Wikipedia talk:External links/workshop

Moved
Hi,

I have moved the page to main Wikipedia space to leave room for discussion and more detailed propositions here. It was quite "bold" of me, but I guess it is quite a minor change, and the benefits are clear.

Regards,

Bravada, talk - 23:10, 11 October 2006 (UTC)

PS. Below is my proposition for the "not a web directory" section, addressing some concerns raised by others and some of my own :D

Wikipedia is not a web directory
While some external links are welcome in the article (see below), one should remember that it is not Wikipedia's purpose to include a list of all external links related to the topic. If the site or page you want to link to includes information that is not yet a part of the article, please consider using it as a source first (see WP:CITE for guidance).

Please also see about the Open Directory Project, which is entirely devoted to creating relevant directories of links pertaining to various topics. It is recommended to link to a related Open Directory category from an article rather to create a long list of external links within the article itself. If there is no relevant category, you may want to create one.

If you belive an external links section has gotten out of hand, then you can mark it for attention of other editors by placing External links at the top of the section.

By not using mid-level section headings, we eliminate the overexposure of DMOZ, while at the same time maintaining the strong recommendation to use it in lieu of WP as a links directory (obviously). I am not that satisfied with my wording, so suggestions on how to brush it up (while retaining the meaning and "spirit") are welcome.

I would also have reservation about this "valuable service" thing, as it might be misunderstood. I would say:


 * "However, adding some relevant external links is permitted and recommended. See below for details."

Simple and, I guess, clear.

Oh, another thing. For the vast majority of any media, be it books, music or visual arts, any site containing a copy would violate copyright. I would move it to "links to be used occassionally", and reword it to express the extreme caution that should be taken before linking to such site - I guess only un-copyrighted works would qualify.

Regards, Bravada, talk - 23:26, 11 October 2006 (UTC)


 * I also wonder if we should create a template to request moving links to DMOZ or creation of an apropriate category. --Barberio 09:20, 12 October 2006 (UTC)


 * That's a very good idea! Bravada, talk - 21:28, 12 October 2006 (UTC)


 * There are a number of editors from ODP who also edit here (like me) who would be eager to help out. Also, don't forget about the dmoz template, which could probably use a mention here. &mdash;Wrathchild (talk) 17:45, 19 October 2006 (UTC)
 * I like the wording re: Directory request. I expect that very often the ODP will have a directory but that people will have trouble negotiating the hierarchy. An ODP editor should probably be able to zero in on the appropriate directory very easily. &mdash;Wrathchild (talk) 16:37, 23 October 2006 (UTC)

Avoid special exceptions, Avoid instruction creep
We realy need to stop putting in complicated special exceptions for circumstances.

Idealy, a rule set should be simple, and applicable to situations as they arise. Especialy considering the general disclaimer of 'guidelines are just guideines', we don't need to put in cavets that you can link to a Fansite if it's especialy notable, or officialy endorsed. We need to ween people off this kind of beuocracy, rather than adding to it. (see Avoid instruction creep)

As an aside, linking to fansites just to link to fansites goes directly against the 'Wikipedia is not a Linkings Directory' policy. It is not a guidelines place to overide policy. I see no reason to recomend people to link to fansites. And they should be included in the general recomendation against 'spurious unrelated links' which applys in a more general way. If there is special reason to link to a fansite, then it's not a spurious unrelated link. No other special wording is needed, so under Avoid instruction creep it doesn't need it's own rule. --Barberio 22:57, 18 October 2006 (UTC)

Again, and I can't stress this enough, we should keep the guidelines offered here to the minimum required! Too many rules for specific cases where a general rule already applies with the same result is a bad thing, and multiple rules that can be consolidated into a single general rule should be consolidated.

Copying over a rule from the current page's and demanding they it be included as stated there defeats the point of the rewrite. As we ask editors to keep links to a minimum because Wikipedia is not a link directory, so we should keep rules to a minimum because Wikipedia is not a Bureaucracy. --Barberio 00:55, 22 October 2006 (UTC)

Undue weight
The following seems to imply that we must have link(s) for each point of view. "On articles with multiple points of view, the number of links dedicated to one point of view should not overwhelm the number dedicated to any other." There are some cases where the minority point of view is so small that it shouldn't be given any links. Or the minority point of view has a separate subarticle (in compliance with WP:SUMMARY), and links dedicated to that point of view go on that subarticle.

