Wikinews talk:Easy Peer Review (2025 version)

Issues with text highlighting in article

@Asked42, I'm not sure if this is the right place to report issues. If it isn't, let me know a better place and I'll move the comment.

I used EzPR 2025 today on Four countries back out of Eurovision 2026 amid ongoing boycott and enabled "Allow highlight in article" and added several phrases and issues. With allow highlight enabled, I received the following error: Inline highlight or unsourced marking couldn't be applied because the sentences are too complex or no matches were found. They will be ignored. All phrases were copy/pasted from the article and do exist as of Special:Diff/4954668.

It produced the following, which I successfully included without enabling allow highlight:

{{SIH
| text1 = The four withdrawals generated major turmoil
| issue1 = Needs attribution.
| text2 = The broadcaster for Slovenia reported
| issue2 = Attribution could be strengthened with a name.
| text3 = The controversy encapsulated complaints stretching for years.
| issue3 = Needs attribution. 
| text4 = Complaints grew after Israel started its military campaign in Gaza
| issue4 = Needs attribution.
| text5 = Due to Israel’s broadcaster being part of the EBU, the country is eligible to compete
| issue5 = Implied causality without attribution
| text6 = Some countries, namely Austria, the host of the Eurovision Song Contest in 2026, Germany, France, Britain, Portugal, and Sweden, supported the inclusion of Israel in the 2026 edition of the contest.
| issue6 = Needs attribution.
| text7 = and the high death count in Gaza.
| issue7 = Multiple sources say the broadcasters cited the humanitarian crisis or humanitarian impact of the war. This is stronger and broader than “death count.”
}}

Let me know how I can help test this further, if needed. Thanks again for a great tool! Michael.C.Wright (^Talk/_Reviewer) 15:35, 9 December 2025 (UTC)

Thank you for explaining the issue. I have tried to understand the cause behind it, but I am having some difficulty identifying it. I will keep looking and try to resolve it once I find the culprit behind this. Sorry for the inconvenience. -- Asked42 (talk) 14:57, 10 December 2025 (UTC)

No worries. The text issues highlight template is still quite useful and visually helpful.Michael.C.Wright (^Talk/_Reviewer) 15:07, 10 December 2025 (UTC)

I think the issue could be related to the difference between straight quotation marks and curly quotation marks, possibly due to how macOS and Windows handle them. I have made some changes to handle this. In future reviews if possible, please try it again and see whether the issue still persists. -- Asked42 (talk) 15:16, 10 December 2025 (UTC)

I just had a successful test of the highlight feature and it works well, though I didn't test it with any quotation marks yet. See Interpol Red Notice eyed against Zaldy Co.

I anticipate a request to make it a one-click removal of template:highlight. I was working on incorporating that feature into my MarkupManager (MM) but was trying to do too much with also being able to hide the template using MM.

Have you already built into EZPR 2025 the ability to remove template:highlight before being published? The template does use Module:DetectDraft to suppress highlights in a published article (in case the template is not removed before being published). So we have a fallback.Michael.C.Wright (^Talk/_Reviewer) 14:44, 12 December 2025 (UTC)

Good to know it works. And yes, EZPR 2025 has a built in function to remove highlights if they are present before publishing. -- Asked42 (talk) 15:15, 12 December 2025 (UTC)

@Michael.C.Wright: One more thing I noticed: in | text5 = Due to Israel’s broadcaster being part of the EBU, the country is eligible to compete, the quoted text uses EBU, while in the article itself it is written as European Broadcasting Union. This mismatch also causes the issue. When I tried using "European Broadcasting Union" in the quoted text, the same as in the article, it seemed to work.-- Asked42 (talk) 16:35, 10 December 2025 (UTC)

For the wish list

I've done a partial review twice now. The first time I did it as a test (successfully) and then immediately followed it up myself with the full review, to see what it looks like and it worked great, it found the partial review, etc.

It might be useful to post a templated notice on the talk page that a partial review exists for that article. We may even want to categorize partial reviews (Category:Partially reviewed?) so they can be easily found. That way if a reviewer has limited time, they could go straight to a partial review if they thought it would take less time to review.

It might also be useful to allow for "partial review notes" that could be 'temporary' in the sense that they are stored as data, and will load into the review space when a reviewer loads the partial review, but they are not otherwise kept and reflected in the final, full review. That way a partial reviewer can communicate things to the next reviewer easily.

We might also consider adding a notice when a partial review exists, to the effect of "The final, publishing reviewer takes full responsibility for the state of published articles. Unlike an edit history, there is no permanent record of what was reported in a partial review."

I understand if you want to wait for the SPTF decision before further development (don't blame you). These are not necessary for current use.Michael.C.Wright (^Talk/_Reviewer) 17:11, 11 December 2025 (UTC)

@Michael.C.Wright: That's a good idea. I can make the changes to implement these three cases, and they are not very complicated to add, so it shouldn't take much time.

For a partial review, do you think the script should add both reviewers' names to the peer review template? Or should it only add the notice "The final, publishing reviewer takes full responsibility..." for the publishing reviewer? -- Asked42 (talk) 17:48, 11 December 2025 (UTC)

> For a partial review, do you think the script should add both reviewers' names to the peer review template?

I’ve been thinking about this and see two competing points.

• The publishing reviewer carries full responsibility for the final state of a published article. Our policies, guidelines, and practices are unevenly documented. Reviewers often rely on institutional knowledge. That makes shared responsibility tricky.
• At the same time, collaboration between reviewers could lighten the workload and make partial reviews more useful.

The difficulty is that our reviewer group doesn’t yet operate with the level of trust and routine collaboration that would make this straightforward. Until that improves, it’s not obvious how much weight a publishing reviewer should place on a colleague’s partial review. We’ve had a high number of corrections and retractions this year, which suggests our review accuracy needs improvement.

