but can be when the entries are external links to webpages that contain useful (not redundant and highly relevant) lists that are not or cannot be used as references within the list article itself. --Ronz (talk) 17:08, 4 June 2015 (UTC)

    I guess that the ==External links== section could contain links to dmoz of similar lists, or when the list in itself has something that can be talked about. But I agree that most of the true list pages will often not need an External links section. --Dirk Beetstra T C 07:42, 7 June 2015 (UTC)

    Since nobody supports a ban on ==External links== sections, and since the sentence could be read that way, I've clarified. I've also added a section heading for convenient linking and an example, in case that might help editors understand what we really don't want. Is everyone satisfied? If not, it's a wiki, and you should feel free to try out a new way to phrase it. WhatamIdoing (talk) 23:36, 7 June 2015 (UTC)

    .onion linking and clearnet gateways

    Hi all

    So I've been writing up a number of tor hidden services so the list now has an increased amount of live services than dead ones. Now I understand the inherent controversy in linking to darknet markets, but I also understand the issue where 'random' often unverifiable links get added - is that the real site or not? You don't know!

    There is a major exception to this issue, onion sites with clearnet gateway sites or full joint web support. (related, one day Clearnet gateway may be a thing) Some examples:

    • DeepDotWeb, features an official onion link with a citation from the clearnet site - I added a reference directly to the page in this case
    • Facebook, has an official onion address but it has low prominence - suggest the above convention?
    • Bitcoin Fog, their gateway exists purely to link to their onion site - so probably work labelling the link as a 'clearnet gateway', without linking directly to the end point
    • Agora (online marketplace), I just added the official gateway to it no it's unofficial, this is exactly the problem!, but I would understand if this would be undesirable due to the site's content

    Onion sites and alternate end points are here to stay, I propose the following conventions for onion sites:

    1. Sites that support Tor access must have a clearnet reference on their external link to verify its authenticity to any Wikipedian. Direct onion linking should go though a small approval process, so allow direct linking via exception only. Bots could automatically reject links without citations out of hand, prior to a human having to approve it.
    2. Official clearnet gateway sites should go through a verification process (not sure how exactly just yet), but should be used instead of linking to the onion site. The link should be labelled a clearnet gateway site, like you would a mobile site.

    Thoughts? Deku-shrub (talk) 18:45, 7 June 2015 (UTC)

    WP:ELNO#EL3 comes immediately to mind for at least some of the websites. There is no requirement to provide an external link at all, even if the subject of the article is web based. In the case of Agora, I would tend to be opposed to linking to its official gateway as it seems to offer no encyclopedic knowledge about the subject.
    Generally though, something should be added to the guideline to address this type of content and your two bullet points above seem to be a good starting point for discussion. - MrX 19:48, 7 June 2015 (UTC)
    Whilst I don't want to distract from the issue for a generalised onion-linking policy, whilst a site containing illegal media (criminal or copyright infringing) should not be linked to under WP:ELNO#EL3, AFAIK it is not illegal to access a darknet market. Otherwise journalists like Andy Greenberg and the creators of Deep Web (film) would be criminals, which they're not. That said, if the onion-linking policy becomes more liberal, relevant safeguards might need to be put in place ahead of any policy change Deku-shrub (talk) 20:31, 7 June 2015 (UTC)
    To clarify, I was thinking more along the lines of child pornography when I referred to EL3.- MrX 20:53, 7 June 2015 (UTC)
    Fair point. First of all there are no active notable such sites on wikipedia right now, so doesn't apply. Should this happen, it would be covered under that point you raised already though. Deku-shrub (talk) 21:25, 7 June 2015 (UTC)

    Deku-shrub, for the links you're talking about, what will happen if I click on them (without installing any special software, configuring anything, etc.)? One of the main reasons for a near-total ban on .onion sites has been WP:ELNO#EL7, "Sites that are inaccessible to a substantial number of users, such as sites that work only with a specific browser or in a specific country." WhatamIdoing (talk) 23:33, 7 June 2015 (UTC)

    @Deku-shrub: the .onion is an official link of facebook, though not thé official link of facebook. We do not include all (official) links that are connected to a subject, only what most people see as the most important one (see WP:ELOFFICIAL. For those cases where the .onion itself is thé official site of the subject of the page, it can be included, after a request for whitelisting (which will be granted if you can show that it is the official site). Do note, that WP:ELOFFICIAL is a guideline-ified form of WP:IAR on WP:NOT (many official sites of subjects fail our inclusion standards), and we could again WP:IAR on WP:ELOFFICIAL if a site is the official site but is too problematic for use - there is no must in linking, even to the official site (and many official sites fail several of the WP:ELNO-rules). --Dirk Beetstra T C 06:16, 8 June 2015 (UTC)
    Is there an existing request for whitelisting process I can follow? Do you have info? Deku-shrub (talk) 14:54, 8 June 2015 (UTC)
    WT:WHITELIST.- MrX 15:28, 8 June 2015 (UTC)
    I'll see what the response is Deku-shrub (talk) 16:10, 8 June 2015 (UTC)
    Nope, all denied. The issue appears to be that alternate links should be rejected in favour of more prominent links. However the onion only sites often have questionable content... Deku-shrub (talk) 14:18, 10 June 2015 (UTC)
    @Deku-shrub: - it is not what it appears to be, that is what we have codified in WP:NOT and WP:ELOFFICIAL, it is what our pillars/policies and guidelines describe. And also, this is not about the content that is being linked to, it is about the problems Wikipedia in the past had with .onion sites (abuse, in this case in the form of insertion of false .onion addresses over the actual official site of the subject). If a .onion is thé official site for a subject, it merits linking to, even if the content on the external site is (in the eyes of some) questionable. --Dirk Beetstra T C 10:25, 11 June 2015 (UTC)
    The problem is, there is no iron-clad way for an onion-only site to advertise it's url, and as a result, it's hard to get that assurance that it's genuine. Only sites with dual clearnet and tor sites can officially advertise their onion urls, these are the ones suggest to be redundant.
    Frankly, this is a problem with the fragmenting nature of name systems in general, I believe Namecoin want to be the successor to .onion domains to manage this better, then one could cite some official namecoin registry viewer to authenticate sites. As mentioned, clearnet gateways attempt to solve this issue, but the have their own problems :/ Deku-shrub (talk) 12:01, 11 June 2015 (UTC)
    @Deku-shrub: - that is close to the reason how .onion ended up on the blacklist, and we will only whitelist them with significant scrutiny and when really needed. The ones you requested whitelisting for are for sure not 'needed' as encyclopedic content. --Dirk Beetstra T C 13:51, 11 June 2015 (UTC)

    I'm trying to work out the best / safest way to ever allow onion linking. I thought it'd be a stronger test case to allow the likes of Facebook, rather than sites like Agora (online marketplace) where I have to cite a reference from DeepDotWeb for the onion link, which suffer from 3rd party reference dependencies and questionable content. I'm looking to pave the way for general, safe onion use on Wikipedia. Deku-shrub (talk) 21:02, 11 June 2015 (UTC)

    The fundamental problem is that for the regular web there's an infrastructure in place to ensure that when someone types "google.com" into their browser, it actually goes to google's servers, and if it doesn't, it can be easily fixed. It's because the regular web isn't anonymous. Google.com is registered to a real company with a real contact address and it's registered with a real registrar, likewise with a real address/contact. If a domain is hijacked or website otherwise compromised, the mechanisms in place make it so it doesn't typically take very long to sort it out except in those cases when the site's content or trade is illegal such that the legitimate owners have an interest in maintaining anonymity (and even then, sometimes). So when there's a reliable non-Tor counterpart to a site which advertises the "official" .onion, that's one thing, but in all other cases, everybody is anonymous regardless of whether their dealings are legal or illegal, and in an anonymous network, there's just not enough accountability to say any .onion link is the official link (again, barring those cases where there's the non-.onion counterpart. — Rhododendrites talk \\ 03:20, 6 July 2015 (UTC)
    Yep, I understand there is a fundamental problem with onion site authentication - who do you trust to vouch for it? But by putting in place a robust Wikipedia policy, Wikipedia itself could aid in the vouching process through reliable sourcing of 1 official, or multiple unofficial links. I think this is a desirable endeavour. Deku-shrub (talk) 14:15, 6 July 2015 (UTC)
    Personally, I lean toward not trusting any source sufficiently to include a .onion link on Wikipedia. Requiring the address to be accompanied by a reliable source has been a good compromise, but because of the tenuous relationship between .onion address and a particular site/owner, it's easy for any source to get it wrong, and harder to determine if something has changed (if that address now points to a phishing site, for example, would we require an updated reliable source in order to remove it?). The sources that write about these sites, published at a particular point in time, are too easily outdated and the consequences too significant. Having a site like Wikipedia "vouch" would indeed be useful, but Wikipedia should never be the most authoritative source since its principles of WP:V, WP:NOR, and WP:RS preclude including anything here that isn't already covered in existing reliable sources. — Rhododendrites talk \\ 22:36, 6 July 2015 (UTC)
    Since the introduction of vanity urls, e.g. facebookcorewwwi.onion the issue of accidentally linking to the wrong url has been diminishing significantly. I would support restrictions to only allow vanity .onion address for instance. Against proactive fraud it doesn't work, e.g. darknetmarkets.org and fake Agora, but this is the same problem that the existing domain name system has, not a new one. Websites are hacked every day, I'm not sure how much more frequently this is on Tor. Saying 'we won't be able to remove it in case it gets hacked' is a little paranoid in my opinion. Also, hacks get pretty well documented. Like with any site, a link could be temporarily removed, pending further information. Also, who says what is an official site anyhow? The whois information? What if there's no whois? A known social media account, a PR announcement? The chain of trust determining what an official site is has to start somewhere, Wikipedia should capture the key signposts and turn this into a policy that allows the safe linking of onion site IMO Deku-shrub (talk) 23:02, 6 July 2015 (UTC)

    Deku-shrub, I don't see an answer to the question I asked above. Let's pretend that we decide that we want to include .onion links. You add one. I want to click on it. What will happen if I click on it, without installing any special software, configuring anything, etc.? Something? Nothing? An error message? Will a completely plain-vanilla web browser even recognize it as being a link? WhatamIdoing (talk) 00:29, 7 July 2015 (UTC)

    Having the links clickable and having them included without apology are different tasks. Assume I just want uncontested inclusion for now Deku-shrub (talk) 09:17, 7 July 2015 (UTC)
    User:Deku-shrub - barring very, very few exceptions, those links are all failing our inclusion standards - those exceptions can be whitelisted and be a clickable link, for the rest, the answer is simply going to be 'no', because the inclusion is contested (per Wikipedia:External links) for all of them. You focus on the problem of the links being false links, but that is by far not the only problem why they are and should be excluded. --Dirk Beetstra T C 11:00, 7 July 2015 (UTC)

    Please comment on the RfC at The Pirate Bay

    There is an RfC at Talk:The Pirate Bay regarding inclusion of external links with little participation from uninvolved editors. If you would like to contribute to the discussion, please do so! Wugapodes (talk) 21:08, 7 August 2015 (UTC)

    .onion linking proposed standard

    Continued from discussion at #.onion linking and clearnet gateways I would like to propose a standard that has emerged from all the various List of Tor hidden services I have documented.

