Wikisource:Requests for comment/Wikilinking policy

From Wikisource
Jump to navigation Jump to search
Proposal to update the Wikisource:Wikilinks policy

The current policy document on Wikilinking was written before Wikidata was created and is out of step with some of our current practices. The proposed text below between the pairs of rules is a complete overhaul that has had input from several experienced English Wikisource editors.

Please add comments in the sections at the end of this RFC.

The proposed closing date for this RFC is when discussion has been dormant for more than 72 hours, or 23:59 UTC on 28 March, 2021, whichever is later.




This policy regulates how links should be used on English Wikisource in order to provide the best value to our readers, reusers, contributors, sister projects, and the Wikimedia movement as a whole. Its primary focus is on links in the texts we host and the user-visible namespaces, but also regulates other aspects of using links on the project and their use in project-internal namespaces. The linking policy defines what links are acceptable in our main texts, but some forms and kinds of linking that are not permitted under this policy may be acceptable under the policy for annotations as a special exception for those texts. Likewise, some links that would otherwise be permitted under this policy may be prohibited by other policies, such as by the copyright policy when the link is to an external site that violates copyright.


Summary[edit]

Wikilinks provide hyperlinks to other pages and sections within Wikisource and sister projects supported by the Wikimedia Foundation.

Light wikilinking inside works is permitted and recommended. Recommended links are links to author pages, titles of other works, and internal cross references within a single work. Permitted links that go outside of Wikisource are links to relevant sister projects such as Wiktionary, English Wikipedia, and Wikidata.[1]

Links that are forbidden include external links inside a hosted text, links that are interpretative, and links that for other reasons do not serve the reader or violate our principle of neutrality. A key concept in this policy is that there should be no surprises for the reader when they click a link.

Different rules apply for links inside a hosted text, in the header section of a page, a text's talk page, and in project-internal wikipages such as policy pages, discussion forums, and user pages.

Links in presentation namespaces[edit]

The use of wikilinks within the presentation namespaces (primarily the Main and Translation namespaces, but also the Page namespace by extension) is limited by policy.

There are two main types of wikilinks: those within the text itself, and those in the header template. See the documentation pages for the {{header}} and {{translation header}} templates for full details of what links are used there.

Basic wikilinks[edit]

Certain very basic wikilinks are permitted and, in fact, recommended:

  • Cross-references to pages or other sections within the work should be linked. For example, references to other articles in an encyclopaedia, references to other chapters in a book, and text such as "See footnote on page xx".
  • In journals cross-references to articles in previous issues should also be linked.
  • References to other works should be linked (whether we currently host them or not). If possible, do a deep link (see the section in the Style Guide), but at least link to the full title.
    This may require research as many common works are referred to in abbreviated form. When we hold multiple versions or editions of a work, if you are unable to determine which one was meant in the work, then a link to the versions page for the work is acceptable.
  • Links to mentioned authors and non-authors who meet the criteria to have a portal are always acceptable. Take care, though, not to overlink. Linking the first mention in a section (chapter or other subdivision) of the text is sufficient. Also, if the work is about a person and the related_author field in the header template has been used, then linking in the text as well is not necessary.

Unacceptable wikilinks[edit]

Links to the Index and Page namespaces and other project-internal pages should not be made within a text. In particular, even though a book's table of contents, list of illustrations, or similar constructs list a specific page number, these should never be linked to individual pages in the Page: namespace.[2] In transcluded texts, links to the relevant Index: and Page: namespace pages are available either by clicking the source button at the top of the page for the Index and the small page numbers in the left hand margin for the Page namespace. If a work mentions another work, the link should be to the main namespace rather than the Index or Page namespace. The same applies if a work mentions a section or passage of itself or another work. Anchors can be used where needed to create a suitable link target if a wikipage contains multiple independent sections.

Interpretative links are not acceptable.

External links and links to non-Wikimedia pages are not acceptable.

Links to other Wikimedia Projects[edit]

Occasional links to Wikipedia for clarity and to Wiktionary for obscure words are permitted. See the section on Context-appropriate links below for more detail.

Links to pages on other Wikimedia-projects are not permitted unless the original text included such links.[3] These include (but are not limited to):

  • files and file description pages
  • pages in the Help, MediaWiki, Module, Project, Special, Template and User namespaces
  • MediaWiki, Meta-Wiki, Wikimedia Commons, Wikimedia Labs, Wikibooks, Wikidata, Wikiversity, Wikispecies, and Wikivoyage pages

Files on Wikimedia Commons that reproduce an image from the work should be included inline in a manner similar to how it appeared in the original work rather than just linked to.[4]

In regard to Wikidata, direct links to individual data items are explicitly prohibited. See the Wikidata section for Wikidata's purpose in the project and the permitted ways to use Wikidata on English Wikisource.

Links in other namespaces[edit]

Index namespace[edit]

See Help:Index pages for more information.
  • Link the Title to the title that is used on the Main namespace page. If the work is transcluded in multiple volumes, link the volume field to the appropriate subpage.
  • Link the Author to the Author: namespace page for that author.
  • Link the Publisher to the Portal namespace page for that publisher, if one exists.
  • If there is an Illustrator, Editor, or Translator, link these to their Author pages.

Portal namespace[edit]

See Help:Portals for more information.

Use the authority control template to link to Wikidata.

The lists of works on a portal should all be linked (whether the work is yet created or not). If a work is not yet scan-backed, and there is an Index or File available, then use the template {{small scan link}}. If an Index or File is not available but a scan is known to exist in an online repository, use {{ext scan link}} to link to it. Once a File or Index page exists any {{ext scan link}} should be removed. Once a work has been fully proofread, the {{small scan link}} template should be removed.

Author namespace[edit]

See Help:Author pages more information.

Adding a link to the corresponding item on Wikidata automatically provides links in the header to the author's pages on Wikipedia, Commons, Wikidata, and Wikibooks. It also allows {{authority control}} to show all available authority control data for that author. Manual links to sister projects in the {{author}} template are therefore not needed and are not permitted by this policy.

Rather than using the notes field use the related_author field to make links to relevant other authors.

Works listed on an Author page should be linked, regardless of whether the target page exists yet or not. The exception is works that are known to not be freely licensed or in the public domain, which should not be linked. Co-authors, translators, editors, and illustrators that are mentioned should be linked.

Wikiproject pages[edit]

Links on these should be managed in a similar fashion to Portals (with the exception of authority control, which is not applicable). Wikiproject pages are explicitly project-internal pages and more leeway on linking is permitted subject to local consensus.

Talk pages[edit]

External links are explicitly permitted on Talk pages, but within the limits of other policies and guidelines such as policies on copyright and acceptable conduct. The {{textinfo}} template has a field for the source of the work, which will usually be offsite. Adding links to the source of biographic data on Author Talk pages is encouraged. Explanatory notes and links may be added to a work's Talk page—provided that such do not become general commentary on the work and its merits.

User and user talk pages[edit]

These namespaces are freer than most, with the limitation of advertising and other forms of self-promotional links. It is recommended that links that will identify the user are not posted on these pages.


Uses of links[edit]

Interpretative vs. non-interpretative links[edit]

A link's target should correspond to the term showing as the link as closely as possible given the context. Links should only be made to the most specific page appropriate to the context of the link.

The linked text and the link target should be essentially the same. No significant amount of interpretation is allowed in adding a wikilink to the body of a text. Doing so is considered to violate our policy of neutrality. As part of this, straight links are preferred whenever possible. Piped links should only be used to hide namespace and project names, to disambiguate, where alternative spelling has been used, different wordings, or other, similar, minor differences.

Example
Status What you type How it appears Notes
Allowed If the mad hatter of "[[Alice in Wonderland]]" had undertaken to write a history of the world, If the mad hatter of "Alice in Wonderland" had undertaken to write a history of the world, The title linked here is clearly marked and no interpretation is necessary in understanding its meaning.
Not allowed ...in Lewis Carroll's [[Alice in Wonderland|most popular work]]. ...in Lewis Carroll's most popular work. This link makes an assumption about the author's original intention. This should be left to the individual reader.

Wikilinks as annotations[edit]

Wikilinks within the body of a text, beyond those mentioned above, are generally considered to be annotations. This only relates to the presentation namespaces (i.e., the main and Translation namespaces) and those the feed into it (e.g., the Page namespace).

Overlinking[edit]

Excessive linking should be avoided. Even where it is allowable to add a wikilink under this policy, too many wikilinks can be distracting and should not be used if they are not necessary.

