When you press the "normalise everything" button, it changes the parameters last1 and first1 to just last and first. I'm not sure whether this a bug or even if it means anything, but just incase I thought I'd mention it. greyzxqtalk19:41, 12 February 2023 (UTC)[reply]
Meaningless. |last= is an exact alias of |last1=. There is no benefit to making this change. And, there are other tools (User:Citation bot comes to mind) that change |last= to |last1= – also a meaningless change. Because the change is meaningless, this tool should stop making these changes.
When normalizing references with ProveIt, the |website= parameter is placed after |access-date= when using {{Cite web}}, but then |magazine= is moved to between |url-status= and |archive-url= when using {{Cite magazine}}, and for some reason |newspaper= is incorrectly converted to |work= and moved to between |title= and |url= when using {{Cite news}}. For consistency's sake, please consider having |website=, |magazine=, |newspaper=, |work=, |publisher=, |agency=, and all other aliases be placed after |access-date= on all citation templates. Thanks. InfiniteNexus (talk) 05:00, 24 February 2023 (UTC)[reply]
For the record: |website=, |magazine=, |newspaper=, and |work= along with |journal= are aliases of |periodical=. |agency= and |publisher= are not aliases of each other nor are they aliases of |periodical=.
cs1|2 does not care about the order of parameter placement in the wikitext of a template. For me, I think that |agency= should follow the author name-list in {{cite news}} templates; the periodical aliases should follow the |title= parameter in {{cite journal}}, {{cite magazine}}, {{cite web}} templates; and |publisher= (if present in periodical templates) should follow the periodical parameter or |location= in non-periodical templates. No doubt others think differently so no matter which ordering is chosen, someone will object so one has to wonder if attempting to 'fix' the OP's complaint will be worth the effort.
I'm referring to web citations. If using {{Cite web}}, ProveIt uses {{Cite web |last= |first= |date= |title= |url= |url-status=live |archive-url= |archive-date= |access-date= |website=}}; if using {{Cite news}}, ProveIt uses {{Cite news |last= |first= |date= |title= |work= |url= |url-status=live |access-date= |archive-url= |archive-date=}}; if using {{Cite magazine}}, ProveIt uses {{Cite magazine |last= |first= |date= |title= |url= |url-status=live |magazine= |archive-url= |archive-date= |access-date=}}. That is terribly inconsistent. InfiniteNexus (talk) 18:02, 24 February 2023 (UTC)[reply]
In this test edit, the previous revision uses a consistent parameter order. You can see what happens when you click "Normalize references", which I thought was supposed to make all citations consistent. InfiniteNexus (talk) 18:04, 24 February 2023 (UTC)[reply]
@1AmNobody24 Hi! Well, it seems to me that the "website" parameter isn't required under all circumstances, so it shouldn't be marked as required and I just unmarked it. Proveit should catch up in a few minutes. Could you confirm is the error persistes then? Cheers! Sophivorus (talk) 11:24, 12 May 2023 (UTC)[reply]
Hi Sophivorus, 'Name of the website' isn't shown in bold anymore, but the CS1 Error still remains if I add the website. Should the status at Template:Cite_ssrn/doc#TemplateData be changed to optional instead of suggested, so it only shows up after hitting 'Show all fields'? Nobody (talk) 11:40, 12 May 2023 (UTC)[reply]
If you look at Template:Cite ssrn § TemplateData you will see a long list of error messages among which is |website= is not a valid parameter. {{cite ssrn}} does not support |website= so that parameter should not appear in the TemplateData. To avoid having my butt chewed (again) because I changed TemplateData, I wrote a Lua script to highlight errors in the TemplateData. The error messaging worked to fix the errors in the TemplateData of the big-four ({{cite book}}, {{cite journal}}, {{cite news}}, {{cite web}}) and many others. Some templates still lag behind. The thing to do is to fix the TemplateData but because of the aforementioned butt chewing, I shall not be the editor to make those fixes.