    1. That if the site has a reliably documented hidden service link, it is to be included. Sites with a hidden service address without a reliable source shall not be included. Such reliable sources of links include:
      1. A link is featured in a notable source, e.g. news article
      2. A link is featured on an official associated website or official clearnet gateway
      3. A link is featured on a reliable 3rd party source of .onion links. The only one I consider reliable at this time is DeepDotWeb
    2. That the link has not subsequently been invalidated by means such as:
      1. The site key has been found to have leaked (e.g. noted on a forum or less reliable source)
      2. The site has moved (and the updated link has not yet been noted in a source from above)
      3. The site has closed (e.g. law enforcement action). In this case the domain(s) may warrant listing in the body rather than external links section.
    3. The standard format of linking shall be the following:
      * <nowiki>http(s)://preferredonionaddress.onion</nowiki><ref>Reference to a reliable source as mentioned above</ref>, [[.onion]] address for [[Tor (network)|Tor]] hidden service

    For the purposes of this I am assuming the Wikipedia-wide ban on direct onion linking shall remain in affect. Deku-shrub (talk) 18:41, 18 August 2015 (UTC)

    @MrX, WhatamIdoing, and Beetstra: just pinging those other than myself and Deku-shrub who participated in the previous thread (hopefully I'll have more time for a real response later) — Rhododendrites talk \\ 21:11, 18 August 2015 (UTC)
    What encyclopedic purpose would be served by advocating use of .onion or listing .onion links? Johnuniq (talk) 01:03, 19 August 2015 (UTC)
    It's an elegant proposal, but why do we need it? Like a similar proposal to include all of The Pirate Bay's URLs as WP:ELOFFICIAL links in its article, this seems to lose sight of the fact that we are an encyclopedia, not a directory of links. What would be lost if we didn't include .onion links in these articles? Maybe a better (meta) solution would be to include a link to the appropriate listing page at deepdotweb.com. - MrX 01:44, 19 August 2015 (UTC)
    I still don't have a plain answer to my question above. What will actually happen if someone clicks on such a link?
    I don't think I'm being clear about this, because I'm not getting useful answers. So let me give you some sample answers, and maybe after a game of multiple choice, I'll actually get an answer.
    1. The user can't click on the link because https://example.оnion will not be recognized as a link. It'll just display in plain text, like "https://example.onion" already does.
    2. The user will click on it, and then absolutely nothing will happen. Maybe a new tab will open (if the browser prefs are set to open links in a new tab), but the link will fail to load because it requires special software.
    3. The user will click on it, and it will work exactly like the user was going to https://example.com
    4. The user will click on it, and a bunch of error messages will appear.
    Now: which one of these is a reasonably accurate description of actual, consensual reality for a person using a plain-vanilla, zero-extensions/plug-ins, simple web browser (i.e., the thing that most people will encounter at an internet café or public library)? Do you even know the answer? WhatamIdoing (talk) 02:08, 19 August 2015 (UTC)

    I get a massive beating-a-dead-horse feeling here. Can we please stop this discussion as it has been answered over and over in the same way, and the same questions have been asked over and over. IF a .onion address is thé (THE, see WP:ELOFFICIAL) official site of the subject of a page, then it should (not must, should) be linked. If that is the case, one can request whitelisting for that site, which will be granted. If it is not, it should not be linked, and whitelisting will not be granted. As for WhatamIdoing's question, if you don't have the software, it will fail to get there, your DNS will not understand. As such, when a link gets whitelisted and linked from a page, then it must (now not should, must) be clear that you need special software to follow that official link. This is all already mentioned in our policies and guidelines, and follows common sense.

    Same goes for references, with the caveat that a) a direct link for a reference is not needed, it is a convenience link. Writing that it was reported in the journal Nature, first issue of 2000, on page 300 allows everybody to verify the data - it is just more work. That goes for .onion as well. b) the information must be totally unique on the .onion site (and if that is so, the mentioning of that information on Wikipedia must be because it is significant). As for the external links, it should be made clear that special software is needed to access the information - more clear than what you propose, per our policies and guidelines. Again, this is again all mentioned in policies and guidelines, and follows common sense.

    In short, we have all procedures in place for linking to .onion, and we do not need to make a special case out of it. Can we now stop beating this dead horse, we have enough bureaucracy here. Please close this thread. --Dirk Beetstra T C 03:34, 19 August 2015 (UTC)

    "IF a .onion address is thé (THE, see WP:ELOFFICIAL) official site of the subject of a page, then it should (not must, should) be linked. If that is the case, one can request whitelisting for that site, which will be granted." I'm not willing to go through the whitelisting procedure for every such link. I disagree that that this would be even be a safe approach. Deku-shrub (talk) 14:35, 19 August 2015 (UTC)
    @Deku-shrub: "for every such link" - that is a handful, we do not have that many notable webpages that exist solely on .onion (or better described: where the .onion is the most prominent official site). Moreover, you don't need to do it alone, there are other editors who will create such pages and have the need for that one (we have no editors who only use examiner.com sources and . And note that the official .onion of facebook is an official site of facebook, not the official site of facebook. We do not include all official links per WP:ELOFFICIAL, only the most prominent one(s). Adding .onion (or myspaces, twitter accounts, etc. etc.) is outside our scope. If you can now show me 20 individual pages on Wikipedia for which the .onion is the most prominent official site, then I am going to be impressed (but still far, far away from the threshold of where I would say that the work of having to whitelist those few outweighs the advantages of control through blacklisting).
    I'm not sure what you think is unsafe on this approach. --Dirk Beetstra T C 03:29, 20 August 2015 (UTC)
    As an addition, the community is not willing to check, nor has the capability to check, for all the hundreds of pages which do have an official .onion next to their official site, whether that is actually a correct one, and to revert changes of that one into spammy/malicious .onion addresses (while we should not have those included in the first place anyway). Nor am I willing to remove all the .onion addresses on pages where the subject does not have a .onion address but where someone sees a chance to insert a .onion to link to a malicious site. Insertion of malicious .onions on pages where the .onion wás the official site was the problem that got the sites blacklisted. There is unfortunately a need for control here, just as for the plethora of legit sites on the blacklist with very limited use because their owners felt the need to abuse Wikipedia to get them linked (I have 2 companies where I keep my eye on which are active already for several years) - that is why we have the blacklist with a whitelist companion. Don't blame the whitelist, don't blame the way that problem is handled on Wikipedia .. blame the people who feel the need to abuse Wikipedia, because those are the ones you need to thank for this. --Dirk Beetstra T C 03:40, 20 August 2015 (UTC)
    "show me 20 individual pages on Wikipedia for which the .onion is the most prominent official site"
    * The Hidden Wiki (though it has its own issues)
    * Bitmessage, Riseup arguable
    * Tor Mail (though the onion is missing right now)
    * Darkode (but is per-user unique)
    * The Hub
    * Russian Anonymous Marketplace, Assassination market, Agora, AlphaBay Market, TheRealDeal (note, these are being expanded on Draft:List of darknet markets
    * Doxbin
    * Bitcoin Fog
    It's not 20, but it's a few and I'm still working on improving coverage in general. Deku-shrub (talk) 11:18, 20 August 2015 (UTC)
    Indeed, 13. And I know that 20 is reasonable. And quite some seem to have a 'regular' official site (I checked 3), next to their .onion (and that regular official site provides much more info to most of the readers than the .onion). Still, 20 articles for which we need to do whitelisting is nothing in comparison to the millions of pages we have. I am looking forward to some whitelisting requests, and to an answer to my question. --Dirk Beetstra T C 03:35, 23 August 2015 (UTC)
    You want to whitelist some and set some precendents prior to creating a policy then? Deku-shrub (talk) 18:09, 23 August 2015 (UTC)
    This is all already perfectly in line with our policies and guidelines. We do not need to create a policy for this. WP:NOT/WP:V/WP:EL/WP:ELOFFICIAL/WP:SPAM, and the handling of the spam blacklist and spam whitelist are all more than enough (we blacklist if they are abused, we whitelist specifics if they pass the bar/are needed). I asked earier to stop beating this dead horse, we are not a bureaucracy, no need for having written down rules for everything. --Dirk Beetstra T C 11:07, 31 August 2015 (UTC)

    Ordering of ELs?

    I glanced over this page but ... is the rule specified anywhere?

    If not, how about adding to WP:ELPOINTS something like the following:

    There are no hard and fast rules governing the ordering of external links. However, if you add an external link to a preexisting list, either consider the best place for it, or place it at the bottom of the list.

    I wonder about this because Ariwara no Narihira and several other articles I recently compiled have ELs "(1) List of X's poems on Y database. (2) Digitized copy of The X Anthology on the same database." and it's possible someone might put something else in between (1) and (2) in the future. In theory that's not a problem, but wouldn't it be better if the guideline explicitly encouraged rewriting (2) as necessary in such a case?

    Hijiri 88 (やや) 09:08, 29 August 2015 (UTC)

    I think it's best to avoid repetition in external links as in speech and would not repeat descriptions and proper names of moderate length. And would avoid the empty "digitized copy of". For the example given, supposing "on" is an appropriate preposition, I would write [1] "List of X's poems on Y database" and either [2] "The X Anthology on Y" or [2a] "The X Anthology on y" where "y" is a short version of the full name "Y". For instance (without any ordinary noun such as "database"), Y = Library of Congress Catalog, y = LC Catalog. I would not be influenced by the possibility that someone might insert one or a dozen other external links between them, nor that someone might re-order the two. --P64 (talk) 17:58, 30 August 2015 (UTC)
    In the case where further links refer to previous links, you've got a sub-list, and so you could indent each of the related links by one level by using an additional asterisk in the list format, so that they're visually subordinate to the first link. {{Nihiltres |talk |edits}} 13:41, 31 August 2015 (UTC)
    We generally try not to write down advice unless there has actually been a problem that wasn't easily resolved by good-faith, common-sense editing. WP:Nobody reads the directions before editing anyway, so writing down your advice would almost certainly be ineffective. If you are worried about those two links, then inserting a friendly <!--hidden HTML comment--> would be more practical. WhatamIdoing (talk) 06:09, 2 September 2015 (UTC)

    Official Facebook links?

    Consider a page whose subject is an organisation, charity or whatever. They meet WP:N. They also have a Facebook page, an official page, run by the group in question. It is used for the publication of content from the group, to an interested audience. Comments take place, but it's primarily a publishing channel, not a forum.

    Should this page be linked? Under ELs, as we would for their website?

    Is there any clear policy statement on this - I can't see one at present. Andy Dingley (talk) 13:07, 3 September 2015 (UTC)

    WP:ELOFFICIAL - we do not list all official pages, only what is deemed the most official one (generally the 'subject.com'-like one). There are cases where the facebook is as an addition of interest (e.g. someone who is a facebook personality, same with someone who is primarily known for their YouTube channel or their twitter activities), but if that is not the case, the facebook is superfluous and not included (note, that generally if the facebook channel is of interest, it is also already prominently linked from their more official website). Then there are cases where the only official online 'presence' is on facebook (though often these tend to be not very notable in themselves).
    Generally, the facebooks/twitters/youtube channels/myspaces/etc. etc. are not passing the bar - we are not writing an internet directory, and sometimes even the official page is already an exception to a lot of WP:ELNO-cases (purely commercial, not informative about the subject in itself, etc.). --Dirk Beetstra T C 13:53, 3 September 2015 (UTC)
    I agree with Dirk Beetstra. If there's an official website we should usually omit links to social networking site per WP:ELNO#social and WP:ELMINOFFICIAL. Of course, there may be rare exceptions to consider. - MrX 13:59, 3 September 2015 (UTC)
    Assuming that it is appropriate to list it, where's our policy on this? Does it need to be made more evident?
    In this case I'm inclined to include it, no doubt others will vehemently remove it. The organisation does seem to be most active through their FB page – a situation that's not unusual these days (my own business FB page is my busiest web communication). As FB encourages rapid communication, it tends to be used for such. I see this as valuable for portraying what the organisation is and what they do. A simple blanket policy of "FB isn't Proper" is just snobbishness: if the group is notable, they deserve coverage of whatever and wherever they do their thing. Andy Dingley (talk) 15:12, 3 September 2015 (UTC)
    That they are most active through means nothing. I mean many companies have an official web presence which is very stable: a front page which has had many hours, days or even months spent on it getting it to look just right and be most useful. That is the best place to start, when looking for information about the company and we should link to it. Other venues may be more active but that’s as they are less important: Facebook, Youtube, Twitter, Instagram, Myspace. They only exist as e.g. Facebook is so pervasive you have to have a presence. If you don’t your customers/fans/supporters will create one and discuss you anyway. Unless it is their main or only online presence it should not be included.--JohnBlackburnewordsdeeds 17:24, 3 September 2015 (UTC)
    @ Andy Dingley: External links are mostly governed by guidelines, not policy, although the NOT policy is clear that Wikipedia is not a place to promote things. That said, if a Facebook page offers a deeper encyclopedic understanding of the subject than the subject's official website, and assuming it's not promotional, then I see no reason not to list it. - MrX 17:40, 3 September 2015 (UTC)
    Here's the link https://www.facebook.com/Fortressstudygroup What do you reckon?
    The problem is that an IP has been adding this as an inline link to the article at Fortress Study Group and so has inadvertently kicked off an edit war. Andy Dingley (talk) 20:40, 3 September 2015 (UTC)
    No per ELMINOFFICIAL and WP:FACEBOOK --Ronz (talk) 21:04, 3 September 2015 (UTC)
    I would include that FB link. The diagrams, maps, and photos, which can't be found on their website, have encyclopedic value per WP:ELYES#YES 3. This is a good example of where principles are more important than rules. - MrX 21:17, 3 September 2015 (UTC)