Generally, a link should only appear once on a single wikipage within the body of a given work. For example, if an author is mentioned by name multiple times within a short story, chapter, or other subdivision of a work, only the first instance should be wikilinked.

An author or other person who is already linked in the header should generally not also be linked in the text of the work.

An exception to this is where the work itself makes frequent or repeated references, in which case those references should still be linked.[5]

Underlinking[edit]

Underlinking is not a problem on Wikisource. It is normal for the body of a text to contain no wikilinks at all.

However, each page should have a {{header}} or {{translation header}} template and this should be as complete as possible. In particular, the header template should link to the author's page in the Author namespace. Subpages that represent chapters should link to the previous and next chapters in the work.

Titles of works and the names of authors within the body of the text should normally be wikilinked (to works in the main or Translation namespace and author pages in the Author namespace, respectively) but this is not required.

Unintended emphasis[edit]

Wikilinks that place an unintended emphasis on the linked term, to the extent that the work is altered and the term exoticised, should be avoided.

Wikilinks look slightly different than the surrounding, unlinked text. This draws the readers' attention and puts more emphasis on those words. In some situations, this emphasis potentially alters the work itself and can affect its reading. This can occur even when the meaning of the linked term and the link target are essentially the same.

For example, in Flight 93 Cockpit Transcript, wikilinking the phrase "Allah is greatest" to Wikipedia:Takbir places additional emphasis on those words, even though it is technically correct.

Context-appropriate links[edit]

The nature of the text may affect whether or not a wikilink is appropriate. This should take into account the author's presumed intentions, the likely audience of the text, and the use to which it may be put.

With academic or scientific works, more wikilinking may be appropriate as academics are more likely to use the text in that fashion. Heavier wikilinking may also be more appropriate in reference works, such as dictionaries and encyclopedias, as the links would enhance the usefulness of such as text.

With poetry or fiction, little or no wikilinking is more appropriate. Archaic or obscure words may be wikilinked to their definitions on Wiktionary to aid the reader. It is more acceptable to wikilink difficult words in children's fiction than in adult-orientated fiction due to the presumed younger readership.

With works of a similar type, the specific use of the work can make a difference. For example, general theological works, such as Tracts for the Times (1834), and biblical commentaries, such as A Critical and Exegetical Commentary on Judges (1910), both contain Bible passages and references that could be wikilinked to a copy of the Bible on Wikisource. However, a biblical commentary is meant to be used with a copy of the Bible already open; wikilinks are less important here as they are more redundant. On the other hand, it may be more appropriate to wikilink references in the general work as the reader is less likely to have a copy of the Bible on hand while reading.

Red links[edit]

If the target page of an internal wikilink does not exist, it will appear red like this.

Sometimes it is useful in editing texts and other pages to create a red link to indicate that a page will be created soon or that an page should be created eventually. Red links show works and authors that could, and should, exist on Wikisource but are currently not present.

A 2008 study relating to the sister project Wikipedia showed that red links helped that project grow.

Wikilinks to Wikipedia and other Wikimedia projects will never appear red. Such a link will appear in the same light-blue color whether the target page exists or not, and therefore inactive links are indistinguishable from active links. External links should be created only if the target page already exists, or if the target page will be created in short order.


Wikidata[edit]

See Wikisource:Wikidata for more information.

Wikidata is a sister site to Wikisource; it is a free, collaborative, multilingual, secondary database, collecting structured data to provide support for Wikimedia projects. Wikidata supports Wikisource with more easily maintainable interlanguage- and interwiki links, author images, and infoboxes, thus reducing the workload in Wikisource and increasing its quality. Links to sister projects (e.g. Wikipedia) and interlanguage links (e.g. French Wikisource) should be maintained through the Wikidata database, wherever possible.

The default item view on Wikidata is not user friendly or useful for most people, and for this reason direct wikilinks to Wikidata are not permitted in presentation namespaces. In some cases, however, it may be useful to identify a person or work for which a Wikidata item exists, but for which there is no suitable link target on Wikisource or the permitted sister projects. In these cases it is acceptable to link to Wikidata using the {{wdl}} template, which dynamically displays a link to the most suitable destination based on which targets are available.


Translations[edit]

See Wikisource:Translations for more details.

Wikisource translations are original translations of non-English works created by Wikisource contributors. For the purposes of this policy, Wikisource translations are just another kind of text even if they are hosted in the Translation: namespace. All parts of this policy that apply to texts in the main namespace apply equally to Wikisource translations.


Born-digital texts[edit]

Works that were originally published in a modern digital format are a special class of text on Wikisource. Such works will often include either web page addresses or have active links within the text, much like a wikipage has. For such works the original links in the form they appeared when published should generally be reproduced when hosted on Wikisource. Restrictions in this policy on acceptable links do not apply directly to these works, but restrictions from other policies, such as the copyright policy, that limit links may still apply.[6]

In works where the original includes live links, additional links should not be introduced even though they would otherwise be permissible under this policy. This is so that there will not be any confusion about which links were in the original and which were added by Wikisource.


Notes[edit]

Note: Footnotes in this page are informative and not normative. If there is a conflict between a footnote and the main text, the main text takes precedence. However, the footnotes are a part of the policy and the guidance and explanations contained in them are intended to inform the interpretation of the policy.

  1. Note that links to Wikidata must always use the {{wdl}} template. See the Wikidata section for details.
  2. For a table of contents the chapter title should be linked to the wikipage in mainspace where that chapter is (or will be) transcluded, and for a list of illustrations the equivalent part should be linked to an anchor where that illustration appears on the mainspace page where it is transcluded. Link targets at arbitrary positions within a page can be created with the {{anchor}} template.
  3. This is to allow for freely licensed modern texts that directly relate to Wikimedia sites or the Wikimedia movement. Examples include research papers and journal articles studying the projects, or books about the free culture movement.
  4. While inline images are not in themselves links, the displayed image is automatically linked to the File: page on Commons. These links as a side-effect of including the image are acceptable, but manually added links to files or file description pages are not.
  5. This applies when, for example, the same or a similar citation occurs in multiple footnotes on a single page or within a single section, chapter, or other subdivision in a text. It will typically not apply within running prose even if the text references the same author or similar multiple times, but might apply to something like a table which includes a column with citations similar to those usually found in footnotes.
  6. In particular, no page on Wikisource may link to a website that hosts copyright violations even if such a link was included in the work reproduced. This restriction applies to live (clickable) links, and not website addresses included as text.


See also[edit]




Support[edit]

Oppose[edit]

While I agree with almost everything in the suggested policy, I strongly disagree with permitting links to Wikipedia for the following reasons:

  • Links are distracting and drive readers away from the work and so should be limited only to advertise other content of Wikisource.
  • Currently it is not possible to distinguish visually internal Wikisource links and external links to other Wikimedia projects. Readers clicking a link on a person’s name usually anticipate that it would take them to an author page and so it should not take them anywhere else. They should not be unexpectedly taken away from Wikisource to another project. For some it may be annoying, others may not return back. They should be taken to Wikipedia only if they know that they are going to be taken to Wikipedia.
  • I absolutely disagree with Heavier wikilinking may also be more appropriate in reference works, such as dictionaries and encyclopedias. An example of heavy wikilinking in encyclopedias is e. g. Encyclopædia Britannica, Ninth Edition/Hudson River. Changing all the extensive internal links into Wikipedia external links would be in accordance with this suggested policy, but the result would be even worse than the current situation.
  • There are often lists of people in some articles and I sometimes meet the situation when some of the people are linked to their author page and others to Wikipedia articles, and you never know where you land prior to trying so. This is imo absolutely undesirable, and this suggested policy would explicitely allow it.
  • We often have person based or topic based portals which can be linked if some clarification is needed. What is more, these portals are interconnected with Wikipedia articles, if the readers need it.
  • If the readers really need to visit Wikipedia, they can quite easily do it without our assistance too.

