This template does not require a rating on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Academic Journals This template is within the scope of WikiProject Academic Journals, a collaborative effort to improve the coverage of Academic Journals on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.Academic JournalsWikipedia:WikiProject Academic JournalsTemplate:WikiProject Academic JournalsAcademic Journal Astronomy This template is within the scope of WikiProject Astronomy, which collaborates on articles related to Astronomy on Wikipedia.AstronomyWikipedia:WikiProject AstronomyTemplate:WikiProject AstronomyAstronomy Physics This template is within the scope of WikiProject Physics, a collaborative effort to improve the coverage of Physics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.PhysicsWikipedia:WikiProject PhysicsTemplate:WikiProject Physicsphysics

I'm a bit confused as to why this template exists - is it not better to use Template:Cite journal, with Template:Arxiv in the ID field? Mike Peel 12:10, 24 August 2006 (UTC)[reply]

I imagine so that one can cite papers that exist on the arxiv but have not (yet) appeared in journals. —David Eppstein 06:48, 16 September 2006 (UTC)[reply]

My point was that it is easy to do something like:

• Marathe, Madhav V. (1994). "Geometry based heuristics for unit disk graphs". arXiv:math.CO/9409226. {{cite journal}}: Cite journal requires |journal= (help); Unknown parameter |coauthors= ignored (|author= suggested) (help)

... which cites arXiv on its own, with no reference to a journal location, while leaving the option open to add in a full journal reference in later. I've agreed with User:Hillman (the creator of this template) to not do anything with this for at least a few months, so that he can sort through other problems he's having with Wikipedia first. Mike Peel 07:55, 16 September 2006 (UTC)[reply]

That does look better. I didn't know about that id trick — maybe it deserves more prominent mention earlier on this talk page? —David Eppstein 14:57, 16 September 2006 (UTC)[reply]

Merge with {{cite paper}}? — Omegatron 02:49, 24 April 2007 (UTC)[reply]

I'd like it if the cite arxiv template looked something like

• Marathe, Madhav V., Breu, Heinz, Hunt, Harry B., III, Ravi, S. S., and Rosenkrantz, Daniel J. (1994). "Geometry based heuristics for unit disk graphs". arXiv:math.CO/9409226.

instead of

• Marathe, Madhav V., Breu, Heinz, Hunt, Harry B., III, Ravi, S. S., and Rosenkrantz, Daniel J. (1994). "Geometry based heuristics for unit disk graphs." .

For one thing, the doubled period looks bad. For a second, unless I hover over the link I can't tell the first kind of cite apart from a random web page anywhere else. We include publishers of other kinds of sources; why should we slight the arxiv in this way?

On an unrelated issue, it's not obvious how or why to fill in the "version =" part of the template for a paper that doesn't have multiple versions. Omitting it doesn't work correctly. The cite above was generated with a blank version. The examples on this page seem to indicate that the version should be filled in with the upload date, but arxiv uses a different numbering scheme in which versions are like "v1", "v2", etc. The year is already visible elsewhere, but if you want a date as well, why not call it a date? And why not allow the "16 Dec 2005" style format appearing on arxiv pages for easier copy and pasting?

—David Eppstein 06:48, 16 September 2006 (UTC)[reply]

Hi,

I've upgraded the template so it uses Template:Citation/core. I've been careful to retain current functionality; in addition, I've integrated the template with User:Citation bot. Therefore it is now only necessary to add an |eprint= parameter to the template, and you can sit back and watch as the bot automatically fills in the rest for you.

Hope that doesn't cause any problems!

Martin (Smith609 – Talk) 02:44, 17 February 2009 (UTC)[reply]

In other citation formats, there is a preference for using a separate parameter for each author. I am aware of two main benefits: firstly, metadata is more readily extracted and more usefully presented; secondly; there is improved compatability with the Harvard style of inline referencing. The template already supports this behaviour, and I consider it non-controversial to update the template documentation accordingly. Please dissent below! Martin (Smith609 – Talk) 19:53, 13 April 2009 (UTC)[reply]

See Template_talk:Cite_web#Upper_case_form_of_template_name. Debresser (talk) 06:37, 21 October 2009 (UTC)[reply]

I commented out the suggestion that citations be migrated to cite journal and the published journal, to "help the bots", firstly that's not a good reason, secondly there is or was a bot doing the actual moving, which suggests that it can deal with either format, thirdly arXiv is freely accessible, whereas many journals charge, WP prefers the free link, and fourthly if there is any difference between the journal version and the arΧiv version, we are citing the wrong source. Rich Farmbrough, 14:34, 4 December 2009 (UTC).[reply]