    Bug

    There seems to be a bug I think. For example [13] <-- the label does not show up when I see it it shows as "[1]". It should have said "Google". I have to add a space after the vertical line for the label to show -> Google but doing so invalidates the url by adding "%7C" to the end of the URL. David Condrey log talk 07:03, 8 September 2015 (UTC)

    Nevermind. I just realized I'm stupid. David Condrey log talk 07:05, 8 September 2015 (UTC)

    Social media site official or not

    Apparently that Facebook link is an official page, hence it clearly belongs in the article. I think the editor does not understand that www.metal-archives.com is the official site while the facebook page is not even a site, just a page on Facebook. Should it be included or not? Perhaps the wording at EL:NO could be tightened-up to clarify what is meant in situations like this. Walter Görlitz (talk) 02:29, 24 September 2015 (UTC)

    And of course, if I'm wrong, feel free to tell me that as well and confirm the other editor's opinion. Walter Görlitz (talk) 03:02, 24 September 2015 (UTC)
    Two quick answers:
    1. This kind of question should be asked at WP:ELN.
    2. WP:ELPEREN will probably tell you the answer. WhatamIdoing (talk) 03:10, 24 September 2015 (UTC)

    Http vs Https

    I've recently been seeing SPAs changing http to https . I'm not clear on when/if this is appropriate. In the past I've seen it as an attempt to hide spam, but I realize there may be legitimate uses of it. Can anyone explain? --Ronz (talk) 16:41, 9 September 2015 (UTC)

    See WP:Village pump (technical)#HTTPS for selected external links, but how?. That's just one of many discussions on the topic over last few months. The OP there used AWB on a massive number of articles to change http to https. I think the AWB right was removed because it was annoying people (it annoyed me). The issue has been brewing strongly since Wikipedia switched to https only. Johnuniq (talk) 01:59, 10 September 2015 (UTC)
    Thanks! If I'm not overlooking something: https is used for Wikipedia, Google services, and Internet Archive. I see some discussion on other uses, but no concensus. --Ronz (talk) 14:53, 10 September 2015 (UTC)
    HTTPS uses more computing power and offers little to no advantage for viewing ELs. If the reader wants to switch to a secure connection, they have the opportunity to do so. In fact, some secure connections won't work if the link is going through a proxy—the link is actually broken as a result. Walter Görlitz (talk) 02:25, 24 September 2015 (UTC)
    Seriously? Deku-shrub (talk) 17:31, 8 October 2015 (UTC)
    Here are two recent, and in my opinion unnecessary, changes to https: [14] [15]. Walter Görlitz (talk) 14:24, 8 October 2015 (UTC)

    Disallowing search results pages

    Disallowing search results pages may be normal, but I hope to get some feedback on making a general exception clearly stated in this guideline, specifically a clarification of where it says: "Links normally to be avoided"

    9. Any search results pages, such as links to individual website searches, search engines, search aggregators, or RSS feeds.

    Somehow it has happened that the search link {{In title}} or its alias {{intitle}} are serving up search results pages on over ten thousand mainspace pages:

    That a search link has worth as an online navigating tool is undebatable. That it might show up in print is of course not appropriate. This is esp. true if they help build articles and navigate articles. (See below.) Yet {{search link}} will not even preview (a very useful tool indeed) in mainspace because of the (9) rule quoted.

    For the sake of normalizing search links, the list simply needs to be broken down into one list for 8, 9, and 19, and "External links" for the rest. In the section heading, the use of "Links" allows for the two types of URL, one "External links", and one "URL-internal links" or "Internal URL links". It needs to be made very clear that search links class themselves at all times as "selfreference" and "noprint" before they are truly not "links to be avoided". The further problem where the search link is placed as a bullet item

    which would be left bare in print is also relevant here. so we would have

    Links to internal search results pages are an exception where properly classed and where not placed as a bullet item.

    Some other points I'd like to make clear in the guideline about URLS:

    • The first sentence defines "external links". Just pointing this out.
    • An external link is not an internal link. A URL can be either. URL should mean "URL-style of internal wikilink", just like "page", "Search", "server", and the like terms that are used (concerning the Internet) also imply an internal reference unless otherwise stated.
    • A URL is not an external link just because it is handled by the servers for any number of good, internal, reasons.
    • URLs are normal, and needed
      • to urlencode certain characters in page names
      • to access internal (Mediawiki) applications such as History, Talk archive, Diff and Search.
      • to become either a wikilink or URL depending on whether they are on Wikipedia, or a mirror, or various types of forks. In other words URLS are needed anywhere a wikilink is needed: {{srlink}} will created either a wiklink or URL but distinguishing its use might be complicated, and is why we need to make an effort to clarify exceptions.
      • and are so important, that for URLs, developers see that they are very generous in allowing them: there can be literally more than 50 ways to make one URL to a single page.

    Searchable URL were recently a top wish of the community.

    A second, but major concern I have is a case in point. It involves a particular article. Please see the tables at "List of Latin and Greek words commonly used in systematic names". I want to replace the search column with parentheticals that don't show up in print, and are marked "selfreference":

    Latin/Greek
    (Search)
    acaulis

    Search links won't show in a print preview, yet the page authors will have the search tool they need, per the talk page. Is this OK with people? — CpiralCpiral 00:55, 24 December 2015 (UTC)

    The first page from your search results above that I clicked was Hanoi, which has such an internal search in the See also section: Hanoi#See_also ('{{nomirror|{{wikibullet}}{{intitle}}}}', resulting in '
    • {{{1}}}All pages with titles containing External links/Archive 35'). I don't know if this is one of the appropriate applications, that is why we have categories (in fact, the search only results in categories, and one page in Wikipedia-space)(first mainly categories, and some pages in Wikipedia-namespace (meetups, articles for deletion, linkreports), many pages which are already linked in the mother-page and would fail being in a See also section for that - and I don't know if ~225 search results are that useful). There may be utility (though I yet fail to think of examples why this would be needed in mainspace), but I think that thousands of transclusions might need cleanup. --Dirk Beetstra T C 06:36, 24 December 2015 (UTC)
    Yes, I too looked into Hanoi, and when I saw it I added the nomirror and wikibullet fix. Yes, categories help target cleanups and violations. But we still have a guideline and mainspace problem.
    Every single one of those 10,000+ pages precisely violates the "no search results" guideline, and are very useful, each and every one. All 10,000 are currently violating this guideline, even Hanoi, even Hanoi fixed; 5000 can be quickly fixed with "noprint" added to the template, but the other 5000 will then, in print view, leave a stray bullet. (See more details on the stray bullet, intitle cleanup.)
    What I'm look for is an understanding acknowledgment and a solid acceptance of the situation, and how it leads inevitably to a wording change to the guideline, and I'm looking for your opinion on a case in point at List of Latin and Greek words commonly used in systematic names, as shown in a mini mockup here on this page of the tables currently on that page. Try the search link in the table here above, then then see this page in a printable version. Thanks. — CpiralCpiral 07:31, 24 December 2015 (UTC)
    I dispute that that search result in Hanoi is 'very useful'. That search result plainly violates what a See also section is for. For Hanoi it leaves you with 225 subjects (of which quite a couple not encyclopedically related to Hanoi, or unrelated to the city Hanoi (Les Hanois?)). If one does the same on London, one gets nearly 8000 results (with again many pages not encyclopedically related to London). If one does the same for plants, one would even get results not related to the subject at hand (the multicellular eukaryotes of the kingdom Plantae - the last time I checked the muticellular eukaryote Robert Plant, he was not part of the kingdom Plantae), but it would fail to find the Christmas rose (which is a multicellular eukaryote of the kingdom Plantae). And I haven't even started about the Christmas turkey, which is not .. coming from Turkey. In short, whether internally or externally, the indiscriminate linking to search engine results should be done with extreme care and properly restricted as it (as demonstrated) rarely results in the results one wants, and as Wikipedia (or the general internet) is expanding, it may now give you the results you want, tomorrow it may include again totally unrelated subjects. And that is why this guideline is suggesting to avoid search engine results.
    More in general, I think that this guideline talks about external links to search results (google, bing, yahoo, whatever), not to our internal search results. It is hence not for the External Links-guideline to exclude internal links to search results as they are by definition not covered anyway. --Dirk Beetstra T C 08:06, 24 December 2015 (UTC)
    • As the bullet point 9 quoted above is within a page about External Links, it seems to me to be obvious that it has nothing to say about internal links to Wikipedia searches. If there is confusion, perhaps the bullet point should be amended to make this explicit, by amending to "Any external search results...". PamD 08:18, 24 December 2015 (UTC)

    No, wait Dirk. You're jumping ahead. I'm not comfortable discussing the cleanup particulars yet until we have a guideline. Yes we may remove intitle from some uses. Indeed we may debate about a sanctioned search link's search results, whether it could return the wrong "plant" or "turkey" . But first the guideline needs to align with reality, before the cleanup to some as-yet-to-codify conformance. I'm sure we agree and can all understand that there is a cluster of related messes. Now, athough the guideline violation warrants intitle removal from article space entirely, I don't think we've agreed yet that a better solution would be to address the guideline's emptiness concerning "use of URL-internal-links in article space". It needs to distinguish between the external links and the entire class of internal URL as represented partially by misplaced rules 8 and 9.

    Firstly, I hope we're going to proceed to acknowledge that there is a problem with the guideline wording. By flaunting the guideline myriad times, intitle is an example of the guidelines lack of clear addressing. By trying to follow the current guideline template search link has the opposite problem, it bans itself from mainspace, which it should not. I may just edit the guideline and say "per talk page". — CpiralCpiral 21:48, 24 December 2015 (UTC)

    • This proposal is quite confusing and the undiscussed addition to the page seems unclear in purpose and appears to make many assumptions. I suggest that addition be removed while discussion proceeds.olderwiser 09:19, 27 December 2015 (UTC)
    A blind guideline says there are "no search links", while there are ten thousand search links. What more could there be to discuss? Something needs to be said about that situation: A blind search link author has no guide on how to make search links. You make them disappear in a print. (They don't!) I made a guideline that says to authors "For print's sake! Make them disappear". And it says for bureaucrats that the technical provisions and other offline media will not be ignored. (They are!) — CpiralCpiral 12:41, 27 December 2015 (UTC)
    I'm sorry to say that I don't understand the proposal either. I don't know what "a blind guideline" means.— Preceding unsigned comment added by Ronz (talkcontribs) 01:22, 28 December 2015‎
    It just seemed to me that it "ignored" (or was blind to) violations of its own guideline. — CpiralCpiral 09:10, 28 December 2015 (UTC)
    • A comment above says "But 8, and 9 are not external links". I must be missing something because everything at Wikipedia:External links concerns external links. Is there any text in WP:EL which relates to internal links? Is this discussion about the use of {{in title}} and {{intitle}} (the latter is a redirect to the former)? If so, this is the wrong page because those templates are for searching within Wikipedia and are not external links. The templates are possibly dubious, but that should be discussed elsewhere. Johnuniq (talk) 05:35, 28 December 2015 (UTC)
    • Plastikspork (talk · contribs) put a direct reference to this page at {{Search link}} at its inception (over 5 yrs). There are zero instances of Search link in article space, while {{In title}} and {{Look from}} skirt the loose wording here. I will now happily remove my misunderstanding, never again accepting the search link self-restriction, and that topic is done here, but not elsewhere. Thank you all very much.
    • Nope. "External" can refer to the markup, or the server. At External links search it says "To search for external links to pages on this site...". Here cover external and interwiki markup as well as sites. We distinguish links that "providing direct information and internal navigation" from external links. We cover templates for external links including {{plainlinks}} and {{srlink}} (which is both internal and external links). The proposed, is in Wikipedia:External links/Archive 35 § How to link, just as it could be proposed at WP:URL and at Help:Linking, and it just adds some much needed "how to link with class semantics" information. That said, yes, this page could be more cleanly delineated, and the proposed moved to WP:URL or Help:URL (along with other things). I see the proposed as a first guideline attempt at getting the noprint and selfreference semantics clear and exemplified. After the current intitle and looksfrom "noprint" debacle, we obviously need the semantics information spread about. — CpiralCpiral 09:10, 28 December 2015 (UTC)

    I've reverted this.