I have in mind here September 11, 2001 attacks and the minority "view" that are the 9/11 conspiracy theories. Per WP:NPOV, there are no conspiracy sites linked on the main 9/11 attacks article. Though, we have a small summary paragraph in the article about these theories, with a link to the subarticle. On 9/11 conspiracy theories, there are conspiracy links (and debunking links). To conspiracy theorists, they view their theories as "alternative views" and see themselves as "9/11 researchers". Though, in pouring over reliable sources, the conspiracy theories have very little weight and should be treated as such on the main article. Number 3 of "links normally to be avoided" might cover conspiracy sites, but might contradict "Avoid undue weight" with its present wording.

I suggest this sentence be reworded, though I'm not sure yet how it should be reworded. --Aude (talk) 15:45, 19 October 2006 (UTC)


 * How about just state the obvious? External links should conform to Wikipedia policy on undue weight. Fagstein 01:46, 20 October 2006 (UTC)

Subsection format in the section itself
Well, I will just paste what I wrote here. Hopefully you will understand why I object using === instead of ; in the external links section.

I heavily contest this section about using === instead of ;. Rationale: Hopefully, we will remove the === syntax from this section and force the usage of ;, as it is much better (at least, from my point of view) than the === one. -- ReyBrujo 05:55, 23 October 2006 (UTC)
 * 1) By using ===, you are cluttering the table of contents of the article. External links should just be a section in there, not a full tree of possibilities. See the current state of the external links section of the Jew article (it has 9 subsections in the table of contents!).
 * 2) It makes much easier for people to add their own spam links. They just need to click the Edit subsection in the "Fansites" list to add their own. In example, see this revision of the Grand Theft Auto: Liberty City Stories article, this revision of the Kabbalah, or the previously mentioned Jew article.
 * 3) Most warnings are put just below the External links heading, either as a template or as a HTML comment. By allowing them to edit a sub section of the heading they are bypassing the warning.
 * 4) When a responsible user wants to add or modify an external link from the section, if he needs to browse through the full external link section because of ;, he will be encouraged to eliminate as many of those as possible to make navigation easier.


 * The rewrite doesn't advise using subsections at all. For the reason that a linklist that is long enough to require subsections is almost certanly too long a list of links. --Barberio 00:44, 27 October 2006 (UTC)

Mulitiple pages
"Editors should try to avoid linking to multiple pages from the same website, instead trying to find an apropriate page to link to that will provide links within the site." Is this not at odds with a need for links to be to a specific page? I can imagine spammers using this as an excuse to link to site which is only generally applicable to the article. I don't have a strong view about this BTW. &mdash; Moondyne 06:19, 23 October 2006 (UTC)
 * I think we want to ballance that against keeping links to a minimum. I have seem several articles which have an over long links section because they link to multiple pages from the same site. --Barberio 09:54, 23 October 2006 (UTC)

Proposed changes
While some of the workshop seems like a nice improvement, some new additions of detail work against its stated aim. I've removed them from the workship and listing them here.
 * 1. The suggestion to create a Dmoz category is not the business of the Wikipedia.
 * No, but there is discussion afoot among ODP editors to work more closely with Wikipedians in this regard. I'm working on putting together a WikiProject. &mdash;Wrathchild (talk) 04:51, 27 October 2006 (UTC)
 * 2. "If you believe that an external links section has gotten out of hand..." is an invitation to trolling, and not appropriate at the top of the document. A paragraph at the bottom, with more text than the self-important encouragement of "if you think..." would be a good idea.
 * 3. The one or two links to reviews makes no sense at all in a document that wants to avoid bureaucratic creep.  Things called "reviews" should be dealt with like any other webpages.  there is no reason to encourage them, nor limit them arbitraily.
 * 4. "...for example, links to various fansites devoted to a TV show, blogs commenting on a news event, or forum posts about a video game..." is both entirely unnecessary and again troll fodder. This sentence: "There is often a great temptation to add sites that are only indirectly related to the article's subject" is both well-concieved and on the point. There is no reason to make general examples, particularly the fansite one which is just plain wrong. Links to various fansites devoted to a TV show will often be excellent linking that follows this (and the old) guideline.
 * 5. "Links to websites run by anonymous individuals." Where did this nonsense come from?  Anonymity is not a comment on usefulness.  It is no criteria at all for many things... while it would be a criteria for new-sy type things.  It certainly is no criteria for lighter subjects.
 * 6. In general, i could go either way on whether the guideline should be rewritten this way. If these five changes were implemented, I'd be fine with moving the workshop over.2005 20:19, 26 October 2006 (UTC)