Hi Trappist, I went ahead and did this[1], which I suggested above: Should the status at Template:Cite_ssrn/doc#TemplateData be changed to optional instead of suggested, so it only shows up after hitting 'Show all fields
It works as intended for ProveIt so I would call this
This toggle appears behind the exact same location as the ProveIt toggle. Is there a way for the ProveIt toggle to shift to the left under Vector 2022? Aaron Liu (talk) 22:42, 14 June 2023 (UTC)[reply]
Thanks! Is there a way to change this so it is at its normal position when the button isn't there i.e. when the viewport is too small? Aaron Liu (talk) 20:55, 15 June 2023 (UTC)[reply]
@Aaron Liu Not sure what you mean, but perhaps something along the following lines:
I have been having problems using ProveIt and I am not sure why. When I select the [ProveIt] button in the bottom right place on my screen, I get a dialog box to add a reference or bibliography, but it doesn't show the list of citation templates and says [no template].
Every now and again, I will see a [reload] or [load] button above ProveIt. And if so, it sometimes fixes the issue.
At the moment, ProveIt or load function + ProveIt are working about 10% of the time. I tried using the classic ProveIt function, but that's not working at all. What can I do to fix it? Or is there a better option for formatting multiple kinds of citations?
I have appended &withJS=MediaWiki:Gadget-ProveIt.js&withCSS=MediaWiki:Gadget-ProveIt.css to Wikipedia URLs, and enabled and disabled ProveIt in preferences, but there is no change and appending &withJS=MediaWiki:Gadget-ProveIt.js&withCSS=MediaWiki:Gadget-ProveIt.css yields "Wikipedia does not have an article with this exact name". I have also tried to follow the instructions on the install page. Chamaemelum (talk) 23:41, 5 July 2023 (UTC)[reply]
@Chamaemelum If the problem is only with the first URL, then the cause may be that it's malformed. The &withJS=MediaWiki:Gadget:ProveIt.js&withCSS=MediaWiki:Gadget-ProveIt.css must follow some other query string parameter, such as action=edit when editing a page. See Query string#Structure for more. I just edited the documentation to clarify this, since it lent itself to confusion. Kind regards, Sophivorus (talk) 13:58, 7 July 2023 (UTC)[reply]
Thanks! Sorry about that. Interesting that it still doesn't show up automatically when I'm logged in with it installed, though. Chamaemelum (talk) 14:06, 7 July 2023 (UTC)[reply]
The ProveIt logo showed up in the bottom right-hand corner of my screen for the first time, and at the same time, it immediately froze the page and my computer kernel panicked, restarted twice, and froze an additional time. Really weird, but I guess that's just me. I have a standard, new computer and browser. Chamaemelum (talk) 23:54, 7 July 2023 (UTC)[reply]
I really like ProveIt, but I very often find the result fails validation in the cite template with
{{cite journal}}: Check date values in: |date= (help)
The ProveIt tool inserts a date like 2009-01, cite flunks it, so I have to manually edit to 2009-01-01.
Can ProveIt just put a valid date in? I always us DOI inputs, so the date is whatever the publisher offered, typically only a month not a day. Johnjbarton (talk) 17:20, 12 July 2023 (UTC)[reply]
the date values in YYYY-MM format could be patched up. The simplest patch would be YYYY-MM -> YYYY-MM-01; the best patch would be to convert the MM to a month, Month YYYY. Johnjbarton (talk) 01:38, 22 February 2024 (UTC)[reply]
The simplest patch would be YYYY-MM -> YYYY-MM-01. Don't do that. Earlier versions of citoid and/or its predecessors did that but, thankfully, they no longer do. Making up a day-specific date to avoid the cs1|2 error resulting from the MOS restriction is a bad practice that should not be restored.