    • First of all, "external links" is defined in the very first sentence as "links to web pages outside Wikipedia (external links)", so links to pages (e.g., search results) inside Wikipedia is completely off-topic for this guideline. If you want to get this accepted, then it needs to go on some other guideline, or in its own. See WP:PROPOSAL for advice.
    • Secondly, I'm not sure that filling articles with links to Wikipedia search results is a good idea. The links could be contaminated by all sorts of factors, including user settings (what if mine are set to search the Wikipedia talk: namespace by default, and the search results are dominated by AFDs and WikiProject discussions?) and other uses for these words? Africana results are contaminated with non-botanical results, Bios is filled with computer technology that is related only through the accident of an acronym, and Album results are utterly useless. WhatamIdoing (talk) 20:09, 28 December 2015 (UTC)
    • Thanks for resolving that any likeness to the proposed is "completely off topic", and that any likeness to the proposed belongs elsewhere.
    • As you may know, :query searches article space. Any search link can override any default search domain. ({{Search link}} defaults to article space in any case for the reasons you say. {{Search link}} will soon join its weaker cousins to provide services invisible to offline communities.)
    • Any other reason anyone could argue about search links in article space is a style issue, which is up to each page. This must be the case, for even if search links are allowed with the highest conceivable restrictions, at the point of acceptance that they, just as this help page says, provide "direct information and internal navigation" services for online-community content-building, its up to them to decide how much noise in their search results they decide to accept. Ten new intitles are added/day. Currently on 2% of dab pages and in over 2000 articles that are not dab pages. Happy new years! — CpiralCpiral 23:50, 28 December 2015 (UTC)

    In summary, this help page does not and cannot disallow Wikipedia-based search links in article space. (That can be done elsewhere.) Nor is this page the best place for the description of HTML noprint or selfreference semantics because only Wikimedia-server-related URLs need them. Thanks all. — CpiralCpiral 23:50, 28 December 2015 (UTC)

    I have argued above, and with that I agree with Cpiral, that this is not the place to disallow internal search links (technically, we are not disallowing even the external search results, but strongly disfavouring them for problems). However, this guideline is for a part build on the pillar WP:NOT, and I do think that that gives a strong argument against external, but also internal, search results. The one example I followed above (Hanoi), and expanding that to some other examples, shows that search results, whether internal or external do not give reliable results, and I do think that WP:NOT disallows these. There may be use for the template, but I believe that there are currently way too many transclusions which should be drastically cleaned out. --Dirk Beetstra T C 07:24, 30 December 2015 (UTC)
    Beetstra, decisions need to be made because the situation is out of hand. Opinions solicited at WP:NOT and MOS:SEEALSO have occurred to me. I am hardly in favor of removing 'em all, but rather lean towards a descriptive over a prescriptive approach. E.g. when or where would an online-only otherwise-invisible element ever be used (at the bottom of a list, the top of a section, a hatnote, an entire section, within a navbox) wherever the MOS:LAYOUT could recommend. — CpiralCpiral 08:59, 30 December 2015 (UTC)
    @Cpiral: I agree that the situation is out of hand, but I think that the decision has already been made in the consensus that resulted in the current version of WP:NOT. This type of searches should IMHO not be in See also sections, general text, lists, external links sections, and only be used very, very sparingly in specific cases (maaaayybbee on every disambig, in an 'infobox-like' setting). --Dirk Beetstra T C 05:21, 4 January 2016 (UTC)

    YouTube

    Just a note about the change I just reverted:

    The "YouTube" section needs to work equally well for both extremes on the whole YouTube-or-not-YouTube continuum. The first sentence is directed at the people who remove too many such links. The second sentence is directed at the people who add too many such links. Both of these positions are wrong, so we don't want to re-write it to favor either of them. The concept is neither "Although they're technically accepted, they're bad", nor is it "These are definitely permitted, as long as there is no LINKVIO here".

    In general, YouTube links are more widely accepted now than they were when this section saw its last major re-write. The increased acceptance is due to an increase in the number of official channels and a change in people's expectations about rich content (i.e., more people have computers and internet connections that can stream video easily). WhatamIdoing (talk) 22:54, 15 November 2015 (UTC)

    @WhatamIdoing: I hear you. Please see how this new revision goes. Fleet Command (talk) 08:38, 16 November 2015 (UTC)
    That looks acceptable to me. Thanks for boldly trying it. WhatamIdoing (talk) 16:21, 16 November 2015 (UTC)
    Yeah there is much more legal official material now than in the old days, when Youtube links where often under default assumption to link to copyright violations. That doesn't really apply anymore and many artists, broadcaster, institutions, universities use Youtube as as distribution media to offer (legal) content to the public.--Kmhkmh (talk) 09:06, 4 January 2016 (UTC)

    There doesn't seem to be any info about how Template:Cite_patent creates an inline external link that by the nature of the code seems to be allowed. Should this and any other coded exceptions be addressed in this article?Timtempleton (talk) 19:35, 20 January 2016 (UTC)

    I certainly wish templates that create external links indicated the policies that apply, EL and REF in this case. At least this one documents recommended usage. --Ronz (talk) 22:22, 20 January 2016 (UTC)
    User:Jonesey95 kindly informed me that the {{cite patent}} template isn't supposed to be used in the body of an article and updated the notes for that. Didn't know that.Timtempleton (talk) 02:51, 21 January 2016 (UTC)
    All I did was link to the existing guideline, specifically the section that includes this: "External links should not normally be used in the body of an article." The guideline seems pretty clear to me. – Jonesey95 (talk) 04:20, 21 January 2016 (UTC)
    Yes - the guidelines for external linking are clear, except when templates are involved that automatically include an external link, as in the case I stumbled across with the cite patent template. More clarification is great in this case.Timtempleton (talk) 08:59, 21 January 2016 (UTC)
    I'm guessing that you didn't notice the bold text in this guideline's lead paragraphs. The external links guideline doesn't apply to citations. Many citation templates use external links, and as long as the templates are placed within <ref> tags, there's nothing to worry about. - Eureka Lott 14:33, 21 January 2016 (UTC)

    Language

    Any objection to this change? Nyttend (talk) 23:12, 5 April 2016 (UTC)

    ELNO gets tossed around as if it were ELNEVER. I agree that normally we should avoid foreign language links, but I'm concerned about the implications of my former statement.... --Izno (talk) 23:22, 5 April 2016 (UTC)

    Please note discussion, which may affect this page, at Wikipedia:Village pump (technical)#Preferred protocol for external links templates. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits

    Having found no consensus there, I have now opened an RfC on the matter. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:30, 8 April 2016 (UTC)

    I came here looking for some guideline regarding whether in-text external links are acceptable. I had always assumed they weren't and so expected to find a statement to that effect on this page, but I didn't. Is there some policy/guideline page that says anything about this subject? Everymorning (talk) 15:25, 14 April 2016 (UTC)

    Point 2 under WP:ELPOINTS. DonIago (talk) 15:30, 14 April 2016 (UTC)
    Thanks. Everymorning (talk) 15:36, 14 April 2016 (UTC)
    People still do it anyway. It says don't do it "normally" but "exceptions are rare", and then gives three examples: links to Wiktionary, Wikisource, {{external media}}. It should say "exceptions should be rare" cuz they're not really that rare, people do it -- mostly because of not knowing better, sometimes out of laziness or not caring. Drives you nuts. Wiktionary links are fine, but other than that external links don't belong in the body of the text. Herostratus (talk) 15:55, 14 April 2016 (UTC)

    What's so special about broadcast media articles?

    • WP:ELPOINTS states that: "In the "External links" section, try to avoid separate links to multiple pages in the same website; instead, try to find an appropriate linking page within the site."
    • WP:LINKSTOAVOID adds that links to be avoided include: "Any search results pages, such as links to individual website searches, search engines, search aggregators, or RSS feeds."

    Yet in numerous broadcast media articles we have such abominations as Iowa Public Television#External links and Wisconsin Public Television#External links. Both fly in the face of the guidelines to avoid links to multiple pages in a single website and links to search results pages. The numerous links in those EL sections are clearly generated by searches of an FCC database and a private company's database. Yet when I try to simplify the External links section by linking only to the databases themselves, I'm told by some media fanboy that linking to multiple search result pages is standard operating procedure. What gives? Why are broadcast media articles allowed to flout the guidelines when one simple link to the database itself could substitute for numerous links to numerous database search results? 32.218.33.187 (talk) 20:25, 18 May 2016 (UTC)

    We want users to find information in only one or two clicks, thus the FCC templates which link right to the station page. Your solution throws a user into the FCC database haphazardly without any guidance, where they have no idea where to type anything. We have them for every station, even a state network with 10-20 stations. Read WP:TVS#Tasks; it's been unilaterally agreed to for a decade that an FCC template for each station is appropriate and meets ELNO just fine (and since the station's histories would just be repeating that it carried a state network over and over, we see no need to break out that monotonous information onto separate station pages at this point). The goal of Wikipedia is making information easy to access, not hard, and if it takes a template repeating 15 times, so be it. I stand by keeping the page as-is, and as for your 'fanboy' comment, I ask that you be WP:CIVIL. Nate (chatter) 22:38, 18 May 2016 (UTC)
    "haphazardly without any guidance, where they have no idea where to type anything."
    Seriously, Mrschimpf? What about "call sign" or "city" are hard to understand? Wikipedia readers aren't morons. I ask that you treat them with a modicum of respect. (BTW, You still haven't addressed why the criteria you made up apply to broadcast media articles, but no others.) 32.218.33.187 (talk) 23:02, 18 May 2016 (UTC)
    You're the only editor in eleven years who has had an objection to the use of TVQ's (along with AMQ/FMQ for their radio sisters), so I stand by the use of them as common MO for all broadcasting pages. Again, the common TVS policy has been linked for you to read; any objections, the talk page is next door. And I never made any aspersions about the ability of readers to understand; we link them to the proper information (in this case a government agency where all of the information is known and true), and the goal of an external link should never to be to confuse or frustrate the reader, no matter their level of understanding. The main FCC query page could easily do that. There is no "made-up criteria", but common policy that all the possible information is linked out in a page. Nate (chatter) 01:32, 19 May 2016 (UTC)
    Look, Mrschimpf, if I had wanted to have a two-way discussion with you about this matter, I would have taken this issue to your talk page. But I already know from your repeated ownership behavior what you think. I brought the matter here so that disinterested parties could weigh in on the matter. So why don't you just stop wikistalking me, sit on your hands for a while, and give some objective editors a chance to comment? 32.218.33.187 (talk) 02:58, 19 May 2016 (UTC)
    If you have an editing issue, you're supposed to inform the conflicting party; you did not, I found it was being discussed and said why I reverted your edits. I also meant WP:TVS's talk page, not mine, and all editors are free to comment on anything here. You're removing common templates without any discussion, which can be considered vandalism, and you're refusing to discuss the issue civilly. Again, I invite you to bring this up at WT:TVS for interested parties in the subject to comment on. Nate (chatter) 04:08, 19 May 2016 (UTC)

    The IP did correctly, and bring the discussion to a relevant noticeboard (this is not an issue where a local wikiproject can overrule policies/guidelines, so this should be brought to a wider audience). I can agree that it would have been good form to notify an editor with whom they ran into this issue. Nate, you are asking the editor to be civil, but "You're removing common templates without any discussion, which can be considered vandalism, and you're refusing to discuss the issue civilly .." is a chilling and contradicting comment in itself, implying that the editor, who in good faith tried to improve the page as they had concerns that the links were improper, is a vandal.