 * 1) Being a directory is not Wikipedia's buisness, however I see no reason at all why we can't send effort towards Dmoz as an apropriate open directory. It serves the purpose of re-enforcing that Wikipedia is not a directory, by clearly saying where the content would be welcome instead. And helps by ensuring that there are apropriate places on Dmoz to link to under our guideline to do so.
 * 2) Propose alternative wording that you feel would work better, but I do think it needs to stay near the top as maintenece mechanisims are important.
 * 3) This is a phrase that has consistantly been in there since 2005 when it was formed as a full guidelien. It would *not* otherwise be suitable to link to reviews under any other rule we have, since they would only be indirectly related links. There is no other generalist rule I could see to merge it into unless you see otherwise. Merging it with 'indirectly related links' would just be moving the rule from one categorary to another, and making the guidelines a little more obscure.
 * 4) Examples are made, because otherwise people will make argument for laxness that was not intended, the intent is to show that we have a large scope defenition of 'indirectly related sites'. Links like the examples may well be acceptable links, if and only if they pass one of the two guides for inclusion. It is not an outright proscriptive listing, but a stricter scruitiny.
 * 5) I'm neither for or against this, since no good arguments have been made either way. (I will note that the fundamental linking requirments are not solely based on utility of the link.) If no one has any good reasons to keep it, it should be dropped.--Barberio 00:40, 27 October 2006 (UTC)


 * 7. And a new seventh point I missed before. The workshop emphasizes "relevance" in the section right at the top but still kind of hidden. As mentioned on the main article talk page, relevance is no criteria to emphasize here as hopelessly rotten sites can be highly relevant to the article. I added the word "valuable" instead, but it could be "appropriate" or something else. The relevant point is made later in distinguishing links directly about the article versus broader ones. 2005 21:24, 27 October 2006 (UTC)

Application of Wikipedia policies
I think perhaps the guideline could be streamlined by adding a consolidated section on how the guideline relates to various Wikipedia policies. WP:NOT is already addressed in the beginning, and WP:NPOV is addressed near the bottom, under "Undue weight". What about WP:V &mdash; any thoughts on how verifiability applies to external links? Schi 21:50, 26 October 2006 (UTC)
 * It doesn't. That would lead to never ending tail chasing of then verifying the content that verifies the external link.  Besides that, the Foundation will never "verify" anything about an external link as the legal ramifications of that are endless.  External links are further reading, unverified and unendorsed by the Wikipedia. 2005 21:58, 26 October 2006 (UTC)
 * That's what I suspected, thanks for the input. Schi 17:53, 27 October 2006 (UTC)

What should be linked to
3 and 4 under "What should be linked to" seem very similar to me: 3. Sites that contain neutral and accurate material not already in the article. Ideally this content should be integrated into the Wikipedia article, in which case the link would remain as a reference. However, in some cases this is not possible because of a site's copyright issues, unencyclopedic level of detail, or other reasons. 4. Sites with other meaningful, relevant content that is not suitable for inclusion in an article, such as professional athlete statistics, screen credits, interviews, or online textbooks. Can someone explain the reason for this distinction? I suspect it has to do with the "neutral and accurate" clause in #3, but I'm not coming up with any convincing examples of what would fall under #4 but not #3. Thanks. Schi 17:53, 27 October 2006 (UTC)


 * I've moved this line down to the 'to be considered' list. They are subtly different, but yes, fulfilling #3 means it fulfils #4, while #4 is a slightly weaker restriction. Links that only satisfy #4 should be 'weaker' than the others, and drooped in preference to them, so it belongs under 'to be considered'. Also because this is a more subjective requirement that would need some consideration. --Barberio 10:29, 2 November 2006 (UTC)