I reverted your commenting, but clarified what was meant by "switch to cite journal". Headbomb {^ταλκ_{κοντριβς} – WP Physics} 14:56, 4 December 2009 (UTC)[reply]

We should definitely encourage keeping the arXiv link in place after formal peer-review acceptance. If authors make important corrections that are only present in the non-open-access version, then they take the risk that they will get cited only for their erroneous, open-access version. Scientists (1) in poor countries and/or (2) outside of a specialty (e.g. astronomy vs particle physics vs a field of mathematics) and/or (3) in a not-so-poor country institution with a low budget tend not to make the special/costly effort required to get non-open-access articles.

i've tried to clarify the wording a little - IMHO we need to be clear that we recommend retaining the arXiv link. Even for old articles in journals that are open-access-for-old-articles, there is still a higher chance of long-term perennity (e.g. 50 years? 100 years?) in arXiv than in individual journals - arXiv is efficient and has dozens of mirrors, while at least some journals claim to delete the original LaTeX source after a short period (2 years or so?), or do not even store it themselves, leaving that to the publisher. In 20 years' time, converting arXiv sources (LaTeX + postscript) to some future format would probably be cleaner than converting html (with files all over the place) to that future format. Boud (talk) 19:38, 12 August 2011 (UTC)[reply]

So we can cite to pages. Paradoctor (talk) 01:06, 29 January 2010 (UTC)[reply]

I am against this for arxivs. If you need to reference a particular page, it is better to write See page 10 in {{cite arxiv|...}}. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 01:13, 29 January 2010 (UTC)[reply]

Maybe, but for a single use it needlessly clutters up the article. There is no need to argue, I'm asking for optional behavior, not a default. If you don't like it, and don't want to do it, I have no problem with that. Just means I'll have to do the work myself. Is there a reason I shouldn't? Paradoctor (talk) 01:39, 29 January 2010 (UTC)[reply]

There is an RFC on the addition of identifier links to citations by bots. Please comment. Headbomb {talk / contribs / physics / books} 16:03, 15 August 2011 (UTC)[reply]

In this version of a Wikipedia article the class parameter of this template is used. The resulting reference (number 1) has an external link to the desired source material, followed by a generic second external link to the part of the archive that is associated with the class. This second link does not seem very useful to me. The documentation for this template states that the class parameter is optional and should be used only with new-style eprint identifiers that do not include the class. Is there some reason why the class parameter should be used in that situation? Can we add some usage guidance to the documentation? Thanks. Peter Chastain (talk) 13:38, 15 December 2011 (UTC)[reply]

{{cite journal}} and {{cite doi}} are now in the navbox. ---— Gadget850 (Ed) ^talk 19:46, 27 February 2012 (UTC)[reply]

{{Citation Style 1}} has been edited to show that this template is no longer bot-filled. What is the status? ---— Gadget850 (Ed) ^talk 00:21, 21 May 2012 (UTC)[reply]

It would be far more useful if {{cite arxiv}} worked in the same fashion as {{cite doi}} and {{cite pmid}}, etc, so that {{cite arxiv|xxxx.yyyy}} pointed to {{cite arxiv/xxxx.yyyy}}. Then we could have the citation completed on demand, and duplicated results taken care of with REDIRECTs between templates.

As far as I can see, if the paper is published, it gets a DOI which can be seen on the arXiv page, which could be used to decide whether to use {{cite journal}} or {{cite paper}}, although since the latter redirects to the former the distinction is reasonably moot.

Thoughts? —Phil | Talk 16:07, 24 July 2012 (UTC)[reply]

See arxiv for details. We may need to change the template. -- phoebe / (talk to me) 22:46, 6 January 2015 (UTC)[reply]

Note, I checked it and the output of doibot with the new 5 digit numbers seems fine. -- phoebe / (talk to me) 22:53, 6 January 2015 (UTC)[reply]

Today in the Sagittarius A* article, I had to remove two parameters in order to fix a red CS1 error; the params were:

• url – In the Title subsection this doc states Can be wikilinked to an existing Wikipedia article or url may be used to add an external link, but not both. Displays in italics. (There was no wikilink to a WP article.) Since I had to remove this param to eliminate the CS1 error, this may be either a doc-update need or a need to tweak the module? If the latter, then I can add the url back in after the tweak. Also, the title does not display in italics; it displays within quotation marks. (?)
• accessdate – This doc does not explicitly exclude support for this param; however, in the Usage section we find The template also supports most of the parameters supported by {{cite journal}} and uses the style of {{cite journal}}. So if the url parameter is included by tweaking the module, then wouldn't this param also be needed? – Paine  17:25, 27 April 2015 (UTC)[reply]