    Now back to the subject of the issue - those linkfarms. And User:Mrschimpf, the criteria you give are in conflict with our inclusion standards - "We want users to find information in only one or two clicks" <-> we are not writing an internet directory for finding database entries for radio stations which are indirectly connected to the subject of the page (see WP:ELNO). Moreover, the technical information provided by the links does not help in helping the understanding of the subject of the page, and much of the necessary data can be incorporated (and sometimes already is) incorporated into the page (and could/should then be used as a reference for that specific piece of information, which it sometimes should but is not). As such, even if those lists are there for years (see WP:OTHERCRAPEXISTS), they can still be in violation of our inclusion standards, and I second the IPs concerns and think that they should be removed (in fact, they should now even be removed until there is consensus on inclusion - and even a unilateral consensus from one wikiproject does NOT overrule policies and guidelines, and this local consensus is completely against the broader consensus of said policies and guidelines). --Dirk Beetstra T C 05:07, 19 May 2016 (UTC)

    So the removal of a common template which has been used without incident for eleven years can be considered 'not vandalism'? The TVQ template leads to government information about television stations (and literally sources all aspects of Infobox TV, so basically we would have to remove it and [citation needed] nearly everything involving the infobox or a station's technical info), and you're asking us to overrule a decade-long consensus for the sake of page beautification brought up by one IP. I've been trained for years to assume blanking a template(s)=vandalism. Apologies if I have been mislead on that and that I have been working off TVS consensus over the years and that it's apparently in conflict of NOT#DIRECTORY I've never heard of before. This should be brought to a broader forum, I agree, but my asking of the proposal to take this to WT:TVS was merely to inform interested editors of the issue so that someone in TVS with template experience might be able to figure out a solution to this for state networks and which would apply in turn to WP:WPRS, which uses the same design for the AMQ/FMQ templates. If my statement was chilling, I do apologize for that but there needs to be some kind of consensus before a widely-used template is removed wholesale from pages. Nate (chatter) 17:21, 19 May 2016 (UTC)
    This is about the lists in the external links sections - you are now suggesting that it is referencing the information in the infoboxes, then it should be a proper reference for that information, not an external link. And on the articles references, they are just a directory and certainly not linked to the infobox (the links removed in this edit are not references to any infobox information - they could be a reference on the individual station's articles, but also there not an external link). --Dirk Beetstra T C 06:31, 20 May 2016 (UTC)

    Semi-protected edit request on 21 May 2016

    The name of the movie used in sample external links should be "Babbette's Feast" rather than "Babette's Feasts" Ccvickery (talk) 22:10, 21 May 2016 (UTC)

    Done — Andy W. (talk ·ctb) 22:51, 21 May 2016 (UTC)
    Church of Saint Giles (pl) Erected in the 1220s at Ostrów Tumski, Wrocław

    Is a link to a foreign language wiki regarded an external link? Should such inline links (like the picture's caption on the right at the Wroclaw article) to foreign wikis be avoided? HerkusMonte (talk) 16:54, 31 May 2016 (UTC)

    No, it's not regarded as an external link, although some interwiki links to non-Wikipedia sites are available, and I think consensus would support treating them with extreme caution (that is, remove them unless very helpful for the article concerned). Sorry I don't have the information at hand, but there is a page somewhere saying that links to other-language Wikipedias are ok if no enwiki article is available. In fact, {{Interlanguage link}} (or its redirect {{Ill}}) can be used for that. Like every piece of text in an article, people might discuss whether an interwiki link is helpful and if it should be removed. However, there is no guideline supporting removal. Johnuniq (talk) 00:27, 1 June 2016 (UTC)
    Thanks, German wiki has such a policy, thought it would be the same here. HerkusMonte (talk) 06:50, 1 June 2016 (UTC)

    Question about {{Allmovie}} (9 July 2016).

    According to WP:ELMAYBE, the recommendation to consider professional reviews as external links was repealed [...]. The reviews should instead be used as sources in a "Reception" section. Why then is {{AllMovie}} the only one professional review to have a template, and where should it be used? Thanks. (Ps: the problem has arisen here). --Mauro Lanari (talk) 11:52, 9 July 2016 (UTC)

    Wikipedia:External links#Templates for external links doesn't explain why AllMovie is an exception with a template of its own. --Mauro Lanari (talk) 15:41, 9 July 2016 (UTC)
    AllMovie isn't one review only. It aggregates film metadata and user rating too. In that respect, it is like IMDb. FleetCommand (Speak your mind!) 17:12, 9 July 2016 (UTC)

    wanting agreement before allowing these changes

    An editor (User:Fuhghettaboutit) has changed "should not" (link to copyrighted material) to "may not" link to copyrighted material, which is stronger language. I'm not saying its wrong I'm just pointing this out.

    Also the editor changed the text to clarify or specify that copyrighted material may not be linked to anywhere in the Wikipedia, including talk pages. This new language would disallow, on a talk page, something like "Look at this link, I think its copyrighted but I'm not sure, what do you guys think?

    I'm not saying any of this is necessarily wrong, but let's get agreement that this is an improvement before doing, and the editors should do this for important rules generally. Herostratus (talk) 22:47, 22 July 2016 (UTC)

    This is an improvement. Walter Görlitz (talk) 04:11, 26 July 2016 (UTC)
    Better, indeed. I suggest to revert again to the improved version (though continuing discussion if needed). --Dirk Beetstra T C 05:00, 26 July 2016 (UTC)

    Victim hotlines

    Was suggested to post this question here. Replies would be most helpful if made on Talk:Domestic violence.

    Regarding this article, is there any precedence for not including a portion of the external links section for large, national, almost certainly non-profit, victim hotlines? I realize WP is not for advocacy, but I'm not sure victim advocacy is really the type that policy is worried about. Comments welcome. TimothyJosephWood 21:49, 28 July 2016 (UTC)

    What do I do if I find a hijacked link?

    The article says that links can be hijacked (i.e. the domain expires and someone unrelated takes it), but doesn't say what to do about it. It's not a dead link, so I shouldn't add Template:Dead link. But what should I do? The article as it is doesn't say. -Thunderforge (talk) 01:08, 17 June 2016 (UTC)

    The quick answer is to remove any external link that is not useful for readers with an encyclopedic interest in the topic (that is, which does not satisfy WP:EL). Sometimes, perhaps on more technical articles, it might be apparent that the link might have had really useful information and I have sometimes Googled trying to find if there is a new location so I can update the link. If a reference includes a hijacked link...well, that's a matter for some other noticeboard I guess, but I might use an HTML comment to remove such a link if I couldn't find an update. Johnuniq (talk) 01:50, 17 June 2016 (UTC)
    The quick answer is to cite the specific version of the page that did have useful information on it, using the method at Help:Using the Wayback Machine if you don't have time to find a better reference. "Do not delete a URL solely because the URL does not work any longer." -- Wikipedia:Link rot. --DavidCary (talk) 17:03, 29 July 2016 (UTC)

    I think we went through this before, a while back? Not sure. Anyway, an editor (User:Fuhghettaboutit) made a change; I've bolded some text just to highlight the change (there's no bolding in the actual text). She changed this:

    Policy: material that violates the copyrights of others per contributors' rights and obligations should not be linked

    to this:

    Policy: material that violates the copyrights of others per contributors' rights and obligations may not be linked anywhere on Wikipedia

    This may be an improvement, but I've reverted it for the time being for the following reasons and invite discussion toward consensus first.

    1. It is almost never a good idea to make substantial changes to policies without discussing beforehand. This is an exception to WP:BOLD. Otherwise there's always the danger of substantive policy changes slipping in without really being broadly accepted. I've seen this.
    2. On the merits of scope, I'm not certain that this page should proclaim a policy that covers all pages and all situations of the Wikipedia, rather than just external links. I would think that WP:External links should perhaps limit itself to just proclaiming rules about external links? So if this sentence should indeed be a rule, perhaps it should go one of the copyright rule pages.
    3. On the merits of the rule itself, this new text would forbid someone writing, on a talk page, "Hey guys take a look at this link [link]. I think it's OK but it might be copyvio, can I get some opinions on whether it is OK to include as an external link?" I think we want to be careful about forbidding questions like that so I would like to see more discussion on that first.

    Since the section ways "editors are restricted from linking to the following, without exception", it seems certain that the new text is OK. Is it helpful? It just doubles and triples down on the restriction which is fine, but makes doubly sure that a #3 type link on a talk page or on the EL noticeboard would be reverted on sight, and that seems the likely main difference. Do we want to be doubly sure of this?

    Also, WP:COPYVIO "Copyright infringing material should also not be linked to" and it then points to this page. This seems odd since these kind of infringing links are also likely to come up as reference citation links. It should link to WP:COPYLINK you'd think, and WP:COPYLINK might be the place for this material, or maybe it should be in both places. Herostratus (talk) 17:00, 19 August 2016 (UTC)

    I have started a discussion about the use of Google translate links in citations and external links sections. Please comment there. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:22, 19 August 2016 (UTC)

    Ancestry.com

    We have around 25 pages with links to the home page of ancestry.com, for example, on Joseph F. Ambrose, we have:

    <ref>{{cite web|title=1920 Census, Census Place: Joliet Ward 3, Will, Illinois; Roll: T625_415; Page: 10A; Enumeration District: 166;|url=http://ancestry.com|accessdate=12 November 2015}}</ref>

    and:

    <ref>{{cite web|title=1940 Census; Census Place: Joliet, Will, Illinois; Roll: T627_907; Page: 62B; Enumeration District: 99-28 Source Information|url=http://ancestry.com|publisher=United States of America, Bureau of the Census. Fifteenth Census of the United States, 1930.|accessdate=12 November 2015}}</ref>

    Is there any reason no to remove the URL from these citations, and convert them to {{Cite census}}, or whatever suits the particular case? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:31, 12 September 2016 (UTC)

    I would indeed strongly suggest to make those links more correct, either deeplink to the correct information, or replace them with a reference that backs up the claim. --Dirk Beetstra T C 20:33, 12 September 2016 (UTC)

    PDFs - Acceptable or not?

    There appears to be a view: Wikipedia:External_links/Noticeboard#Are_PDFs_allowed_in_.22Further_reading.22_or_.22External_links.22_sections.3F that PDFs are unacceptable as ELs. This is ridiculous, let us clarify forthwith that they are acceptable. Andy Dingley (talk) 00:23, 28 August 2016 (UTC)

    The most recent thread I found, from 2012, was Wikipedia_talk:External_links/Archive_33#PDFs_should_be_exempt_from_no-link_rule, where the only concern seemed to be "browser-based exploits". But we allow them in citations, which readers are also expected to click on to veify material. Allowing them in one place but not another doesn't make sense to me. Disallowing PDFs as sources (assuling they otherwise meet the standards) doesn't make sense to me, nor does having a technical prohibition that applies to one set of links but not another, placed right next to them. The current dispute aside, this prohibition may be outdated or poorly conceived. Felsic2 (talk) 00:31, 28 August 2016 (UTC)
    It should be a source if it's that type of document. EL's are supposed to elucidate on the article. Opening up EL's to documents, videos, etc, just get around the WP:RS requirement. It would be a BLP nightmare if we allow exceptions that aren't easily verifiable as directly topical to the article. --DHeyward (talk) 01:38, 28 August 2016 (UTC)
    Not always. For example, I know of many large government projects that are notable, and that can be sourced to very detailed documents, but the gov't agency overseeing the project also offers a "easy" fact sheet. We'd likely never reference the fact sheet, opting to use the more technical sources, but an EL to the fact sheet seems completely appropriate. I agree that many PDFs can be used directly as references, but not always, and there's clearly no disallowance to link to PDFs as long as they are being published by the agency that has the copyright to them (eg no links to things like academia.com copies). --MASEM (t) 02:00, 28 August 2016 (UTC)
    I don't see how allowing PDFs would alter the BLP landscape. The same rules should apply to content regardless of whether the link goes to a PDF or HTML page. A PDF is just as verifiable, or unverifiable, as a webpage. Felsic2 (talk) 02:09, 28 August 2016 (UTC)
    it's easily seen in the case you are trying to shoehorn in. vpc.org is not linkable at the root level as it is a partisan, political organization with a prominent "Donate Here" button. Outside the WP article on vpc.org where it would be ELOFFICIAL, it should never be an external link. The fact that they created a pdf doesn't separate them from that problem. That's quite different from Masem's example of a summary from reliable sources (though our article should summarize such a document making linking unnecessary as it is included in the article and sourced to the same detailed, technical source that the summary used - e.g. a "good article" on global warming would include the essential "Summary for policy Makers" statements in the article, sourced to the details. The EL would be to IPCC as a whole, not the pdf summary for policy makers.). The pdf Felsic2 is proposing doesn't pass muster at the site level even though the pdf is written to look like "research." vpc.org is a partisan advocacy site and linking to the site brings up political donation requests. The pdf is buried and hides it but doesn't change the character of the authors. Youtube hosts a number of videos in very common formats and there are very few that I would trust as a direct EL for a living person simply because it's so easy to publish without review. Ubiquity of the player is not a reason to bypass the guideline and we should be linking to straightforward sites that can provide the pdf or youtube link as their own reference. Directly linking and deep liking is problematic and is one of the reasons ELNO exists. --DHeyward (talk) 10:48, 28 August 2016 (UTC)

    I don't see why we should exclude material just because it is a pdf. We do have discouraging rules for material that needs special software to be installed, but since pdf displaying software is quite widespread that should not be too much of a hurdle here. We do of course have to comply with the rest of the inclusion standards as described by this guideline (e.g., no linkfarms of 50 pdfs, or links to tangentially related pdfs after we have here decided that pdfs are fine). --Dirk Beetstra T C 03:32, 28 August 2016 (UTC)