 * The two are completely different and of equal importance. #3 addresses sites with great detail, like a 200 page site about everything having to do with the Boer War. #4 addresses sites with things we don't include, like a list of a baseball player's batting averages and fielding statistics over a 20 year career.  In other words, #4 talks about things that should never be in an article, like huge lists of statistics, while #3 is massive volume of stuff that could conceivably be used but there is just far more of it than we can use.  Agai, these are completely different and no less important than each other. 2005 22:18, 2 November 2006 (UTC)
 * I interpret your explanation as defining #3 and #4 as sites presenting data not to be included in the article for the same reason: (massive volumes of) un-encyclopedic level of detail, which is covered by #3. Statistics like batting averages should be in articles when they are notable and meaningful statistics (a lifetime average would probably be useful; I can imagine cases where a batting average for one season could be particularly notable and singled out). For example, see Mickey Mantle, a listed Good Article, which includes various relevant and notable statistics (batting averages in specific years, lengths of various home runs); or featured article, Sandy Koufax, which includes many relevant statistics (ERAs, winning percentages).  Another featured article, Wayne Gretzky, has a whole section on career statistics.  Statistics should be included in articles (where appropriate), but obviously, huge lists of them, like huge lists of anything, should not be included. Same goes for screen credits or interviews &mdash; they're neutral and accurate material but not suitable for inclusion because of their un-encyclopedic level of detail. Schi 22:54, 2 November 2006 (UTC)
 * I think you're making a distinction without a difference here. There is no functional difference between the two lines, except that one is a laxer version of the other. Your given reason is that #4 covers items "we would include but there is too much of it", which seems to me to be "we would include it, but for the 'wikipedia is not an indiscriminate collection of information' policy". This is actually the a subset of "information we can't include". --Barberio 23:07, 2 November 2006 (UTC)
 * 3 says material that could be in the article. 4 says stuff not suitable for the article. Please explain what you are not understanding since the differeence is plainly obvious, and both are vital to include right next to each other in bullet points.  As always language could possibly be more clear, but these are 180 degree opposite points being made. 2005 23:27, 2 November 2006 (UTC)
 * No. #3 does not say material that could be in the article.  #3 says material that should be integrated into the article if possible, but may not be possible due to other reasons, like copyright issues or unencyclopedic level of detail.  #4 says meaningful relevant content not suitable for inclusion.  Some examples are given (statistics, screen credits), but no explanation for what makes them unsuitable.  I have suggested that what makes them unsuitable is their un-encyclopedic level of detail, covered under #3.  As I said above, I would like to see an example of #4 that doesn't fall under #3.  I do think #3 could be better worded in order to avoid misreadings like your own, for example: "3. Sites that contain neutral and accurate material that cannot be integrated into the Wikipedia article due to a site's copyright issues, unencyclopedic level of detail, or other reasons." Schi 23:47, 2 November 2006 (UTC)


 * It says "Ideally this content should be integrated into the Wikipedia article" not "Ideally this content should not be integrated into the Wikipedia article". It seems your confusion over the clear language leads to your confusion over why they are distinct.  Instead of reading the words for obscure meaninings, just read what they say. 3 relates to level of detail.  4 relates miscellaneous lists, pov interviews and so on. They are completely different emphasis.  They could be combined in one paragraph of course, but they are talking about completely different things. 2005 09:15, 3 November 2006 (UTC)
 * That's the problem. These line's don't *explicitly* refer to their intent, or at least not the intent you describe. They are, as demonstrated, open to differing interpenetration. I suspect this is the fault of a drive to make these into "A Simple Bulleted List" and confusing terseness for simplicity.
 * I suggest that both of you provide text's that would describe what you belive the current guideline's intent is, and then we can see if there any differences that need to be reconciled. --Barberio 10:59, 3 November 2006 (UTC)

I'm not reading the words for obscure meanings, I'm just reading the whole guideline and what that says. It's true, #3 starts out saying you should link to information that should ideally be included in the article, but if you read past the first sentence, you should note that it says that you should link to information that should not be included in the article because of reasons of copyright or level of detail. The point is that if it should ideally be incorporated in the article, then why should we encourage (by saying such sites should be linked to) non-ideal behavior? This was also discussed here; also note how it's phrased in the current, non-workshop guideline - under "What to link": "An article can link to pages that contain neutral and accurate material not already in the article. When possible, the information on the linked page should eventually be brought into the Wikipedia article, and the link should then be moved into article's References section. Sometimes this isn't possible and the external link should remain, such as when the linked page contains copyrighted material which can not be brought into Wikipedia, or has a level of detail which isn't appropriate for Wikipedia (such as professional athlete statistics, movie or television credits, interview transcripts, or online textbooks)." I see that 2005 has added "pov interviews" in explaining what distinguishes #3 and #4. This is the only example I've seen that satisfies the criteria of #4 but not #3, and it turns on "neutral and accurate", which is what I originally inquired about. I suggest the following re-write: 3. Sites that contain neutral and accurate material that cannot be integrated into the Wikipedia article due to the site's copyright issues, un-encyclopedic level of detail, or other reasons. 4. Sites with other meaningful, relevant content that is not suitable for inclusion in an article, such as reviews and interviews. I'm concerned about #4 because it seems to leave a big hole for POV linking, so perhaps it should be amended with a note to keep in mind WP:NPOV. I'd prefer to strike the "material not already in the article. Ideally this content should be integrated into the Wikipedia article, in which case the link would remain as a reference." bit because it encourages lazy article-writing; if the site contains good material that should be in the article, then it should be in the article and the link should be kept as a reference. If the site contains neutral and accurate material that shouldn't be in the article (because of un-encyclopedic level of detail for example), then it should be included in the external links. In the case that someone just found a good link but doesn't have time to incorporate the new material in the article, I like Brian Kendig's suggestion here that the guideline encourage dropping the link on the talk page until it can be incorporated. Schi 18:10, 3 November 2006 (UTC)

Hijacked links
I think the "What to do with a dead link" section should expanded to cover the issue of hijacked links. These are links which may have been appropriate when added, but now are not appropriate to be linked to. This may be due to a simple "bait and switch" by the site owner, a domain that expired and was snapped up by a domain speculator, or some other reason.