Here is a comparison between the {{citation/core}} and Module:Citation/CS1 versions of {{cite arxiv}}. As you can see, in neither case is the title linked:

Cite arXiv comparison

Wikitext	{{cite arXiv|accessdate=6 October 2012|arxiv=1210.1294|author10=K. Matthews|author2=A. M. Ghez|author3=R. Schödel|author4=S. Yelda|author5=A. Boehle|author6=J. R. Lu|author7=T. Do|author8=M. R. Morris|author9=E. E. Becklin|date=4 October 2012|first=L.|last=Meyer|title=The Shortest Known Period Star Orbiting our Galaxy’s Supermassive Black Hole|url=http://arxiv.org/pdf/1210.1294v1.pdf}}
Live	Meyer, L.; A. M. Ghez; R. Schödel; S. Yelda; A. Boehle; J. R. Lu; T. Do; M. R. Morris; E. E. Becklin; K. Matthews (4 October 2012). "The Shortest Known Period Star Orbiting our Galaxy's Supermassive Black Hole" (PDF). arXiv:1210.1294. {{cite arXiv}}: Unknown parameter |accessdate= ignored (help)
Sandbox	Meyer, L.; A. M. Ghez; R. Schödel; S. Yelda; A. Boehle; J. R. Lu; T. Do; M. R. Morris; E. E. Becklin; K. Matthews (4 October 2012). "The Shortest Known Period Star Orbiting our Galaxy's Supermassive Black Hole" (PDF). arXiv:1210.1294. {{cite arXiv}}: Unknown parameter |accessdate= ignored (help)

Because the {{citation/core}} version does not support |url= it had no reason to support |access-date=.

{{cite arxiv}} is intended to cite a preprint at arXiv. If editors are citing the final published version of the paper, then they should consider using {{cite journal}}.

Title is quoted because it is a paper or article preprint.

The documentation is wrong and is in need of fixing. In part this arises from the common documentation at {{csdoc}}.

—Trappist the monk (talk) 17:57, 27 April 2015 (UTC)[reply]

Thank you very much for that, Trappist the monk! So we can leave the article as is and tweak the {{csdoc}} template by omitting the ...or url may be used to add an external link, but not both. Displays in italics.? And maybe change that last part to "Displays within quotation marks."? According to the doc ({{csdoc}}), the wikilink can be added with the title-link parameter. Is this still the case? or should that also be omitted? And of course this all relies on whether or not the {{csdoc}} should be edited at all due to applications where these params still apply? If that is the case, then perhaps these params should be omitted and added manually to doc pages where they do apply? In other words, is this an easy or a gnarly fix? – Paine  18:54, 27 April 2015 (UTC)[reply]

I have mostly updated the "title" documentation. I tried to add a switch called "link" to exclude the wikilink/url text from the Cite arxiv documentation, but it's not working for me. Trappist or Gadget850, can you find my error? – Jonesey95 (talk) 21:07, 27 April 2015 (UTC)[reply]

The title documentation displays correctly now. – Jonesey95 (talk) 12:38, 30 April 2015 (UTC)[reply]

Virtually all arXiv documents can now be displayed in simple webpage (HTML) format, using the online tool "Arxiv Vanity". That could be quite useful for anyone using a device without access to Adobe Acrobat Reader, because they find it hard to read pdf doucuments or for whatever other reason. For example, the arXiv version of the paper above viewed as an HTML web page through Arxiv Vanity: The Shortest Known Period Star Orbiting our Galaxy’s Supermassive Black Hole.

If what is published at arXiv are preprints and there's no review process but "moderation" to categorize them correctly, that would seem to make them low-quality, primary sources, maybe too low for science/medicine citations at least (maybe okay in the humanities, as directly attributed primary sources used with caution for non-controversial claims and only when necessary. They would seem to have the reliability level of, e.g., a thesis/dissertation, maybe less.

An argument can be made that they're self-published sources, since arXiv is acting as a self-publishing house for academics (in WP:RS terms, if not in intent; I gather their intent is more like that of WikiSource and Project Gutenberg and Archive.org, but narrowly tailored).

Another argument can be made that it's WP:UGC; people all over the world uploading what they've written to a website (with an administrative user class) where it is categorized for public, free consumption ... that's a description of a Wiki.