    This is moot. PDFs are already in widespread use on Wikipedia as references and external links. There is no impending "BLP nightmare" to fear. PDF is just a document format, like HTML, XML, doc, txt, etc.- MrX 03:33, 28 August 2016 (UTC)
    That's not the full story. There was a time when viewing a pdf could install malware on the computer of anyone using the Adobe reader. Adobe would patch the bug, and a few days later a new disaster would be discovered. That went on for a couple of years and is an indication that Adobe software is fundamentally broken and should not be relied on for any purpose on the Internet. That consideration does not rule out pdf links because obviously they are useful and are used, but the story is more complicated than "PDF is just a document format". Johnuniq (talk) 03:42, 28 August 2016 (UTC)
    You can say the same, only worse, for links to HTML documents. If there is a security problem with the web at large, then we must have no ELs whatsoever, anywhere. We don't even link to "HTML" anyway, we link to a HTTP URL and that can return anything it likes. Andy Dingley (talk) 09:22, 28 August 2016 (UTC)
    (edit conflict) @MrX: with respect to this guideline, pdf is not a document format like HTML, txt - the browser you are using to display this page is interpreting HTML and txt. The browser is not capable of reading a pdf or a doc, by far most browsers do need an external piece of software for that to be installed. That most people have that installed (or even that that is standard installed) is not a reason to plainly include it, it still falls under the #8 rule and editors still need to keep that in mind when adding external links, even to pdfs. --Dirk Beetstra T C 03:43, 28 August 2016 (UTC)
    Thanks, but I understand quite well how documents work and PDF is a document format and technically HTML is markup language. My point about it being extremely common still applies. This particular guideline serves no practical purpose and is apparently stuck in c. 2003.- MrX 04:36, 28 August 2016 (UTC)
    @MrX: No, it does generalise - there are places where computer technology is not that advanced and where pdfs are not necessarily easy to access (we had a discussion along those lines not too long ago). But I fully agree that pdfs are extremely widely used and commonly installed (I said something along those lines in my initial post). That does not make this a 'html, pdf, xml, txt, etc. etc. are all free to add here'-place. Even for pages which are fully html are restricted under parts of our 'links to avoid'. --Dirk Beetstra T C 05:18, 28 August 2016 (UTC)
    There is also the concern that direct linking bypasses the publisher. A pdf can be generated to look like it is a Nature article but if it's hosted and buried in a psuedoscience site, our readers won't know. Where it lives is important context both in evaluation for the reader and other editors. Direct linked PDFs hide the nature of the parent as do other formats. Their should be a page that is simple html on the parent website that references these pdfs or videos. If those pages are press releases or partisan/pov sites, it's a much better indication of the usefuleness of the document for an external link. For example we wouldn't want a direct link to a creationist websites pdf that looks like reviewed and neutral material on evolution. Coupled with the realization that n EL to the top level creationist site would be innapropriate in a evolution article, it becomes pretty obvious why we want simple, direct links that show the source of the material and not a link that triggers a reader that hides the source. --DHeyward (talk) 11:06, 28 August 2016 (UTC)
    Most of those are better reasons for excluding PDFs from citation than from external links. Is there any reason why we'd allow them one place but not another? Felsic2 (talk) 21:32, 17 September 2016 (UTC)

    Can we come up with some language to express when a PDF link would be acceptable in an external link or citation? Felsic2 (talk) 19:30, 28 August 2016 (UTC)

    Best archive date

    Hi there,

    When choosing an archive url for a citation, should I use a date closest to the access-date parameter (in the cite web template) or the date in which the url first appears in the article? -Tim1357 talk|poke 04:03, 24 September 2016 (UTC)

    Closest to the access date. Edits can be made to web sources for corrections. Walter Görlitz (talk) 18:02, 24 September 2016 (UTC)
    @Walter Görlitz: But in the case when there is no accessdate, then I should find an archive closest to the date of first addition to the article? Tim1357 talk|poke 02:02, 25 September 2016 (UTC)
    Also, is it appropriate to set the access-date to the date of first addition to the article, if it was not present in the template before? Tim1357 talk|poke 02:07, 25 September 2016 (UTC)
    If there is no accessdate... access the article and then find the closest archive. If you did not access the article on the date of first addition, then the accessdate is not an accessdate for you, and if you didn't add it then, you don't know what it said then. --Izno (talk) 03:26, 25 September 2016 (UTC)

    Need help figuring out how to best edit the hatnote for this article, to allow a more intuitive redirect

    I was editing a company article and saw that someone had put in multiple external links to different product sections on the company website. I knew that was a no-no, and went to find the guidelines on external links to confirm. I typed in "external link" in the Wikipedia search box, and clicked on the result, which was a redirect to Internal link. I thought - that's strange - and changed the redirect to Wikipedia:External links. I then realized that people looking for information about external links outside of the context of Wikipedia were going to be left hanging. There's nothing in the hatnote for this article to help those readers. In addition, since WP:EL already links to this article, I can't figure out how to add a note saying "external links redirects here." If someone could fix the hatnote to let readers know a) that both the WP:EL and external links articles redirect here, and 2) that for information about external links in web coding, they should go to Hyperlink, I could then make the redirect more intuitive. The internal links and Hyperlink articles should probably also be merged, with a subsection in the Hyperlink article covering the difference between external versus internal, but that's a bigger task.Timtempleton (talk) 20:03, 26 October 2016 (UTC)

    External link redirects to Internal link which already has a hatnote at the top with a link to Wikipedia:External links, so nothing needs to be done. "Wikipedia:" in front of Wikipedia:External links means it's not an article in the encyclopedia but a page in the Wikipedia:Project namespace which is for editors and not readers. The search box is mainly for readers and only searches articles by default. Add "wp:" in front of a search to search project space instead. There generally shouldn't be redirects from mainspace (articles) to other namespaces except in rare cases for terms which are implausible search terms for readers. See Wikipedia:Cross-namespace redirects. PrimeHunter (talk) 22:43, 26 October 2016 (UTC)
    Based on what you're saying, then the only thing to do would be to merge the internal links and hyperlink articles.Timtempleton (talk) 18:08, 27 October 2016 (UTC)

    Can the template {{cite web}} be used for External links section? For example, instead of:

    External links


    wouldn't it be better to have provide details like:

    External links
    • "Province: Abra". Philippine Standard Geographic Code. Philippine Statistics Authority. Retrieved 11 October 2016.


    Sanglahi86 (talk) 07:27, 11 October 2016 (UTC)

    @Sanglahi86: there are no restrictions to using a template to provide the link (we also have specialised external-links templates, like {{official website}}), and I agree that that in some cases gives more clarity. --Dirk Beetstra T C 07:37, 11 October 2016 (UTC)
    Thank you very much for the information. Sanglahi86 (talk) 07:47, 11 October 2016 (UTC)
    @Sanglahi86: If you need to use {{Cite web}}, you might like have a "Further reading" section instead. See MOS:LAYOUT for details. FleetCommand (Speak your mind!) 12:41, 11 October 2016 (UTC)
    I see no reason why {{Cite web}} cannot be used on regular external links either .. "Official website". Wikimedia. Retrieved 11 October 2016.. --Dirk Beetstra T C 13:31, 11 October 2016 (UTC)
    Then its a good thing I didn't say "don't use it there". But if you have anything against reading MOS:LAYOUT or just considering a "Further reading" section, please don't hold back. FleetCommand (Speak your mind!) 22:54, 12 October 2016 (UTC)

    Sanglahi86, please see WP:ELCITE (especially the paragraph under the colorful example), which discourages the use of citation templates. In particular, a visible access date is inappropriate for these links, because ==External links== need to be working every day, and get removed immediately if the link is dead, so the fact that it worked on a particular date is irrelevant. WhatamIdoing (talk) 07:10, 29 October 2016 (UTC)

    Thank you very much for the valuable information. I will keep that in mind. Sanglahi86 (talk) 14:00, 29 October 2016 (UTC)
    There are multiple "correct" ways to do this. If I were adding the link, then I might choose something like this:
    But this is only one option among many. WhatamIdoing (talk) 19:38, 1 November 2016 (UTC)

    User:Comfr, the problem with saying "the external links in the mobile phone article are acceptable" is that we don't know that they're acceptable in every single version of that page. We know that they were acceptable (in your opinion) in the particular revision of the article that you happened to see. We also know that another editor, who looked at exactly same set of external links as you, disagreed with your claim that those links were all acceptable. But we don't know if editors who read this guideline tomorrow, or next month, or next year, and who go to that article to see examples of what's acceptable will actually find only "acceptable" links there. Based on the article's history, they will probably find only acceptable links, but they might find a pile of spam.

    So we're not going to tell them "here's a good example (90% of the time)". We're instead going to say "Don't put this kind of bad link in that kind of article, because we're tired of reverting spam". WhatamIdoing (talk) 19:35, 1 November 2016 (UTC)

    User:WhatamIdoing, thanks for bringing this up. I agree with everything you say. My only problem is dealing with a double negative. The first negative is the section title, "Links normally to be avoided." Everything is that section is about what not to do. Then I get to item 5, containing an example with a second negative, "does not link." The implication here is that the example contains only good links, which as you pointed out, might always be true.
    I reworded the sentence, only because I was confused the first time I read it. No matter how the sentence is worded, the external links in the phone article may be changed at any time. (Thanks for pointing this out.) For consistency, I suggest either remove the example entirely, or provide an example of something that is not allowed. The section contains two other examples. One example proposes alternatives, and the other example demonstrates prohibited links.
    How about, "For example, an article about cell-phones must not link to web pages that mostly promote or advertise cell-phone products or services."
    Thanks for your many contributions to Wikipedia. Comfr (talk) 23:31, 1 November 2016 (UTC)
    We used to have much more trouble with this sort of link. I haven't seen so much of it recently, though – although User:Beetstra will have a much more informed opinion. What do you think, Dirk? Can we omit the example altogether, or should we try to find a way to make it clearer? WhatamIdoing (talk) 17:23, 2 November 2016 (UTC)
    The 05:59, 31 October 2016 edit changed the example to read In contrast, the external links in the mobile phone article are acceptable, because they do not link to web pages that mostly promote or advertise cell-phone products or services. That is not satisfactory for the reason explained by WhatamIdoing above. There is no problem with the current wording, and it is not a problem if a reader has to pause and think about the issue to understand WP:ELNO#5. Johnuniq (talk) 01:18, 3 November 2016 (UTC)
    @WhatamIdoing: As sole links I do not see this too much either (anymore). I do see sometimes that they are added as extra links (one official page, and then a sales page on the same domain), but that is also infrequent. Of course, many official homepages of subjects will fail, strictly reading, WP:ELNO#5 (often an about page is more neutral, as we tend to demand on whitelisting).
    {{rto|all}} I am thinking .. For example, an article about mobile phones should not link to a page selling or promoting many different mobile phones, but rather could link to a page giving a neutral overview of many different mobile phones (suggested sentence for further discussion). --Dirk Beetstra T C 03:21, 3 November 2016 (UTC)
    {{rto|all}} What initially confused me was the word, "does," instead of "must" or "should." The above suggestion is much better, because User:Beetstra uses the word "should." I would support any change that eliminates the word "does." Comfr (talk) 16:56, 3 November 2016 (UTC)
    Have we reached consensus? If will wait a few more days, and if no one objects, I will change "does not" to "should not." Comfr (talk) 04:04, 8 November 2016 (UTC)

    academia.edu

    This diff is my response to Witger who has been posting several links to the personal profile of Frans Van Droogenbroeck on the registration-required academic social (or professional) networking site academia.edu. I have asked the user for a rationale for such links, but in the meantime have reverted them. It might be that there are open access versions of sources authored by that individual available through his profile on that site, but even if this is the case I am very leery of external links to the social (or professional) networking sites of individuals who are not the subject of the article in which the link is placed. If these are open access publications, one would assume they would be available on official sites of publishers or institutions, as well as on the author's personal social networking page. It seems to me that this might skirt the line of acceptability, and I would appreciate a word about consensus of which side of the line it falls. --Andreas Philopater (talk) 13:42, 11 November 2016 (UTC)

    There's an example at this diff of somebody linking to resources on academia.edu rather than to a profile. That seems to me to come down just the other side of acceptability, but what is the consensus? --Andreas Philopater (talk) 15:28, 11 November 2016 (UTC)
    It's blatant spamming by Witger. --Ronz (talk) 18:06, 11 November 2016 (UTC)