For these reasons I suggest to set the Wikipedia links rule thus: "Links to Wikipedia are generally discouraged and should be used only very rarely in cases when modern readers may need clarification of the topic and when such clarification is really needed for understanding the work hosted in Wikisource. Creating a topic based or person based portal where possible and linking to it is encouraged instead. Mixing internal links and Wikipedia links in lists of items is explicitely forbidden." --Jan Kameníček (talk) 10:56, 14 March 2021 (UTC)[reply]

  • Oppose. Wikipedia might link here, as a source for one of the articles there. 'Adding value' to texts here is redundant at best, and subverts the best purpose for a library of transcriptions. CYGNIS INSIGNIS 17:53, 23 March 2021 (UTC)[reply]
I'll try to clarify (my head is swimming lately) that I am opposed to wikipedia, wiktionary and other ext links in the text, everything else proposed seems to be clear and broadly agreeable. CYGNIS INSIGNIS 03:46, 25 March 2021 (UTC)[reply]
Re: "Links are distracting and drive readers away from the work" – is there any evidence of this? Readers do not appear to be driven away from even heavily linked works on Wikipedia. I don't know why this would be any different with a moderate population of links here. BD2412 T 05:12, 25 March 2021 (UTC)[reply]
@BD2412: Are you quoting me, I don't recall where I said that. CYGNIS INSIGNIS 15:52, 25 March 2021 (UTC)[reply]
@Cygnis insignis: I was quoting the first bullet point from Jan Kameníček above. BD2412 T 16:03, 25 March 2021 (UTC)[reply]
Relevant xkcd? Wikipedia rabbit holes are a well-known phenomenon, and not conducive to linearly reading a text. Linking doesn't drive readers away per se, but may distract them. BethNaught (talk) 12:08, 25 March 2021 (UTC)[reply]
Curious what was searched on to find that xkcd gem, or was it bookmarked :) CYGNIS INSIGNIS 14:50, 25 March 2021 (UTC)[reply]
What may be true for fiction or poetry may not be true of say, the DNB. If someone comes into a DNB article about John Doe and then there is WP link in something like "he died in the Battle of ——" or if you are in a Dictionary of Quotations or Scientific Article or "Books in the —— Library" and there are 30 links on the page that break up linearly reading the text, that may be a good thing. We may not want to have the same thing, in say, the catalog of ships in the Iliad because that may be read linearly and distracting. MarkLSteadman (talk) 12:57, 25 March 2021 (UTC)[reply]
I would expect a fact like that to be in article at wp, it is the kind of thing I've greatly enjoyed: adding facts from the DNB and so on to wikipedia. CYGNIS INSIGNIS 14:50, 25 March 2021 (UTC)[reply]
If we don't allow WP directly from mainspace you would need to click author --> wp article author --> wp fact for every fact about a person (with no obvious way to do it for an article about say a regiment or ship without creating a portal) MarkLSteadman (talk) 16:51, 25 March 2021 (UTC)[reply]

Neutral[edit]

Comments[edit]

Extended meta discussion about decision process collapsed for readability. --Xover (talk) 10:18, 15 March 2021 (UTC)[reply]

Quote from Wikisource:Requests for comment: RfCs are not votes. This is only a request of comment not final voting, so I do not understand dividing the contributors to discussion into supporters and opposers. There are points which I agree with and there are points which I strongly disagree with and I want to discuss them. Sections "Support" and "Oppose" which are typical for voting, not discussions, need to be removed, it is too early to express support for something prior to hearing other opinions. --Jan Kameníček (talk) 10:15, 14 March 2021 (UTC)[reply]

@Jan.Kamenicek: RfCs are not votes. means that they are not decided by mere majority vote (i.e. mere numbers alone are not determinative, though overwhelming numbers or mere majority without strong opposing arguments usually determine the outcome). That's why we tend to call them !votes: they're not strictly votes, but function very similarly. The long and short of it is, if you have comments, reservations, or arguments pro or con, this section (Comments) is where to post them. Note that you can express your vote above tentatively above, preferably outlining your main reason for that stance, and then change it later (typically as a result of discussions here in the comments section). --Xover (talk) 10:25, 14 March 2021 (UTC)[reply]
People should not be encouraged to assume supporting or opposing position before hearing all opinions. That should follow only after discussing everything. --Jan Kameníček (talk) 10:34, 14 March 2021 (UTC)[reply]
@Jan.Kamenicek: They support or oppose based on the proposal, in this case the proposed policy text. If arguments are presented that they find persuasive they can (and should) change their position (signalled by the !vote above). Usually this is people who support based on the proposal, but are then presented with an argument for why that policy is a bad idea; or they oppose it but are then persuaded by subsequent arguments. So far no such arguments have been presented, so there is no basis for anyone to change their minds. --Xover (talk) 10:44, 14 March 2021 (UTC)[reply]
@Jan.Kamenicek: !votes with a brief summary rationale (e.g. "Oppose, see comments below") in the Support/Oppose sections above; extended arguments and discussion here in the comments section. :) --Xover (talk) 11:03, 14 March 2021 (UTC)[reply]
I prefer not to divide my disaproval into two sections. It is not in accordance with the current practice, see other open RfCs. --Jan Kameníček (talk) 12:02, 14 March 2021 (UTC)[reply]
I will probably later suggest a change in the practice of accepting new policies, but suggesting a long and elaborate policy followed by immediate voting, though with parallel discussions, expecting people to keep changing their votes (possibly also several times) as new arguments come, is imo bad practice. There should be a phase of mere discussing and contemplating of pros and against and only after this phase when all arguments are on the table the voting should began. But it is not a story of this RfC only. --Jan Kameníček (talk) 12:23, 14 March 2021 (UTC)[reply]
@Jan.Kamenicek: If it helps you can feel free to blame me for the lack of a period of open discussion before making the proposal. I had the opportunity to insist on that and didn't (because I didn't think it was necessary in this case: it's an update to WS:Links, not a completely new policy, and the changes are to address the community's previously expressed concerns with the old policy). The only actual problem with that is that it may lead to people opposing the proposal over issues that could otherwise have been addressed ahead of time. That may mean we have to go back to revise it and then make a new proposal. But the one issue you specifically object to is not one such: whether or not to allow links to Wikipedia is one of the primary things that this RFC will determine. It is not an objection that could have been addressed ahead of time through editing. It is however something which it would be easy to change after the fact (if the community supports that), if one were to support the other improvements to the policy in spite of disagreeing with this single point.
Not to put too fine a point on it, right now you're just coming across as hostile to the very effort of improving our policies and addressing the concerns expressed by the community on WS:S, by dragging your feet on procedural grounds even though you so far have only articulated a single issue with which you actually disagree. I get that you feel that it has been sprung on you without warning, and that you feel strongly about the issue of links to sister projects; but the proposer has made a good faith effort to improve the project and address the concerns of the community, and should be applauded for that regardless of any individual points of disagreement. Please keep in mind that other members of the community feel just as strongly about this issue as you do, only in the opposite direction.
No, the other RFCs currently open do not employ this format, which is why they are lingering open for months and yet further months after the last substantial comment. Speaking as someone that spends way more than my share of time on closing various community discussions and assessing consensus (not counting votes), that format is absolutely hopeless both for the poor schmuck trying to do the close and for general community members trying to wrap their heads around the issue and decide where they fall. That we have poorly structured discussions elsewhere is not a good argument that all discussions should be poorly structured. So I'll ask once more, please keep the discussion down here and leave the Support/Oppose sections for votes and brief rationales. I'd like to address some of the concerns you have expressed but I am, for the reasons just articulated, not going to do it in the Support/Oppose sections above. --Xover (talk) 15:08, 14 March 2021 (UTC)[reply]
I am sorry if I sound hostile, it was not my intention. However, I really think that this way is a very bad one and the reason is not that there is a single issue in this particular proposal which I disagree with. I disagree with this way generally for all proposals of policies, because it encourages people to vote before all arguments were given. I do not think we can rely on that everybody will keep changing their vote as new arguments come. Votes should always come after arguments, not vice versa. Psychologically, once somebody has expressed their support or opposition, it is always more difficult to convince them to change their opinion than convincing them before they made the support/opposition statement. But let’s imagine that I bring an argument that overcomes this and they cross out their initial vote and vote the other way, and then somebody brings another convincing argument against mine: the chance that people would cross out their crosses and return to their previous vote is minimal, there is a big danger that the fear of making oneself ridiculous for changing opinions too often is bigger. So discussion and laying arguments on the table has to precede voting, it cannot be done before or parallelly. --Jan Kameníček (talk) 17:52, 14 March 2021 (UTC)[reply]


Ad Wikidata[edit]

The fact that it is forbiddent to link directly to Wikidata items from the transcribed texts should be reflected also in the usage of the template {{wdl}}, discussed in the section #Wikidata. I suggest limiting its usage only to "notes" in the header of a page in the main namespace and to portal namespace (and possibly author namespace???), besides the meta namespaces like Wikisource:, Talk:, User talk: etc. --Jan Kameníček (talk) 14:10, 14 March 2021 (UTC)[reply]

The proposal as it stands permits links to Wikidata but requires them to be made through {{wdl}} instead of directly using [[d:Q…]] syntax. The #Wikidata section reflects this. --Xover (talk) 07:07, 15 March 2021 (UTC)[reply]

Withdrawn suggestion about use of {{wdl}}. Inductiveloadtalk/contribs 18:16, 15 March 2021 (UTC)[reply]

Withdrawn I suggest expanding the recommendation for {{wdl}} to include all non-WS WMF links: where a link using {{wdl}} points away from Wikisource (e.g. an author that has no Author page, but does that a Wikipedia or Wikidata page), it is encouraged to create the Wikisource page. Links to authors, works and subjects that point off-wiki should always use {{wdl}} so that they will automatically resolve to a Wikisource link upon creation of a suitable page, without further maintenance.