Exception made for material that is actually published in a journal by the time we need to cite it; the arXiv URL might be the only one we have, if the journal is behind a paywall.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  23:32, 22 July 2015 (UTC)[reply]

Same reason why we might want to cite anything. To establish a claim, a fact, support a thing, or whatever. This is no different than using {{cite web}}. Headbomb {talk / contribs / physics / books} 02:34, 23 July 2015 (UTC)[reply]

Not responsive to a single word I wrote. I'm tempted to delete that as trolling, per WP:REFACTOR.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  07:51, 24 July 2015 (UTC)[reply]

I told you exactly why anyone would cite something from the arxiv. If that doesn't answer your question, then the problem is with your question, not my answer. Headbomb {talk / contribs / physics / books} 14:26, 24 July 2015 (UTC)[reply]

The respondents at the current discussion get the point; it isn't actually necessary that you do. [shrug]  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  03:58, 27 July 2015 (UTC)[reply]

It seems to me that either the documentation or the template needs updating because, unless I am missing something, |orig-year=, |access-date=, and |version= result in ignored parameter errors. Given |orig-year= is ignored, {{csdoc|date}} is not entirely applicable to this template. Perhaps it also needs a |arxiv=yes if there is some reason to exclude |orig-year= (it would be a useful parameter to have)? The TemplateData is also outdated, as well. —Nøkkenbuer (talk • contribs) 08:04, 28 June 2018 (UTC)[reply]

I have tweaked {{csdoc}}. I shall not touch the template data.

—Trappist the monk (talk) 10:53, 28 June 2018 (UTC)[reply]

Similarly |publisher= gives the unhelpful error "Unknown parameter |publisher= ignored (|publisher= suggested)" (see this version of Nvidia Quadro, ref 20. I've replaced it with |collaboration= there, which makes more sense anyway). › Mortee _talk 10:08, 23 August 2018 (UTC)[reply]

Yeah it shouldn't suggest publisher as an alternative to publisher. (Technically it's Cornell University Library, but that's not something that needs to be mentioned). Headbomb {t · c · p · b} 10:57, 23 August 2018 (UTC)[reply]

Fixed in the sandbox. |collaboration= as a 'replacement' for |publisher= seems inappropriate here because the authors are affiliated with NVIDIA (the authors' employer) – papers published by university professors typically don't credit the university that employs the professor (the affiliation) as a collaborator. Also, |author5=Dietger van Antwerpen should be |last5=van Antwerpen |first5=Dietger.

—Trappist the monk (talk) 11:39, 23 August 2018 (UTC)[reply]

See also

Cite arxiv comparison

Wikitext	{{cite arxiv|author1=Behnam Farid|bibcode=2007arXiv0711.0952F|class=cond-mat.str-el|eprint=0711.0952|title=On the Luttinger theorem concerning number of particles in the ground states of systems of interacting fermions|year=2007}}
Live	Behnam Farid (2007). "On the Luttinger theorem concerning number of particles in the ground states of systems of interacting fermions". arXiv:0711.0952 [cond-mat.str-el]. {{cite arXiv}}: Unknown parameter |bibcode= ignored (help)

from Luttinger's theorem.--

Auric talk 22:46, 11 September 2018 (UTC)[reply]

@Auric: those were due to a bot error. This has been fixed a while ago. Headbomb {t · c · p · b} 23:52, 11 September 2018 (UTC)[reply]

This template produces a green lock which I find displayed behind the eprint parameter in both Mozilla Firefox and Microsoft Edge, while in Internet Explorer it is displayed to the right. Hyacinth (talk) 22:27, 18 November 2018 (UTC)[reply]

Where are you seeing this?

—Trappist the monk (talk) 22:39, 18 November 2018 (UTC)[reply]

In R-algebroid arXiv:0804.2451. also has CiteSeerX 10.1.1.312.7226, so must use cite paper rather than cite arxiv. If easy please add the other likely parameters. RDBrown (talk) 01:19, 2 December 2020 (UTC)[reply]

This is not needed. This template is specifically for citing a preprint from arxiv. That the preprint can also be found on CiteSeerX is of no relevance. Headbomb {t · c · p · b} 01:26, 2 December 2020 (UTC)[reply]

It may be appropriate to include a 'doi' parameter, since that is intended to be stable. Praemonitus (talk) 16:39, 28 April 2023 (UTC)[reply]

Again no, because that's fully redundant to the arxiv identifier. Headbomb {t · c · p · b} 14:49, 30 April 2023 (UTC)[reply]

It seems like this template has no support for specifying a translator? –jacobolus (t) 19:02, 3 July 2023 (UTC)[reply]