    Under § Links normally to be avoided, should "Blogs, personal web pages and most fansites (negative ones included), except those written by a recognized authority" also specify largely user-generated content sites such as the Internet Movie Database? Wikipedia:Identifying reliable sources includes such sites under Self-published sources, which are "largely not acceptable". —Sangdeboeuf (talk) 01:06, 21 November 2016 (UTC)

    I tend to regard IMDB as a 'wiki'. IMDB is not reliable, and not suitable as a reference. Although I do personally not fully agree with it, IMDB on a subject generally contains more information and detail than what Wikipedia articles on the same subject should include, and therefore IMDB is generally included as an external link. See WP:ELPEREN. --Dirk Beetstra T C 03:12, 21 November 2016 (UTC)

    Rapidly changing data

    I suggest that "Links to be considered" should inlcude information that is subject to frequent change. One often sees lists and tables on Wikipedia that are a maintenance nightmare. They either need constant attention to keep up to date, or any ever growing procession of snapshots have to be posted periodically. It is much better for our readers to refer them to the accurate source information. SpinningSpark 15:48, 26 November 2016 (UTC)

    Good point. In my experience, "rapidly" just means faster than editors will or can be expected to maintain. This is a huge problem, and it would be nice to develop guidelines to address it. --Ronz (talk) 16:24, 26 November 2016 (UTC)
    The article that prompted this is FTSE MIB which gets reviewed every three months. It is currently entirely out of date and no hope that it will ever be properly maintained. Things that change no more than annually may have a case to be here; it is possible that someone might take an annual snapshot and that kind of historic information is what an encyclopaedia should be about, but any more frequently and we are just posting WP:NOTNEWS. That's not the only example, several others spring to mind. Rubik's cube contains information on current solver record holders. That changes very frequently, but I have to admit it is kept reasonably up to date because the holders are always keen to see it is. Other bad ones are various lists of military strength and hardware by country (sorry, I can't give you any links, I took them off my watchlist years ago because they were such a chore to watch over). Those are a bit more of a difficult case because they tend to be referenced (if they are referenced at all) from a great variety of sources. SpinningSpark 17:18, 26 November 2016 (UTC)
    I think you should consider talking to WP:WikiProject Wikidata about this issue. WhatamIdoing (talk) 00:10, 5 January 2017 (UTC)
    @WhatamIdoing: Wikidata is not and was not set up to combat rapidly changing data. Rapidly changing data is one of the motivating factors for the Commons data sets projects which was just deployed to Commons. --Izno (talk) 00:28, 5 January 2017 (UTC)
    I don't think that this is rapidly changing data. I think this is just a lot of data – with possibly a new number each day. The fact that the stock index will be X tomorrow does not change the fact that it was A last Tuesday and B last Wednesday. WhatamIdoing (talk) 04:13, 5 January 2017 (UTC)
    @WhatamIdoing: For the purposes of Wikidata, rapidly changing is basically anything changing on the basis of monthly periods or smaller. (I don't know why you're arguing your opinion—I'm telling you this as a fact from the Wikidata point of view.) --Izno (talk) 13:08, 5 January 2017 (UTC)
    I'm trying to figure out why you say that this data is changing. The number for 31 December 2016 will never change. Once entered, that data point should never be changed. You will get more data each day, but the prior data should remain unchanged.
    Now, perhaps Wikidata does not want to have datasets that contain daily information (e.g., amount of precipitation in a given location each day, historical stock or commodity prices, etc.). But "we don't do daily data" is not the same thing as "we don't do information that changes". Historical prices don't change. WhatamIdoing (talk) 17:26, 5 January 2017 (UTC)
    The number for 31 December 2016 will never change.  Once entered, that data point should never be changed.  You will get more data each day, but the prior data should remain unchanged. That's a rather pedantic way at looking at what I said originally. But sure, I don't disagree with you now. --Izno (talk) 17:45, 5 January 2017 (UTC)

    The Lists of Soviet films article shows people -- directors and actors -- either without links at all, or redded-out -- who actually have significant articles in Russian language Wikipedia. Certainly better than nothing, so I tried inserting such a link (with the language flag). Unfortunately when I paste it into the edit box, Cyrillic becomes gobbledygook: https://ru.wikipedia.org/wiki/%D0%9C%D0%B5%D0%BB%D1%8C%D0%BD%D0%B8%D0%BA%D0%BE%D0%B2,_%D0%92%D0%B8%D1%82%D0%B0%D0%BB%D0%B8%D0%B9_%D0%92%D1%8F%D1%87%D0%B5%D1%81%D0%BB%D0%B0%D0%B2%D0%BE%D0%B2%D0%B8%D1%87 (this is for the Russian article for director Мельников, Виталий Вячеславович -- film director Vitaly Melnikov). Is there a way to add links to other language articles so they remain legible (in that language) in the edit box AND functional? LADave (talk) 20:06, 10 January 2017 (UTC)

    Try ru:Мельников,_Виталий_Вячеславович. - MrOllie (talk) 20:28, 10 January 2017 (UTC)
    You can still use the ru article's title. That said, maybe you should instead use Template:ill with |WD= and/or the Russian link. --Izno (talk) 20:43, 10 January 2017 (UTC)

    Official website template

    A bot is changing all instances of links to official websites to {{official website}}. [16] It has something to do with Wikidata.

    The guideline says: "Use of the template {{official website}} is optional." Because of that, I've asked the bot operator to stop until consensus is established, but he argues that this is a trivial issue. Is he correct that no one will mind, do you know? SarahSV (talk) 16:25, 19 December 2016 (UTC)

    In general, I'm against forcing editors to use templates such as these. They add another layer of objects that an editor has to keep an eye on, and any future changes to the format of the template may make the display different to other external links in an article. I haven't seen a convincing argument as to why mandating the use of templates on external links would help this wiki. Hchc2009 (talk) 16:29, 19 December 2016 (UTC)
    Thanks Hchc2009. I believe the eventual aim is that they would be controlled by Wikidata. Where is the best place for the bot operator to ask for consensus if he wants to continue? SarahSV (talk) 17:05, 19 December 2016 (UTC)
    I'd suggest an RfC here? It is this guideline which would need changing, and an RfC would give it wider visibility. Hchc2009 (talk) 17:12, 19 December 2016 (UTC)
    Good idea. Okay, I'll suggest that. SarahSV (talk) 17:43, 19 December 2016 (UTC)
    The discussion with the bot operator is at User talk:Ladsgroup#Cosmetic edits. SarahSV (talk) 16:42, 20 December 2016 (UTC)
    Before noticing this, I put some details at WP:External links/Noticeboard#Official website uses Wikidata. I haven't thought much about the issue but there are dubious features. First, someone monitoring a few articles would have to also monitor the Wikidata pages (not easy in practice!) if they wanted to see when a passing editor or spammer had changed the official website. Second, the move increases the general techno-babble associated with editing—wikitext is gloriously simple and using unnecessary templates is, well, unnecessary. Third, it is only a matter of time before someone cleverly makes {{official website}} also show the "official" Facebook/Twitter/Fad-of-the-day URLs, not to mention the "official" websites for every continent and more. Johnuniq (talk) 00:12, 20 December 2016 (UTC)
    I would revert that last problem as a pillar violation, though I can see that we then get a discussion over that which is going to be a battle. And that over and over if it at first doesn't succeed.
    It does indeed also increase the number of targets for spammers. WikiData will now have to handle that for all wikis ...
    Then I noticed the following issue, which is believe is going to make editors mad beyond believe (it already had me) .. blacklisting the official site can result in funny problems. It is now technically possible to add an official site (through WikiData) while it is blacklisted here. And the 'next editor' on the page will run into the spam-blacklist filter while not adding external links. --Dirk Beetstra T C 03:39, 20 December 2016 (UTC)
    "...it is only a matter of time before..." please avoid such FUD. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:50, 20 December 2016 (UTC)
    @Pigsonthewing: the first copy of {{official website}} has already sprung in existence, where practically every transclusion that I have seen was a violation of our inclusion standards .. I don't think that this is FUD, it is reality/Murphies Law. --Dirk Beetstra T C 17:57, 20 December 2016 (UTC)

    Te template has many benefits. It populates categories such as

    We can't enforce people use it but we continue using a bot to convert links. -- Magioladitis (talk) 13:12, 20 December 2016 (UTC)

    1 and 3 are not beneficial to Wikipedia. I don't care at all if the website is different here or there. And 2 is empty (but that may be due to other bot work of course, I don't know). 1 and 3 each already have some 20,000 articles in them, so that's another artificial backlog of 40,000 articles thanks to Wikidata and some bots. 60 Minutes supposedly has a different official website here and on Wikidata. Ours is http://www.cbsnews.com/60-minutes/ Wikidata has http://www.cbsnews.com/60-minutes. Hurrah, lost a minute to find out that we have a "/" at the end which they are missing, but which makes zero difference. What an utter waste of time and resources. I guess some general discussion about the Wikidata invasion is needed (at Wikiproject Cycling, they have trouble enough stopping enthusiastic Wikidata spammers from taking over our layout as well, even though the Wikidata layout doesn't even follow the accessability guidelines), but for now let's simply and clearly oppose this template, these categories, and the bots adding the template everywhere. Fram (talk) 13:45, 20 December 2016 (UTC)
    Fram Adding this step in the conversion, protects Wikidata too. In the past we had problems with data insertion from English Wikipedia.
    The code was improved to recognise minor changes as slashes. Opposing the categories is not beneficial in any way.
    Wikidata was created in order to reduce inconsistency between various Wikipedia sites. -- Magioladitis (talk) 13:47, 20 December 2016 (UTC)
    I couldn't give a flying fuck about protecting Wikidata from bots, it's hard enough protecting enwiki from bots and semi-automated editing. Wikidata was a minor good idea when it was used for interwikilinks. Now, not so much (if at all). "The code was improved to recognise minor changes as slashes." Really? Strange, as my check above was done at the time of writing. The improvements seem to have very little effect if the very first example I check, on a not really obscure article, already gives this result. Pushing these maintenance categories is not beneficial in any way. Another example, the second one I check: 2010 Australian Open. We have http://www.australianopen.com/en_AU/index.html, they have http://www.ausopen.com/index.html Both are "wrong" (well, not really, both work) as they result in http://event.ausopen.com/ Will changing the article or the Wikidata entry improve anything? No, not at all. If people at Wikidata want such categories, they can run bots there and add categories there. But who cares that the official website on our site is different from the official website on another unreliable wiki? Maintenance categories where no actual maintenance is needed are not beneficial. Fram (talk) 13:59, 20 December 2016 (UTC)
    1 and 3 are both beneficial to Wikipedia. As for "Wikidata invasion", please refer to WP:AGF. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:50, 20 December 2016 (UTC)

    @Frietjes and Ladsgroup: to help with the tracking categories if possible. -- Magioladitis (talk) 14:33, 20 December 2016 (UTC)

    Magioladitis, what tracking categories are missing, or needing modification? Frietjes (talk) 14:36, 20 December 2016 (UTC)
    Frietjes http://foo.com and http://foo.com/ should be recognised as the same item. Is this possible? -- Magioladitis (talk) 14:41, 20 December 2016 (UTC)
    Magioladitis, yes, just use
    {{#ifeq:{{#invoke:string|replace|http://foo.com|/$||plain=false}}|{{#invoke:string|replace|http://foo.com/|/$||plain=false}}|true|false}}
    , replacing the two URLs with parameters and replacing the 'true' and 'false' with the categories or lack of categories as desired. Frietjes (talk) 14:46, 20 December 2016 (UTC)
    Magioladitis, added something similar to the LUA module. Frietjes (talk) 14:50, 20 December 2016 (UTC)

    Another issue is dead links. External links last on overage about 7 years before going dead, so in time most instances of {{official}} will contain a dead link. How is this dealt with? Do we just replace the URL with a URL to wayback? If so, the link is no longer the official URL, but a URL to somewhere else (Wayback, Webcite etc). It creates problems with mismatch at Wikidata since the link rot bot InternetArchiveBot does not run on Wikidata (IABot is adding about 5000 waybacklinks to enwiki a day, including replacing links in {{official}}). Or when a link dies do we replace {{official}} with {{webarchive}}? Or tack on an extra {{webarchive}}? Or should {{official}} contain a new argument |archiveurl= and |archivedate= like CS1|2? These questions apply to all external links templates. -- GreenC 15:33, 20 December 2016 (UTC)

    Even this would be easier to handle with a template. We'll find a way. -- Magioladitis (talk) 16:06, 20 December 2016 (UTC)