The idea being that off-wiki links are only done because we have no better target, but we would link to an author/etc if we had a page to link to. {{wdl}} will auto-resolve to the corresponding Wikisource Author/Portal page as soon as it exists.

Linking directly to Wikidata, or Wikipedia for that matter, with the [[d/w:]] syntax is a pain because it's quite a task to scan for "links that could now point to Wikisource". But {{wdl}} does it seamlessly. Inductiveloadtalk/contribs 15:40, 14 March 2021 (UTC)[reply]

@Inductiveload: That works well for things for which Wikisource can eventually have a page, but not for other things (places, concepts, scientific principles, historical eras, whatever...) or even the links to Wiktionary that are permitted under the current policy. If we are to permit links to Wikipedia at all I think going by way of {{wdl}} is needless complexity, so I would at most be in favour of scoping it to apply to links for things that are within scope for Wikisource. If we do not end up permitting Wikipedia links the issue is obviously moot. --Xover (talk) 16:16, 14 March 2021 (UTC)[reply]
@Xover: Fair enough. Perhaps ... Links to authors, works and subjects that point off-wiki should always use {{wdl}} if a valid Wikisource page could exist, but currently does not, so that.... Probably with a footnote like For example, Qian Yiji should use {{wdl}}, as he could have an Author page in future, but w:li (unit) can link directly to Wikipedia with [[w:li (unit)]]. (feel free to substitute "li" with any other link that you think is unsuitable for a Portal, but still merits a WP link.)
As an aside, many of "places, concepts, scientific principles, historical era" can validly have portals: Portal:London, Portal:Rationalism, Portal:Relativity and Portal:Dutch America all seem cromulent to me. Obviously, there is a limit to what we can sensibly have portals for (w:Vela X-1 will probably never get its own Portal unless something dramatic happens like that's where aliens come from), but the "big boys" can at least plausibly be imagined to accrete Portals one day. Until that happens, {{wdl}} and [[w:...]] are equivalent, but the day it does, one auto-updates to an internal link, and one will point to enWP until it is noticed and fixed. AFAIK, there's no way to generate a list of links to Wikipedia which could be to Portals without trawling every single page that contains a WP link, so there isn't even a simple automatic way to chuck it on a backlog. Inductiveloadtalk/contribs 22:09, 14 March 2021 (UTC)[reply]
Sure. But, 1) do we really want to encourage mirroring every (5M+) Wikipedia articles as Portals that, let's be real, are not going to actually explain the concept (Portals are essentially categories in this context, not encyclopedia articles, and often poorly maintained to boot), and 2) do we really want to overload everything into {{wdl}}? To me this is needless complexity for both contributors and readers to very little benefit. Are we also going to force external links to open in a new tab or in a frame in order to keep people on Wikisource? Why do we not force Wiktionary through this mechanism then?
This smacks of wanting to keep Wikipedia at arms length because they feel like a "threat", but Wiktionary is ok because, frankly, nobody considers that project a threat. Authors and works are our bread and butter, spiced up with the occasional publisher portal, and everything else is an exception. We have some great portals but they are not ipso facto good link targets from inside our hosted works, and our portals are not generally designed for that purpose (you have yourself complained about their unclear role). The reason the proposal forces Wikidata links to go through {{wdl}} is because the Wikidata item view (or Reasonator for that matter) is a really crappy and confusing destination for 99% of our readers (and 98% of our contributors), not that it wasn't invented here. Those reasons do not apply to Wikitionary and not to Wikipedia.
What I'm saying is, if we're going to expand the {{wdl}} mandate beyond Wikidata (which is all the current proposal does), we need to split a lot of hairs on this phrasing to get the balance exactly right; and I don't really see that the benefits are anywhere near meriting that effort or the resultant complexity. On the flip side, nothing in the proposal prevents you from using {{wdl}} for other links if the instance at hand is one of the obvious cases where we might eventually have a good target locally. How about delegating this kind of detailed guidance to a help page instead? --Xover (talk) 07:03, 15 March 2021 (UTC)[reply]
@Xover: I think that is eminently sensible. On reflection this is too fine a point for policy. If we end up with endemic major overlinking or too much of a chilling effect on sensible linking (both fairly unlikely as written, IMO) we can revisit from a more secure footing. Inductiveloadtalk/contribs 16:20, 15 March 2021 (UTC)[reply]

Wording quibble: I think the mention of Wikidata in "Links to other Wikimedia Projects" is a bit confusing. At first, they're included in the prohibited list under Links to pages on other Wikimedia-projects are not permitted, then apparently specifically banned again (the meaning of "direct" isn't clear here): In regard to Wikidata, direct links to individual data items are explicitly prohibited, and finally in a different section, [in some cases] it is acceptable to link to Wikidata using the {{wdl}} template (finally, the "direct" meaning becomes a bit clearer, though it's still oblique).

Perhaps Wikidata should not be included in the global prohibition (like WP and Wiktionary are not). Then, add a parenthetical like "direct links to Wikidata (i.e. of the form [[d:Qxxx]])" to make it clear up-front what "direct" means and that there is such a thing as "indirect", which will be handled by the Wikidata section.

I don't think this changes the proposed policy, just makes clearer that it's direct links that are prohibited (and define what that means), and then say "but there are still rules for indirect links". Inductiveloadtalk/contribs 18:15, 15 March 2021 (UTC)[reply]

@Inductiveload: I don't agree there's any actual ambiguity in the text as it stands, but agree that this both could and should be made clearer for the benefit of the poor schmucks trying to read this going forward. But as it won't change the meaning I think that's acceptable copy-editing we can do once the policy is adopted. Changing the proposal once voting is under way is… not optimal. --Xover (talk) 07:37, 23 March 2021 (UTC)[reply]

Ad comments on Wikipedia links[edit]

@Jan.Kamenicek: Responding here to keep the Support/Oppose/Neutral sections manageable.

The main point of your comments appear to be a philosophical objection to linking to Wikipedia at all. I did not anticipate that (concerns about possible downsides etc., yes; but not flat out opposition to it), but that’s on me. The position is certainly a valid one and not one which I believe discussion is likely sway, which somewhat makes this an “all or nothing” issue: either the community wants to permit such links, or they do not, and those of the opposite opinion to the community’s decision will essentially just have to live with it.

However, some of the additional points you adduce are not entirely correct:

  • Interwiki links have the CSS class .iwlink and can be styled to have a different appearance than local wikilinks (subtly or drastically), either per user, in a Gadget, or globally. I don’t personally think that’s needed, but neither do I have any objections to it in principle if the community wants that. It is even technically possible to completely remove link formatting and disable active links for individual users if some outcome of discussions made that necessary.
  • Encyclopædia Britannica, Ninth Edition/Hudson River is the very definition of overlinked, but it was bad practice under the old policy and is forbidden under the new policy. However, overlinking is an inherently subjective judgement so a policy can’t regulate in detail just where the line is: it can only give the general principle and regulating it in detail has to be up to community discussions and practice. If anybody wants to try to codify this in more detail (as a help page or as an expansion of the proposed section in the policy) I’d generally support that.

Regarding mixing destinations for links in lists or tables, I’m not convinced that’s a major problem or one that needs policy qua policy to regulate. I wouldn’t be opposed to putting some recommendations to avoid it in the style guide, but based on my current experience I don’t think it’ll happen very often and I don’t expect it to be a big problem when it does. But that assessment will likely be heavily influenced by one’s philosophical stance on interwiki links in general so opinions will surely differ on that point.

Regarding Links to Wikipedia are generally discouraged and should be used only very rarely …, that is, to me, just a difference in degree to what the proposal says (but one that tips it too far). The overlinking section and other parts limit going overboard (and can be tightened further if Hudson River becomes a problem in new texts) and nothing in the proposal actively encourages or requires excessive linking: it just permits links, with tight limitations, subject to local consensus and community practice. The distinction between actively encouraged (should) internal links (works, authors, etc.) and merely allowed (permitted) interwiki links is deliberate. The actual phrasing in the proposal is: Occasional links to Wikipedia for clarity and to Wiktionary for obscure words are permitted. [emphasis added].