Wikipedia's primitive tools wouldn't flag any of these, as they're not 404 or redirects. &mdash;Wrathchild (talk) 19:14, 1 November 2006 (UTC)


 * I've added text addressing this. --Barberio 10:46, 2 November 2006 (UTC)

Links normally to be avoided
I would recommend that we not start the "Links normally to be avoided" section with the following intro: "Unless there are specific, clear and overriding reasons that meet the criteria of what should be linked to, the following links are usually to be avoided:"

The reasons are twofold:
 * The heading "Links normally to be avoided" says everything that needs be said; in general, less text is better than more text.
 * "Unless there are specific, clear and overriding reasons that meet the criteria of What should be linked to" is redundant: every external link must meet those criteria. Worse, the "unless" implies that, so long as a link meets the What-should-be-linked-to criteria, it doesn't matter if it also falls into the Links-to-avoid criteria.  Thus, with this lead-in in place, if I have a site that "contain neutral and accurate material not already in the article", then I can add it even if I am adding it "to promote a site, that primarily exist to sell products or services..."

--Mwanner | Talk 19:26, 15 November 2006 (UTC)
 * I agree that less text is better, but in the past, I've found that it's useful to have a disclaimer like the contested lead-in to point to for conciseness. I don't care for redundancy either, I've just seen plenty of people remove official blogs/MySpace pages/etc. because they see blogs/MySpace included in the listed and don't read the whole page.
 * But you raise a good point about the neutral/accurate material criteria vs. promoting a site/other criteria for links normally to be avoided. At the moment, I can only think of one situation where the disclaimer is useful, which is when you want to include a link to an official blog/promotion site/factually inaccurate site etc.  So of the three criteria in "What should be linked to", I think only #1 can legitimately override the criteria in "Links normally to be avoided".
 * I suppose I would like to see language that makes explicit that "normally to be avoided" is just that, "normally", and is (always?) overrided by #1 of "What should be linked to", but I'm not married to the idea. I agree that the points you raise are sufficiently problematic to merit removing (or re-writing) the lead-in, so I'll remove it again for now until (and if) we come up with a better solution. Schi 19:56, 15 November 2006 (UTC)
 * OK. I have changed the blog/forum/etc entry under Links to Avoid to "Links to blogs, social networking sites (such as MySpace), video sharing sites, discussion forums or wikis, unless the linked page is the subject of the article or is an official page of the subject of the article.
 * I think that handles your concern? -- Mwanner | Talk 20:12, 15 November 2006 (UTC)
 * Don't you think "unless the linked page is the subject of the article or is an official page of the subject of the article" applies to every entry to "Links to avoid", not just the blogs/MySpace one? How about changing the lead-in to "Unless the linked site is the subject of the article or is an official page of the subject of the article, the following links are usually to be avoided:" ? Schi 22:58, 15 November 2006 (UTC)
 * The official site exception should apply to everything. The above line or something similar is a good idea. 2005 23:52, 15 November 2006 (UTC)
 * OK, I have added "Except for a link to a page that is the subject of the article or is an official page of the subject of the article, one should avoid:" as an intro, and deleted the phrase from the forums/blogs entry. -- Mwanner | Talk 02:37, 16 November 2006 (UTC)

Numbering points
A small beef, but can we change the (*) bullets back to (#) numbered bullets? It's easier to refer to one or the other point when there's a number attached to it. Schi 23:42, 15 November 2006 (UTC)
 * I would not object to that. I only changed it for consistancy. - Mike | Talk 23:46, 15 November 2006 (UTC)
 * I've changed it back, but I understand your concern about consistency. I left some of the bullets as * bullets, as I believe it was before your edit, but am open to someone else converting those to numbers if it makes sense. Schi 00:04, 16 November 2006 (UTC)
 * OK, I have changed it to the numbered format in the "important points to remember" section. Other than that, looks good. - Mike | Talk 00:16, 16 November 2006 (UTC)

Exclude wikis?
Discussion moved to Wikipedia talk:External links