    As a Wikidata-editor, I feel that the main spam-problem is import from Wikipedia, among them en.wikipedia are among the greatest! Your page here: Malmö tells that it has a "website". But, does bricks on the ground have websites? Well, they can have fansites, tourist information sites, but official website? No! www.malmo.se is the official website of Malmö City, but as you can see, you have a completely different page for that! Malmö is also located in Burlöv Municipality, which have another official website: www.burlov.se. Why haven't you spammed that to the article? That there is a parameter website in a template, does not mean that it has to be filled every time the template is installed into a page. -- Innocent bystander (talk) 17:04, 20 December 2016 (UTC)

    Innocent bystander True but we don't discuss a parameter website in a template. We discuss a template that every time you type it it will display the official site. -- Magioladitis (talk) 17:15, 20 December 2016 (UTC)
    It is however an important observation. Malmö has the website set here in the template, but the same website is also available on WikiData, and may not be correct. I think this is very similar to what User:Fram means by '...the official website on our site is different from the official website on another unreliable wiki'. We have enough troubles with unreliability on en.wikipedia, now we get automagical transclusion of more unreliable information from another wiki. Similar to automatic translation of unreliable pages from another wiki. --Dirk Beetstra T C 17:21, 20 December 2016 (UTC)
    @Magioladitis: You should be aware of that the "official website"-property includes a little to much of "non-official websites", since it has been filled with data from the "website"-parameters in the Wikipedia sites. People fail to recognize the difference between "official website" and "other websites". I have used this property very little (not at all actually) on svwiki, since I know it to often is full of sh*t.
    @Beetstra: You should be aware of how many times I have removed this property for places in Sweden. But the users who adds them outnumbers me...
    The last thing I have added to our templates on svwiki, is the option to only transclude WD-claims that has a real source. To not use those claims who have those "imported from:"-references. -- Innocent bystander (talk) 19:38, 20 December 2016 (UTC)

    Innocent bystander There is a more general problem with external link anyway. Thank for the heads up though. -- Magioladitis (talk) 19:40, 20 December 2016 (UTC)

    Previous discussions

    -- Magioladitis (talk) 22:30, 20 December 2016 (UTC)

    @Magioladitis: so this is the first time that we have a wider discussion about it? No VP discussions whether this should be implemented? --Dirk Beetstra T C 03:36, 21 December 2016 (UTC)
    Dirk Beetstra Full implementation has not discussed yet. See Wikipedia:Bots/Requests for approval/KasparBot for instance. We need this conversion so as GoingBatty said "so humans can go back and make the appropriate fixes". The conversion only enables the tracking categories without any other further change. It is true thought that a lot of editors already use the Wikidata values. This also happens in Infoboxes while it has not fully decided yet. As you may see this is a slow discussion that spans in a period of 2.5 years.-- Magioladitis (talk) 07:45, 21 December 2016 (UTC)
    Maybe Pigsonthewing or Od Mishehu know more and want to comment. -- Magioladitis (talk) 07:53, 21 December 2016 (UTC)
    Recall that the tracking categories were created in 2015. So, the arguments here should be whether we want to track differences between en.wp and Wikidata. No migration has ben decided nor implemented in large scale. -- Magioladitis (talk) 07:55, 21 December 2016 (UTC)
    @Magioladitis: What I meant was that there does not seem to be a wider discussion whether this template should use/display WikiData values (whether that, for this template, is desirable, whether it is 'correct', etc.). I am in favour of using WikiData for identifiers and certain immutable data, but that does not go for every possible datapoint that is available from WikiData. For this template, I am weary about several problems (and I see several concerns above), and I think that this specific data is not good data to transclude from WikiData due to those concerns. Whether you want to track is a different question (for that, except for the category flood at the bottom that I generally ignore anyway, I couldn't care less). --Dirk Beetstra T C 09:25, 21 December 2016 (UTC)
    @Beetstra: The fat of the template is not for me to judge. I understand your concerns. I wish we ha more feedback from the Wikidata team on the matter. -- Magioladitis (talk) 09:28, 21 December 2016 (UTC)
    Ehm, sorry, but who cares what the WikiData team (or even, WMF) has to say. I am more concerned about what the en.wikipedia community has to say. --Dirk Beetstra T C 09:47, 21 December 2016 (UTC)

    "We need this conversion so as GoingBatty said "so humans can go back and make the appropriate fixes". Then implement it on Wikidata. It's not logical that every wiki-language version will check whether the wikidata value and their own value are the same. Wikidata can do this centrally. And then the Wikidata editors can go through these thousands of "problems" to check whether any really are problematic and need changing. 2016 Summer Olympics is included in this category, but needs no change on enwiki. 2016 Paris–Roubaix is on that list because we link to the English language version, Wikidata links to the French language version. No change on enwiki needed. 2016 Ryder Cup: we have the right one, Wikidata is outdated (but works as well, it just redirects to what we have). No action on enwiki needed. Académie des Inscriptions et Belles-Lettres: we have it without the http, they have it with the http, no change needed. Nicolas Achten, Didier Agathe, all the same issue. Aerodynos JA 177 Evolution, link doesn't work on Wikidata, ours (to webarchive) does...

    Basically, this is a lot of bot edits, a lot of system checks needing to be done to populate the categories, and then thousands of manual checks to do the "necessary maintenance", all for extremely little actual benefit (yes, there are undoubtedly some cases where the link on enwiki no longer works, and the one on wikidata does, but this seems to be very rare compared to cases where no change is needed or cases where Wikidata is wrong). Add to this the extra chances of errors and vandalism being introduced through Wikidata, which will go largely unnoticed (I will not start checking what "(‎Created claim: Property:P3143: 817530; ‎Created claim: Property:P3138: 294549)" means), and you end up with a net negative.

    Don't add the template, or change the template to ignore wikidata completely (no maintenance categories, no picking up information from Wikidata) Fram (talk) 09:33, 21 December 2016 (UTC)

    I do actually agree. Remove the WikiData-date retrieval, and then, there is actually no need to have the maintenance categories. This is just not the type of data that you want to transclude from there, it is too much editor specific, often incorrect, and gives editing problems. Write a script to pick up the official websites from every page here, and compare that with the WikiData value if you want to improve the data coverage on WikiData, don't abuse en.wikipedia for that. --Dirk Beetstra T C 09:56, 21 December 2016 (UTC)

    I don't agree. There is zero reason not to get the official-x link/ID from Wikidata, and quite frankly the above looks more like anti-Wikidata rabble rousing than it does like useful concerns. Someone has the problem, and whether that problem is here or there (really it's objectively mostly here) is really just a question of "We (certain specific persons) just don't want to deal with it". So don't. People who are interested will make the appropriate changes as necessary and the worst you'll have is edits you can already hide on your watch list.

    Specifically:

    1. I see no editing problems besides the rabble rousing above and perhaps some unfortunate BRFA-approved bots.
    2. I don't see how the information is editor specific.
    3. The statements above re: checking data there: either these are misinformed (in that the best you can do is set up some bots and update static pages) or they are trying to shift work away. As I said, the work should be done regardless, and saying "it's not my problem" is certainly not in the spirit of collaboration.
    4. Regarding languages on Wikidata, those actually do need change such that the official website has a language of work qualifier, so that there are both official websites there. Last I checked the module here correctly looks for the English qualified version first.
    In general: everybody here wants Wikipedia (and Wikidata) to be better (so I might advise a dose of WP:AGF). I'm happy to answer questions if you (plural) have them, but the above indicates some (possibly pre-)judgement on some points or another, rather than fair assessment of the work these people are doing. Specifically, people can work on whatever they want here, and Fram, it is not your job to tell them what they can or cannot do if there is indeed value to their work. That you elect not to consider it valuable is irrelevant to the facts of the situation. --Izno (talk) 10:55, 21 December 2016 (UTC)
    And one more thing: this discussion is largely premature, and wrongly placed. If we want to talk about this, the right place is Template talk:official website, so that editors interested in that template are tuned in to this discussion. --Izno (talk) 11:00, 21 December 2016 (UTC)
    @Izno: no, I would say that this needs to be discussed on a global forum like a Village Pump. Does the community agree that ALL possible data should be transcluded from WikiData, does the community agree that NO data whatsoever should be transcluded from WikiData, or does the community think that that has to be evaluated for each situation independently. If it is the first or the second, then we are ready, if it is the latter, the community needs to see whether there are general guidelines that can show what data should be transcluded and which not, or whether that really needs to be discussed project/community wide for each parameter. Seen the concerns here, and the scope of pages (all of them), I would suggest that this needs broad community input.
    "I see no editing problems besides the rabble rousing above" Nice way to get people to collaborate. "There is zero reason not to get the official-x link/ID from Wikidata" except for all the reasons that have been explained in vain above. "Someone has the problem, and whether that problem is here or there (really it's objectively mostly here)": using "objectively" is all very well, but the facts seem to contradict you. ""We (certain specific persons) just don't want to deal with it". No, we don't believe there is anything to deal with, or certainly not in this labour-intensive, vandalism-prone way with very little result. "I don't see how the information is editor specific." I don't think anyone claimed it was editor-specific, seems like a strawman. "As I said, the work should be done regardless, and saying "it's not my problem" is certainly not in the spirit of collaboration." You start from a wrong position: "the work should be done regardless" is simply not true. Letting a deadlinkbot loose on official websites will be a lot more productive than this. "everybody here wants Wikipedia (and Wikidata) to be better": no, I don't. I don't care how good or (usually) bad Wikidata is, as far as I am concerned everything but the interwikilinks may be deleted from it, as in most cases all I notice is people trying to push data (and layout) from the unreliable Wikidata upon enwiki, making it harder to edit, track changes, prevent vandalism. If people want to use enwiki as a source for Wikidata and then use Wikidata to populate other wiki-languages, and those are happy with that, be my guest. But I see no reason to mass-import data from Wikidata here; most of the things there are unsourced and the quality isn't any better than what we have here.
    The discussion, by the way, doesn't belong on a template talk page either, it belongs at VPP.
    "Specifically, people can work on whatever they want here, and Fram, it is not your job to tell them what they can or cannot do if there is indeed value to their work. That you elect not to consider it valuable is irrelevant to the facts of the situation." I see that my opinion of whether their work adds value is irrelevant, while yours is objective and universally shared and should be accepted as gospel. On the other hand, you have not shown any bit of criticism to be actually invalid, you have only claimed it to be so. Basically, all you have done is some ad hominem attacks (arbble rousing) and some strawman arguments, coupled with arguments from self-declared authority. Yep, that's very convincing. Fram (talk) 11:13, 21 December 2016 (UTC)

    We should think globally and act locally. Not think only about the English Wikipedia. I am an editor in multiple Wikipedias. I try to spend as much time I spend here to other Wikipedias too. -- Magioladitis (talk) 10:56, 21 December 2016 (UTC)

    Zero reason?

    1. The data needs to be locally overridden in case of local blacklisting - causing discrepancies that we have to deal with.
    2. local blacklisting of WikiData-transcluded domains does give editing problems (and very unclear ones!) here.
    3. Official websites of subjects are not always 1-on-1 defined - something that our editors have to deal with
    4. It is difficult to control here what is happening on WikiData, e.g. spam - I have just reverted and blacklisted a massive set of links of a massive sockpuppet set of official homepage altering spam accounts - and that turned out to be the second set from them - we have to deal with that.

    If you think that the community has different thoughts, then you would urge you to gauge what the community at large thinks. This one has been locally implemented, has only been discussed in local forums (and sometimes hardly discussed, just implemented), and has, until now, even not been notified to the policy and guideline pages that govern the data that should be transcluded, let alone global forums. As I said earlier, I am not anti-WikiData - there is a lot of information that can be transcluded and which I (have) encourage(d) to be transcluded. This is not one of them. --Dirk Beetstra T C 11:22, 21 December 2016 (UTC)

    Magioladitis, can the template/module check the spam blacklist? WhatamIdoing (talk) 04:00, 5 January 2017 (UTC)
    WhatamIdoing I think Frietjes could answer this. -- Magioladitis (talk) 17:40, 11 January 2017 (UTC)
    a lua module on this site should be able to load [17]. Frietjes (talk) 19:05, 11 January 2017 (UTC)
    Thank you. Magioladitis, I think it would be a good idea to update {{official website}} to check for spam. This is a problem that could affect all of the active communities, so if we (i.e., not me ;-) could solve it once here and "export" the Lua module everywhere, we'd save people the trouble of re-inventing the wheel. WhatamIdoing (talk) 18:21, 12 January 2017 (UTC)