In the decade+ that we’ve had a links policy that permitted Wiktionary links there hasn’t been an issue with overdoing it, and I spend a lot of time trawling through old texts so I feel confident I would have run across at least some such if they existed. In fact, across ~850k content pages (~3.4M total pages), we have a grand total of ~5.5k links to Wiktionary. That’s something like ~0.65% even if each link is on a different page, and much lower if we assume that multiple Wiktionary links appear on each page. --Xover (talk) 10:16, 15 March 2021 (UTC)[reply]

As for visual distinguishing between internal and external links, I can see two possibilitites, both of which are bad: 1) Allow external wikilinks without visual distinguishing and the readers will not know whether the linked name they will send them to a WS author page (as usually expected) or they will be driven away to another project. 2) Distinguish them visually and thus make the links even more disturbing in the text. My preference is to avoid both problems by avoiding external links, it should be simply clear that links from names go to author pages and if readers want to check the name in Wikipedia, they can do it without the link too.
As for mixing destinations for links in lists or tables: I saw such mixtures several times, an example is a page with the US Declaration of Independence, containing list with 7 links to author pages and the rest to Wikipedia, which is very confusing. However, it is true that if the policy explicitely made usage of external wikilinks rare, additional ban of such mixtures would be redundant.
I agree that the Hudson river is an example of clear overlinking, but the new policy explicitely encourages “heavier linking” in reference works. This should be completely omitted from the policy, as authors of reference works often tell us how heavily linked they want their work using various [q. v.] hints and there is no need to make it heavier with more external links to Wikipedia.
One more argument: Although printed books did not have a possibility of hyperlinks, their authors used other ways of referring to other works if they wanted, e. g. by footnotes with messages like "For details see an excellent article in Enc. Brit. …" But if they chose not to refer to anything in their work, we should not be adding any explanative references (in the form of links to encyclopedia articles) either. It is quite enough that we add there our internal links, why adding more external links too? Not speaking about the fact that many WP articles are too far from excellent and I am convinced that many authors would have never chosen some of the current WP texts to be referred to in their work (many current authors, especially academics, do not do it either, although they do have such a possibility). Current Encyclopedia Britannica does not refer to Wikipedia, why should we add such links to their 1911 edition? I believe that original EB11 authors would never have such intentions either.
I do not want to forbid readers seeking further information in WP if they want to, I just think that adding such links to original works is redundant and adds there something which was not there. If somebody wants to go to WP, they do not need our specific encouragement. --Jan Kameníček (talk) 12:30, 15 March 2021 (UTC)[reply]
One thing that WP links allow is linking of things which are substantially less obvious now that they might have been at the time. A common thing I run into is old and non-standard transliterations of Chinese, for example “Hoo Quang”, which is a now-defunct province that is transliterated “Huguang” today. Or “Yaou-jin”, which is now written “Yao ren” and means w:Yao people. Figuring out "ℨû-lâi” meant “Culai (徂徕)" was a major pain. Sometime, the modern equivalent is no obvious at all and takes a bit of digging to work out. I occasionally use WP links for these (when I can be bothered), because a tooltip can’t be used to select the text. Adding the modern equivalent or the Chinese text has previously been described as annotation: cf. the end of Wikisource:Scriptorium/Archives/2019-09#Wikilinks, so I have been rather gun-shy about it. Inductiveloadtalk/contribs 13:05, 15 March 2021 (UTC)[reply]
Agree… That is why I suggested that WP links …should be used only very rarely in cases when modern readers may need clarification of the topic and when such clarification is really needed for understanding the work hosted in Wikisource. … The wording might not be as comprehensible as I wished and may need improvement. --Jan Kameníček (talk) 13:54, 15 March 2021 (UTC)[reply]
@Inductiveload:I can see you have already voted for the original proposal, but I think that the point you raised is covered by my suggestion. --Jan Kameníček (talk) 14:11, 15 March 2021 (UTC)[reply]
@Jan.Kamenicek: The phrasing in the proposal, much like the identical phrasing in the original policy, and which is given in the context of cautioning against too many links and specifying that judgement must be employed before adding a link, is With academic or scientific works, more wikilinking may be appropriate … followed by With poetry or fiction, little or no wikilinking is more appropriate.. That is, in a scientific work that frequently makes references to other works and authors it may be more appropriate to link these, than it is in poetry and fiction where it is almost never appropriate to add a link. It in no way shape or form suggests “Link every word to Wikipedia!”—in fact it has nothing to do with Wikipedia—but rather it talks about and tries to prevent overlinking (no matter the destination of the link). It is a limitation on linking, not an expansion. --Xover (talk) 14:57, 15 March 2021 (UTC)[reply]
@Xover: I see, thanks for explanation. However, the section #Links to other Wikimedia Projects says that details connected with linking to other Wikimedia projects are written in the section #Context-appropriate links, and so it gives impression that everything written under this header applies to linking to other Wikimedia projects including Wikipedia too. If it really has nothing to do with Wikipedia, would you agree with removing the reference from the first section to the other one and could the section Context-appropriate links mention that it applies only to internal links? Thus it would be much clearer, what do you think? In fact, if we adopted my suggestion written above, no more specifications for Wikipedia links would be needed. --Jan Kameníček (talk) 16:37, 15 March 2021 (UTC)[reply]
@Jan.Kamenicek: My apologies: I have not expressed myself clearly. The section #Context-appropriate links applies to all links, including links to other Wikimedia projects such as Wikipedia (and Wiktionary). What I was trying to say was that the section doesn’t give you permission to add more links. Other sections say what types of links are allowed, and then #Context-appropriate links places limits on how many links and how to link. If the outcome of the vote is that links to Wikipedia should not be permitted, the #Context-appropriate links section will still apply to all the other permitted link types. --Xover (talk) 17:46, 15 March 2021 (UTC)[reply]
I see. So it is exactly what I was afraid of. I admit that links to Wikipedia may make sense in rare cases when modern readers do not understand a term which once used to be clear to everybody, and so the word “rare linking”, not “heavier linking”, should be repeated everywhere throughout the rule. The gates are going to be open too wide if this gets through… --Jan Kameníček (talk) 18:16, 15 March 2021 (UTC)[reply]
  • Frankly, I don’t see why we would not want more extensive links to Wikipedia, to the extent that works contained here mention people, places, and concepts that are covered in Wikipedia articles. For example, our current POTM, Women of the West, collects biographies of noted women from Western states, some of whom would presumably have Wikipedia articles (which might offer a different perspective on their biographies). They all detail places where the subjects lived and organizations to which they belonged. What is the downside to linking them? BD2412 T 03:21, 20 March 2021 (UTC)[reply]
    The downside is that we would be adding what the author did not add to the work. For more reasons see here and here. Extensive linking to WP articles about people, places and concepts that you are considering should be prevented for reasons I have mentioned above, not supported. --Jan Kameníček (talk) 09:01, 20 March 2021 (UTC)[reply]
    Hosting the works on this platform at all is already adding things the author did not add to the work. BD2412 T 15:34, 20 March 2021 (UTC)[reply]
    @BD2412: What do you mean? CYGNIS INSIGNIS 07:03, 23 March 2021 (UTC)[reply]
    I think it’s pretty obvious. If we wanted to, we could just present the original page images like you would see on Google Books or the Internet Archives collection of scanned books. We don’t. We transcribe the works and present them in our own formatting, presumably to enhance the readability experience for our readers. At the top of every page, we bother to link the author’s name to their author page because it may be of some informational benefit for readers to be able to click over to that. We close words split across lines or pages in the original work; we mark obvious typos with {{SIC}}; we do allow links for sufficiently obscure terms, for which the cutoff will always be an arbitrary line. BD2412 T 15:33, 23 March 2021 (UTC)[reply]
    That bears little resemblance to my conception of producing type facsimiles, so what do you mean when you present your opinions couched in “our” and “we”? CYGNIS INSIGNIS 16:31, 23 March 2021 (UTC)[reply]
    This project, as a whole, is already changing the format and presentation of texts to suit readability and facilitate informing readers beyond the limits of the original texts. Everyone who participates in this project can plainly see that, and is therefore assenting to participating in that. BD2412 T 17:24, 23 March 2021 (UTC)[reply]
    The discussion is over? CYGNIS INSIGNIS 17:40, 23 March 2021 (UTC)[reply]
    Well, we do make some changes in the format and way of presentation, but not in the contents, and these changes are often necessary if we want to make the text machine readable etc. Adding explanatory links to WP articles is a different level, and it is absolutely unnecessary, as anybody can visit external encyclopedias of their choice without our assistance too. But it is only one of many arguments that I raised, anyway. --Jan Kameníček (talk) 18:56, 23 March 2021 (UTC)[reply]

I would not want to see us begin adding lots of links in the body text of works of literature. For example, linking lines of Shakespeare to Wikipedia articles in the body of his pays, except on an Annotated version created for that purpose. The body text of literature should, in my opinion, be largely if not entirely devoid of links. (1) When reading a work of fiction, or a play, the links are distracting to many readers. (2) Links in the body of the text can cause problems for people scrolling on a touch screen, as they will open to the target when the reader touches them and take them out of the work they are trying to read.