It was once proposed that citoid use the Library of Congress EDTF format (subsequently made part of ISO 8601-2019) YYYY-MM-XX (T132308) and support for that was provided in Module:Citation/CS1/Date validation. The cs1|2 module auto-translated EDTF dates to Month YYYY. Ultimately, the the proposal was abandoned.
{{#time:F Y|YYYY-MM|<local language tag>}} → {{#time:F Y|2001-02|en}} → February 2001
That should return the proper date for most wikipedias which you can then insert in to the cs1|2 template's |date= parameter. You may need to use {{#time:xg Y|YYYY-MM|<local language tag>}} for those languages that distinguish genitive from nominative (whatever that is – grammar in my own language is difficult enough for me so I don't even try to understand grammar in other languages).
The above not being possible, I suppose that you could create some sort of data array that maps language tag and month number to month name for that language.
As there is an option in the Automatic Reference field and the "Load" button, it would be great if there were something like a "Get latest web.archive data" button on front of the URL field. It could use their API as described at [2] to get the latest archive URL and date, it would be really helpful. :) — Arthurfragoso (talk) 10:52, 14 July 2023 (UTC)[reply]
I asked Claude AI to implement the changes, but I still had to do some manual changes. A bit hackish, but it works. Add the following code:
if (inputName === 'url') {
$button = $('<button>').text('getArchive');
$div.prepend($button);
$button.on('click', $input, function(event) {
$.getJSON('https://archive.org/wayback/available?url=' + encodeURIComponent(event.data.val()))
.done(function(data) {
// Check if archived
if (data.archived_snapshots.closest) {
// Get closest snapshot
var snapshot = data.archived_snapshots.closest;
var curstatus = $('#proveit [name=url-status]');
// Populate fields
if(!curstatus.val()) {
curstatus.val('live');
}
$('#proveit [name=archive-url]').val(snapshot.url.replace(/^http:\/\//, "https://"));
$('#proveit [name=archive-date]').val(snapshot.timestamp.slice(0, 8).replace(/(\d{4})(\d{2})(\d{2})/, '$1-$2-$3'));
} else {
alert('URL not archived');
}
})
.fail(function() {
alert('Error accessing Wayback Machine API');
});
});
}
just before the if ( paramData.type === 'date' ) {
@Arthurfragoso Hi! I just implemented an Archive button based on your code. Unfortunately your code makes several assumptions about the names of parameters and only works for the English Wikipedia, so I had to modify it to make it cross-wiki, which for now means it only shows you the archived URL and you have to copy it to the appropriate field. But it's way better than nothing and at least it works for all wikis. Sophivorus (talk) 14:59, 20 July 2023 (UTC)[reply]
If this tool works for {{cite book}} then, if we apply the same wikilink 'trick', this tool should also work for {{cite map}}. Seems to me that the correct 'fix' here is to make {{cite map}} like {{cite book}}. Someone who uses this tool should make that test and report back.