But there are works and parts of works that benefit greatly from informative linking:

  • Ex.1: The Appendices to volumes in the Yale Shakespeare series, e.g. Henry V (1918) Yale/Appendix B which covers the history behind Shakespeare’s play. Linking to further information about the sources and events adds significant value to the reader making use of the Appendix.
  • Ex.2: Longfellow’s notes to Dante’s Divine Comedy, which gives information about the referents, individuals, quotations, and events that appear throughout the work. The notes are specifically intended to provide more information, and linking key persons, events, etc. will add considerable value to anyone making use of those Notes.

There are quite likely many other examples like these, but my time at the moment is short, so I will stop here for now. --EncycloPetey (talk) 16:51, 25 March 2021 (UTC)[reply]

@EncycloPetey: I do not think that it is our task to add more informative value to works than their authors gave them. Longfellow said what he wanted to say, we should not be adding more. If somebody needs to consult more sources, they can do it and they can choose an encyclopedia they want, why should we prescribe them Wikipedia? We also should not be linking the old works to modern ones which can be written from such a perspective, which the original author may but also may not agree with. Current academic authors have the possibility to link their electronic books to Wikipedia but almost never do it, so why should we do it for Longfellow? --Jan Kameníček (talk) 17:11, 25 March 2021 (UTC)[reply]
It sounds as though you have not actually looked at Longfellow’s notes. Many of his notes explicitly direct the reader to another source for more information: “See Troya, Del Veltro Allegorico di Dante". Linking to such sources makes sense. What Longfellow is saying in his notes is “go look here”. Why should we not therefore link to outside sources? Other notes have no context whatsoever, and assume the reader has studied the relevant literature; vide his note on Canto I, line 22, which reads simply “Beatrice”. Linking to Wikipedia’s article about the relevant Beatrice in that instance will allow the reader to know what is meant. And to answer your question regarding current academic authors: Those authors are writing a new text and have complete control over their content. Instead of linking to an outside source, they can incorporate as much additional information into their text as they wish. On Wikisource, we cannot amend the texts in that way, but must present the text as it was published. The two situations are not comparable, since one has the ability to rewrite the text to include additional information, and the other cannot do so but has to rely on linking or annotation to provide additional context. --EncycloPetey (talk) 19:03, 25 March 2021 (UTC)[reply]
Exactly, Longfellow chose his sources to refer to, and he chose such that corresponded to what he wanted to say. We should not add ours which may not correspond to what he wanted to say at all.
I was not comparing work of academics with ours. I wanted to say that modern academics avoid linking to Wikipedia and so it is reasonable to assume that old authors would not choose such a source either, and so we should not do it instead of them and should let readers choose their sources. --Jan Kameníček (talk) 21:13, 25 March 2021 (UTC)[reply]
I read that argument and interpret it as: 19th-century authors didn't choose to wikilink, so we shouldn't either. Works published in books had limitations that the web does not. Any reader can still see exactly what Longfellow wrote, but an additional century of scholarship and research provides information that was not available. Forcing readers to do things the 19th-century way on a website in the 21st century is a non-argument for me. --EncycloPetey (talk) 22:13, 25 March 2021 (UTC)[reply]
No, you did not read well. I was not writing about wikilinking in the 19th c., I was writing about referring to other works. We use (among others) wikilinks to refer to other works, they used other means and we should keep only what they referred to and not refer to anything else. Especially we should not refer to modern works which interpret many things differently than the works which the original authors really chose to refer to. --Jan Kameníček (talk) 22:43, 25 March 2021 (UTC)[reply]

Red links on Author and Portal pages[edit]

"Works listed on an Author page should be linked, regardless of whether the target page exists yet or not."
"The lists of works on a portal should all be linked (whether the work is yet created or not)."

I think that red links on these pages should not be explicitly recommended. The risk is (and I have seen this actually happen, though I can't recall an example right now) that somebody transcribes a different work of the same name and doesn't check Special:WhatLinksHere: then the Author/Portal page links to the wrong work, so readers get confused and contributors who are interested in transcribing the work may assume it has already been done and go away again.

However, editors who know what they are doing should not be prevented from adding such links if they consider it appropriate (for example, the work is under construction in the Index: namespace). BethNaught (talk) 16:26, 21 March 2021 (UTC)[reply]

@BethNaught: Hmm. I see the risk, but I don't really see how omitting the links prevents it. The targets of links can be moved or replaced for all sorts of reasons and at any point in the link's lifetime. If the one doing the moving etc. doesn't check incoming links then we'll get that problem anyway. I don't think we can prevent it from happening, and I don't think it's a huge problem when it does happen. But we should perhaps develop some guidance on how to construct such links so that we don't have a ton of extremely generic titles linked from author pages to the same target. --Xover (talk) 07:32, 23 March 2021 (UTC)[reply]
For an Author page, it can make sense to have a redlink, especially if the title's target page is being coordinated and/or disambiguated across several pages, such as a title that is part of a series, or which shares a name with the work of another author. But I see no point in having red links on a Portal page, since a Portal is meant to organize what we have. --EncycloPetey (talk) 16:39, 25 March 2021 (UTC)[reply]

Scan links on Author pages[edit]

The "Portal namespace" section has useful recommendations on using {{small scan link}} and {{ext scan link}}. I believe these should also apply to the Author namespace. It would provide a useful resource to readers as well as helping newbies find texts to contribute with a lower barrier to entry (because they can just transcribe, without needing to find and upload scans). Indeed, this is how I got hooked on Wikisource: I found on Author:David Herbert Lawrence that an index for The Rainbow had already been created. BethNaught (talk) 16:26, 21 March 2021 (UTC)[reply]

I agree that this should be harmonised between the Portal and Author sections. Inductiveloadtalk/contribs 11:02, 22 March 2021 (UTC)[reply]
@BethNaught: I believe these are not mentioned for Author: because they are assumed by default; and called out on Portal: because portals should allow much of the same as author pages. In other words, their omission from the section on Author: pages is accidental. --Xover (talk) 07:26, 23 March 2021 (UTC)[reply]

"Essentially the Same"[edit]

I am unclear what this means in the following context: "As an author once wrote: 'A single man in possession of a good fortune, must be in want of a wife'". Is this meant to both prohibit doing the following: "As an author once wrote: 'A single man in possession of a good fortune, must be in want of a wife.'" and to encourage creating anchor to the actual text to link to for the actual text "As a female author once wrote: 'A single man in possession of a good fortune, must be in want of a wife'" (or adding an anchor to the specific line...)? Basically a. is linking "author", "critic", "book", "wrote" disambugation or violate "essentially the same? I can see how something like "Jane's Austen 1813 romantic novel of manners with Elizabeth Bennet and Fitzwilliam Darcy" and "Pride and Prejudice" are essentially the same in the sense they refer to the same thing unambigously but also not since they way they refer to it is different and I am not sure which way is meant b. how does this cover content within a work rather than titles which generally are not "essentially the same"? MarkLSteadman (talk) 23:15, 22 March 2021 (UTC)[reply]