ok i did it to match {{cite map}} like {{cite book}} - it had been a while since I'd edited doc pages when I made the edit request and I was a bit rusty. Right now it's still saying "no template data," but iirc it's pretty typical for changes to need time to propagate through the wiki before it can be tested. -Furicorn (talk) 04:04, 25 November 2023 (UTC)[reply]
@ActivelyDisinterested I think not. I just tried editing the old version of the article and editing the reference with ProveIt, and the date disambiguation was preserved. Thus, I think its removal was a mistake by the editor, rather than ProveIt. Sophivorus (talk) 13:41, 4 October 2023 (UTC)[reply]
I am once again asking for {{Cite magazine}} and {{Cite news}} to place |magazine= and |newspaper= after |access-date=, as {{Cite web}} currently does with |website= and |publisher=. This is the only parameter whose placement ProveIt is inconsistent with when normalizing references. InfiniteNexus (talk) 22:39, 19 November 2023 (UTC)[reply]
Will ProveIt allow an editor to publish an edit using it if the only function it is performing is the reordering of template parameters? This would seem to violate WP:COSMETICBOT if so. I've seen it come close: in this diff, it reorders some template parameters while misparameterising two values, doing nothing constructive.A follow up thread related to the OP here is active at Help talk:Citation Style 1/Archive 91#Consistency in parameter order, should the maintainers wish to engage. Happy Friday, Folly Mox (talk) 11:49, 24 November 2023 (UTC)[reply]
I'd gently suggest that disabling the functionality to publish edits that solely reorder parameters be incorporated into future builds. Not all editors using the gadget will be familiar with bot policy, and it seems safer to prevent that behaviour in the software than to course-correct editors through education. Folly Mox (talk) 12:23, 24 November 2023 (UTC)[reply]
Detecting if the only changes done are cosmetic, would be disproportionately difficult to code and prone to errors. A simpler solution may be to introduce a config option to disable the "Normalize all" button for the English Wikipedia (since it's the only wiki with this concern so far), or showing a warning with a link to WP:COSMETICBOT or similar when the button is clicked. I lean towards the second because some users may still find the functionality useful and I anticipate that if we disable it, we'll soon have people in this talk page asking about it. Sophivorus (talk) 13:24, 25 November 2023 (UTC)[reply]
Would it be possible to have a "don't show me again" button when using the normalize feature to skip the WP:COSMETIC warning? I imagine most times that button is clicked, the editor is fully aware of the potential issues and policy, and is operating responsibly. Remsense留17:45, 2 December 2023 (UTC)[reply]
Thanks for adding the new "archiving" feature. (The one that puts an "Archive" button in the URL field and searches the Internet Archive for archived revisions). I noticed, however, that the button also appears in the Archive-URL field, which seems a but odd. Was this deliberate? 🌺 Cremastra (talk) 14:05, 25 February 2024 (UTC)[reply]
@Cremastra Hi! No, it's not deliberate, but a side effect of the fact that Proveit adds the Archive button to every field of type "url" (as defined by the template data of each citation template). I can't think of a fix that works cross-wiki yet. Sophivorus (talk) 20:51, 7 March 2024 (UTC)[reply]
Anyone know how ProveIt could have gotten this idea? It's taken OCLC 922086108 (A.C. Fox-Davies, A Complete Guide to Heraldry), and mapped it to DOI 10.1016/0006-8993(79)90456-6 (Palacios, Niehoff, Kuhar, "Ontogeny of GABA and benzodiazepine receptors: Effects of Triton X-100, bromide and muscimol"). Aware this error is almost two years old at this point, but thought I'd bring it here just in case. Folly Mox (talk) 12:48, 21 May 2024 (UTC)[reply]
Surprised this hasn't been brought up before, but ProveIt's interface is coming up unreadable for me with the site's appearance selector set to Dark mode; all of the text is white-on-near-white. The main issue seems to be that many of ProveIt's styles don't follow the Recommendations for night mode compatibility on Wikimedia wikis, which say:
Always define color when defining background
When defining a background color, it may be tempting not to define the color if it is the same as the article text color. However, when different themes e.g. night mode are applied, this could have unintended consequences (e.g. white text on a yellow background). It is thus recommended that you always define the two together.
In Special:Diff/1242637481, this script mangled a reference containing |editor3-first=Deyuan <!--|year=1994 onwards--> by reordering the parameters so |year=1994 onwards--> came before |editor3-first=Deyuan <!--, leaving the reference with an unclosed comment and breaking display of that and all subsequent refs in the article. It also seems to have somehow duplicated some of the other parameters. Anomie⚔13:56, 8 September 2024 (UTC)[reply]
Would it be feasible to automate in normalization, per Help:Citation Style 1#Work and publisher, the removal of redundant locations e.g. publisher=Cambridge University Press |location=Cambridge? Of course, the one hang-up I can immediately imagine is when multiple locations are specified, but that seems easy to consistently check for. Remsense ‥ 论18:03, 22 October 2024 (UTC)[reply]
Resolved