@MarkLSteadman: Of all your examples here, I would say the only unequivocally acceptable link would be Jane Austen. The "… 1813 romantic novel of manners …" is prohibited on its face by the policy, but I would say this sort of case is within the bounds of what the community could choose to permit on a case by case basis (or it could choose not to: it would take discussion to find out). The point of the policy is to prevent excessive interpretation: there are contexts where things other than author's names and titles of works can be linked, but that should happen sparingly and only when the specific context is such that the amount of interpretation is below an acceptable amount. And where exactly that line is will always ultimately have to be decided case by case through community consensus.
In thinking about this, keep in mind that when a given work doesn't give you obvious hooks like author names and titles that you can hang a link on, that is often a deliberate style choice (either by the author or by the fashion of the time period). This could, for example, be a form of elitism ("if you don't get my oblique reference then this book isn't for you") or a way to signal belonging to a group ("See how good I am at making references that only my in-group will understand"). However one may feel about such writing, in terms of this policy it's a red flag when you start fighting authorial intent. Not automatically and unequivocally forbidden, but you should definitely be prepared to be challenged on any links by other members of the community. --Xover (talk) 07:22, 23 March 2021 (UTC)[reply]
That is why I wanted to comment since the "Lewis Caroll's most popular work" example seemed to focus author intent and interpretation on ambiguity in the signified ("most popular work", maybe it is Alice, maybe Looking-Glass?, that is what is being interpreted, "essentially the same" don't link characters to their books even if sharing the same name: "she was a model for Lewis Carroll's Alice" is not ok but "Lewis Carroll's Alice was her favorite book" is ok) rather than intent in choice of signifier (using "Alice" vs. "1865 work" vs. "his first book with Macmillan" etc.) and whether these were meant to be covered a strong explicit policy statement ("essentially the same", unless it explicitly says "Pride" no link to "Pride and Prejudice") or something to be worked out by community discussion ("significant interpretation", "author intent") MarkLSteadman (talk) 09:23, 23 March 2021 (UTC)[reply]

I think this simply is an example of the policy wording being ambiguous between "topic" and "word choice" (as you say, signified and signifier). I'd propose changing the wording to something like The linked text and the link target should refer to essentially the same topic. Thus, "Alice-the-book" and "Alice-the-character" are not the same thing, while (to crib from below) "Martha Washington" and "George Washington's wife" are the same thing (and IMO, both would be a valid Wikilink to Author:Martha Washington. Stretching to find a place to squeeze in a Wikilink is another matter, IMO.

However, I think as Xover says elsewhere, that perhaps this can be a follow-on discussion if/when this discussion resolves. Regardless of the outcome of the current RFC discussion, the "essentially the same" wording isn't part of the proposed changes, so this section will be relevant for follow-up even if this RFC fails. Inductiveloadtalk/contribs 15:59, 24 March 2021 (UTC)[reply]

Defining it later is perfect in my opinion and I don't think we need to solve everything now. I just wanted to make sure that the intent of what the policy is meant to cover is clear. MarkLSteadman (talk) 16:14, 24 March 2021 (UTC)[reply]
The existing wording of the policy (which the current proposal does not change) is written to prohibit linking "George Washington's wife" to "Martha Washington". It is intended to permit, barely, linking Alice to Alice in Wonderland (iff that is the clear intent in context) because it is merely an abbreviated title, or to pages in Author: or Portal: even though those prefixes do not appear in the text, but not any other circumspect or oblique references. It is very much word choice and not topic (it is hyper-conservative). Whether this should be loosened, and how, is, as Inductiveload suggests, best handled in a separate (later) discussion (I think permitting topic-equality is probably a good starting point for that discussion).
Incidentally, a blanket statement that such circumspect references are expressions of misogyny does not hold in the general case (that's applying an anachronistic lens), and why we need to be careful about how much scope we allow contributors through the policy. There are certainly works where that is the case, but in general that style is mere fashion of the times, applied more often to avoid implying excessive familiarity or be perceived as rude in naming some personage outright. Or to put it another way, the references are more often to "Mr. ———" (or "the K———", or "Earl ———") than "Mrs. ———". --Xover (talk) 18:44, 24 March 2021 (UTC)[reply]
I think the existing wording of the policy can be read both ways as saying "George Washington's wife" and "Martha Washington" are essentially the same (topic-equivalant) with "different wordings" (maybe open to debate about how far before it bleeds into non-neutrality or is no longer "minor": "Martha Washington" vs. "Mrs. Washington" vs. "Mrs. George Washington" vs. "George Washington's wife" vs. "the mistress of Mount Vernon" vs. "the first first lady" etc.) or as you read it, it clearly prohibits almost all of them as not "essentially the sanme" with a narrow reading of "different wordings" being only, say "The Earl ———" vs. "The Earl of ———". MarkLSteadman (talk) 19:25, 24 March 2021 (UTC)[reply]
Like I said, I think this is ambiguous (since it doesn't actually say "wording", it can be read in different ways), and nothing spills ink like people trying to decide after the fact which way an ambiguous thing was intended. We have 24,000 SCOTUS decisions at WS alone because the US Constitution is ambiguous. We have ten-thousand word arguments at wS:PD because the WS:WWI is ambiguous. The difference between us and a Supreme Court is that we can change the thing we're !arguing over: rather than trying to work out what the immutable policy says, we should rather work out what we want it to say. And then say it as unambiguously as we can. Inductiveloadtalk/contribs 19:50, 24 March 2021 (UTC)[reply]
Given their isn't a strong prescriptive intent around this particular topic in the proposal, more reflecting existing wording I agree it is a good follow-on discussion to have around this particular topic given the much broader scope of this policy and given it is a) exisiting wording and b) it is more interpretation than the wording itself. So I am comfortable calling this resolved for this RFC. MarkLSteadman (talk) 20:23, 24 March 2021 (UTC)[reply]

Proposal to allow wikilinks to dispel harmful falsehoods[edit]

If a source describes as true a harmful false statement, we should be able to link to the Wikipedia article that explains why this statement is false. For example, if a source says, "Jews are known to kidnap Christian children to use their blood to make matzoh", we should link that statement to W:Blood libel, which explains the falsity of the statement. This is not a circumstance where we should honor the intent of the author. BD2412 T 01:34, 24 March 2021 (UTC)[reply]

@BD2412: I don't know whether to agree or disagree. I'm more inclined to agree with you since there is a serious problem today of falsehoods being presented as truth in the media etc. and people believing those dubious sources unquestionably. But on the other hand, are we in the business to debunk false statements presented in works, or would it be more honorable to present them as they were without biases? Since works transcribed here are usually in the public domain, anyone else can debunk them externally from Wikisource if they so desired, without any legal implications whatsoever.
Also, what would you say about entire works that were written with the literal purpose of being racist, for example Thomas Dixon's The Clansman, a handful of his other similar novels, The Birth of a Nation, or The International Jew? There are bound to be many false and dangerous statements within, some that even promote violent racist acts. In those cases would you condone a disclaimer on the header in the "notes" section that explains that the works are preserved here for historical purposes, and not to promote those views? We could make a disclaimer template for works that are very likely to promote dangerous views, and I think I would be in support of that. (Since if we didn't do something like this the entirety of those transcluded white supremacist books would be filled with wikilinks.) PseudoSkull (talk) 14:23, 24 March 2021 (UTC)[reply]
There are some falsehoods for which there is no honor in letting them be presented as facts. I have seen in my time here a number of debunked racist tropes casually peppered into a work, as well as false claims in other areas, such as old anti-vaccination tropes. Some of these are perhaps merely offensive rather than harmful, but that is always going to be something of a judgment call. A disclaimer is not at all a bad idea for works of sufficiently disreputable tone. BD2412 T 15:44, 24 March 2021 (UTC)[reply]
I don't think any works, especially ones that are over 95 years old as applies to most works that are here, should be trusted by anyone as arbiters of absolute truth. I always gathered that we keep them here for historical purposes. Unfortunately some people will come to Wikisource, read some bits of old texts, and assume that everything they just read is the absolute truth just because it seemed reasonable enough to them for whatever reason. But I would say those people are missing the point of Wikisource, that point being to present the information as it was written, and not to convince anyone of anything. To make things worse, many people when coming to a sister project of Wikipedia literally think they're still at Wikipedia when navigating here, so may read our books thinking they are Wikipedia articles... This is why I have suggested in the past that we make these things far clearer in our general disclaimer page; that is, we should probably make very clear that we don't adopt any particular views presented in works transcribed here, and don't work to encourage any of them. PseudoSkull (talk) 16:29, 24 March 2021 (UTC)[reply]
I also don't think we should have a blanket policy of not being able to link to authors or works because the author wrote "John Doe's wife wrote a book of poems" because "wife" and "book" are not valid target and the "author intent" was that a book by a women author was not worthy of mentioning by name. MarkLSteadman (talk) 15:19, 24 March 2021 (UTC)[reply]
I don't know that I'd agree. If "George Washington's wife" for example was unambiguously referring to one specific author, it should probably link to that author page. And if she only ever wrote one book of poems, that's an unambiguous enough reference to warrant a wikilink to that text. Of course where there is ambiguity involved (for example if George Washington's wife wrote many books of poems) I wouldn't say to guess which one. PseudoSkull (talk) 15:30, 24 March 2021 (UTC)[reply]
"If "George Washington's wife" for example was unambiguously referring to one specific author." If this means "the only unambigous author reference is George Washington" we disagree. If it is "the only unambigous author reference is Martha Washington" we agree actually. My point isn't about ambiguity, it is about author intent and that we shouldn't respect "women don't matter, only their husbands do."MarkLSteadman (talk) 15:46, 24 March 2021 (UTC)[reply]
I have similar ambiguous feelings as PseudoSkull. Old works are full of harmful false statements, besides antisemitic and rasistic ones also statements about women, immigration, child labour, homosexuality and many others. I think we could find harmful and false statements in works being gathered at Portal:State Council of the People's Republic of China/The Record of Human Rights Violations in the United States. Is it our task to keep looking such statements up and link them to explanatory articles? Maybe a note in the header for the worst cases could solve it better.
However, no matter whether the answer to this question is positivive or negative, my suggestion of the rule in fact does not disallow such links if really needed, speaking about rare cases when modern readers may need clarification of the topic and when such clarification is really needed for understanding the work hosted in Wikisource. --Jan Kameníček (talk) 15:38, 24 March 2021 (UTC)[reply]
I would not have read "clarification" to include situations where the author's intent is clear but the information conveyed is false and harmful. However, if that is the intent, that's fine. BD2412 T 15:51, 24 March 2021 (UTC)[reply]
I'm not sure in-lining such disclaimers in the body of a text is the best way to approach this. It's another example of overloading of wikilinks to cover two things:
  • Type-A: a "classic" wikilink: a link to somewhere where you can find out more, see a definition, etc.
  • Type-B: to disambiguate/correct something that is, intentionally, accidentally, or through the passage of time not clear or, as in this case, flat-out wrong and needing refutation. Type-B links are not used because you think the reader wants to follow the link as such, they're just a way to say "the text says X, what it means is Y"; that the link can be used to get to "Y" incidental
The latter is where we start really getting into annotation territory and, I think, where much of the push-back on wikilinks is coming from (not only in this section).
The problem with using wikilinks for this kind of annotation is that it's not clear what the Wikisource editor's intention is, and that's a violation of the Principle of least astonishment (←Type-A link). Say the example above is linked to "blood libel". Why is there a link there? Is it going to take me to an example of that actually happening (as it would if it was a Type-A link)? Is it therefore confirming the statement? Only once you click the link do you end up at "blood libel", and, assuming you didn't know what that was, you have to read the intro at least of that article. Only then will you be able to make the connection (and if people are dumb enough to fall for 100-year old anti-Semitic tropes, even this is not guaranteed), that this is why the Wikisource editor left that link there: as an annotation to debunk a false statement.
This is something, which if we are going to allow/encourage/whatever, is IMO better served by something that's clearly an annotation, so that the debunking/refutation intent is clear. Using something like {{user annotation}} with a message like This is an example of w:blood libel(←now it's an obvious Type-A link) and is an anti-Semitic canard. is far more obvious than dropping a wikilink and hoping readers make the deductive leap that the link is itself supposed to be a debunking.
Furthermore, remember that upon export, these links will very likely not work at all, so if your intention was to clarify, debunk or annotate something, you have failed if the user can't follow that link (e.g. if it's a printed PDF, you have no idea whatsoever where that link even goes, you can't hover over paper).
Wikipedia covers the gist of much of all this at w:MOS:LINK, but specifically I think w:MOS:FORCELINK applies even at enWS: …but as far as possible do not force a reader to use that link to understand the sentence. The text needs to make sense to readers who cannot follow links…. w:MOS:LINKCLARITY is also relevant here.
Tl;dr is overloading wikilinks as annotations really the best way to deal with this? Inductiveloadtalk/contribs 17:00, 24 March 2021 (UTC)[reply]
@BD2412: Does this need to be hashed out in this proposal? Right now we don't allow links to Wikipedia at all (which is one of the things the current proposal is trying to address), and some members of the community find the very notion of permitting these offensive. By trying to also add links that reflect the Wikisource contributor's personal moral judgement and with a goal of refuting the very text they're embedded in, you're stretching this so far that even a pretty linking-positive contributor like me is apt to oppose just because we're way far into slippery slope territory.
What do we do when the North Macedonian nationalists start insisting on linking to refute any historical work discussing Greece? Or the Polish are offended at a work that might conceivably be read to suggest Nazi tendencies in Poland if you squint just right? How many fights are we going to have over Sheruchij's personal collection of erotica (some of which shocks and offends me, and I've been online since the early nineties)? How many links will we need to add to Nobokov's œuvre before it's safe to host when that time comes? How about Catullus? Do we linkbomb Kipling's A School History of England because of the colonialist and racist attitudes of its author, that are on full display for the modern reader but was probably entirely invisible to its original intended audience?
These are all rhetorical, of course; but they all need to be answered in any proposal to permit such links. That's going to be a long and complicated discussion, and, I would assert, one better kept separate from all others. (And ditto for related proposals such as disclaimers and trigger warnings; let's not bring those cans of worms into discussions that don't absolutely need them.) --Xover (talk) 17:32, 24 March 2021 (UTC)[reply]
"Fuzzy-Wuzzy" (unindexed) by Kipling et al. CYGNIS INSIGNIS 17:16, 25 March 2021 (UTC)[reply]
This is definitely a tricky problem, but I think that in principle I disagree with the idea that texts on wikisource should be fact-checked or rebutted. We don't have a policy similar to w:Wikipedia:What Wikipedia is Not but I suppose the closest is what can be inferred from our inclusion guidelines. I would say that generally speaking Wikisource is a place for faithful reproductions of verifiable source texts--a place for sources--and not a place for user-generated interpretive annotations of those sources, even when the interpretation is uncontroversial or trivially demonstrated (eg, stating "this is an example of the blood-libel conspiracy theory" or "the event being referenced here did not happen"). It's a tricky thing, as we obviously do minor, uncontroversial annotations all the time via Template:SIC, changing ſ to s sometimes, or linking to referenced authors/sources as per this proposed policy. In my opinion, commentary on sources makes the most sense to include and point to when that commentary is itself also a "source". I think it would be more appropriate for a page with a source to link to a second source providing critical commentary on that source, either in the header or near an "Editions"/"Versions" list (something like a "Works about X" section), than to do any degree of "inline" commentary. (This all being said, I'm flexible on this. If the consensus is that footnotes or margin-notes that use {{User annotation}} are sufficient to separate the "source" from "commentary," I would support that). Mathmitch7 (talk) 15:35, 29 April 2021 (UTC)[reply]
Oh, and to add to myself here -- the proposal for a general disclaimer at the top of an objectionable or harmful work is one that I would definitely get behind, but I think I agree with User:Xover that that should probably be a separate discussion. Mathmitch7 (talk) 15:38, 29 April 2021 (UTC)[reply]

Deeplinking and versions[edit]

A emergent problem as this site has developed, that has a bearing on this bit:

Basic wikilinks
[…]
 *  References to other works should be linked (whether we currently host them or not). If possible, do a deep link (see the section in the Style Guide), but at least link to the full title.

I've done a bit of deeplinking, but notice that in some cases this may be over-ridden by linking to a versions page; something I realised was going to be a problem a while back (works like journals are probably okay, thankfully). Perhaps some mention in guidelines or policy that this might occur. @EncycloPetey:, I think, can confirm that occurred during his efforts to reconcile some Shakespeare works. CYGNIS INSIGNIS 16:12, 25 March 2021 (UTC)[reply]

Yes, we have had that problem many times over with the works of Shakespeare. A well-meaning author created many section links for specific quotations from various plays and linked them into the only existing play copies we had, most of which were unsourced and had to be moved. Many problems have resulted, including: (1) the unsourced target texts have lots of invisible anchors for those quotes, (2) when the unsourced texts were moved to make way for a better copy, or were split into pages by "Acts", the links were broken, (3) when those links are preserved, they imply that the unsourced text is somehow authoritative over and other copy of Shakespeare's plays, since they are the ones being linked, (4) sometimes the quote in the quoting source does not match the quote in the play, because there are different editions of some plays, and these editions can sometimes be very different from each other. And there are other problems besides, but the biggest headache is that of setting up the disambiguation pages, which required moving all the target pages and then editing all the links to those targets. --EncycloPetey (talk) 16:21, 25 March 2021 (UTC)[reply]

Adopt this?[edit]

It seems like this got pretty good consensus (except for the question of linking to Wikipedia). Are we good to adopt it, and update the original page? JesseW (talk) 12:26, 2 August 2022 (UTC)[reply]