I’d like to hear from other reviewers on this. In the meantime, using the notice as a stop-gap may be the safest approach while the community gets used to partial reviews and has time to discuss how they should function.Michael.C.Wright (^Talk/_Reviewer) 15:02, 12 December 2025 (UTC)

Sure. Let's move forward with a notice approach then. I will integrate a notice into the conform review dialog. -- Asked42 (talk) 15:20, 12 December 2025 (UTC)

Done: Now when saving a partial review, the script will post a notice using the {{Partial review log}} template on talk page. A notice will be shown when publishing a continuation of a partial review, allowing the reviewer to leave a note for the next reviewer. This way, multiple reviewers can have a small back-and-forth conversation by leaving notes for each other.

Let me know if there is anything I missed or might have misinterpreted. I will update the documentation soon. -- Asked42 (talk) 10:37, 14 December 2025 (UTC)

Tested notification of multiple contributors

Notifying multiple contributors with the ability to select which ones worked successfully! That's a very useful addition I think, as it could encourage more collaboration. 👍Michael.C.Wright (^Talk/_Reviewer) 16:03, 14 December 2025 (UTC)

Request for minor change to ezpr2

Hi @Asked42

1. In heading of user talk message i suggest writing "Published: Article Name (Month Day)" instead of "your submission (Month Day)". Replace "Published" with "Revision requested" (not "not ready" - the revision requested is more like a call for action) if review did not pass.

2. In any place where "not ready" any software or template I propose to write "revision requested".

Regards, -- Gryllida 05:20, 17 December 2025 (UTC)

On my way to implementing this, if there are no objections. -- Asked42 (talk) 14:53, 18 December 2025 (UTC)

@Gryllida: Done, Now the user talk page notification title will be as you suggested. -- Asked42 (talk) 15:16, 18 December 2025 (UTC)

Suggested comments

@Michael.C.Wright, Gryllida: I have added some suggested comments for the criteria specific comment fields. Please check them out and let me know if any of them are unnecessary or should be removed. If you have any suggestions for additional comment text, text to be removed, or anything else, please let me know as well, along with the relevant criteria. -- Asked42 (talk) 16:27, 18 December 2025 (UTC)

Will do! Thanks for all the work on this. 👍Michael.C.Wright (^Talk/_Reviewer) 16:29, 18 December 2025 (UTC)

The current suggested comments are:

:criteriaSuggestions: {
:	    copyright: [
:	        'Please check the comment below.',
:	        'Potential copyright violation detected (high similarity score).',
:	        'Close paraphrasing of the source text found; needs rewriting.',
:	        'Text copied directly from a non-free source.'
:	    ],
:	    newsworthiness: [
:	        'Article is stale or outdated and no longer newsworthy.',
:	        'Content is an opinion piece or editorial, not a news report.'
:	    ],
:	    verifiability: [
:	        'One or more links are dead or inaccessible.',
:	        'Claims are not supported by the provided sources.',
:	        'Reliance on unreliable sources (blogs, social media).',
:	        'Missing sources for key facts or statistics.'
:	    ],
:	    npov: [
:	        'Article contains loaded language or editorializing.',
:	        'Promotional tone detected.',
:	        'Article fails to represent all significant viewpoints fairly.',
:	        'Title is sensationalized or biased.'
:	    ],
:	    style: [
:	        'Lead sentence does not follow the 5W & H format.',
:	        'Please use the {{source}} template properly.',
:	        'Please consider the inverted pyramid style for the structure.',
:	        'Improper attribution for opinions, unverified claims, or speculation.'
:	    ]
:	},
:

-- Asked42 (talk) 16:29, 18 December 2025 (UTC)

Some quick observations:

Copyright: For "Text copied directly from a non-free source." we could just say "Text copied directly from another article." That covers both WN:Plagiarism and copyvio and avoids the argument 'the article isn't paywalled.'

Newsworthiness: Maybe add "Focal event backgrounded by more recent developments."Michael.C.Wright (^Talk/_Reviewer) 16:37, 18 December 2025 (UTC)

Done -- Asked42 (talk) 16:46, 18 December 2025 (UTC)

AI assistance mode

@Michael.C.Wright, @Gryllida: I have added a new feature, AI assistance mode. You can activate the AI assisted review in the "Review Details and Tools" section. Documentation on how to enable the mode is available there. It uses Gemini's API because they offer free tier usage for integration into the script. There is still room for improvement in this new feature. If you have any objections to it please let me know, including whether the AI feature should be disabled or not.

And yes, Happy New Year. -- Asked42 (talk) 16:58, 31 December 2025 (UTC)

Looking forward to trying this feature! I'll give it a go soon. Thanks again for providing useful tools!Michael.C.Wright (^Talk/_Reviewer) 16:27, 1 January 2026 (UTC)

I tested this today but didn't use the results (because I didn't finish the review). Some feedback/questions:

1. Is there a way to allow the reviewer to manually manipulate the prompt?
2. Could we make it more 'ala carte', i.e., allow for checking only Neutrality/Balance or only for style?
3. Should we include a statement at the end of an AI-generated review summary that it was an AI-generated response? Or maybe a simple check box for the reviewer to select that reports AI was used to assist with the review.

I use AI routinely with reviews and here is the process I use that works well for me. Maybe this could help us develop how we use AI better:

1. I start by feeding it just the lede and ask it to evaluate the 5Ws and state the focal event. I then evaluate its findings.
2. I will use it to reword sentences for clarity, conciseness, and/or distance from source.
3. Once I've verified all statements against the listed sources, I'll feed it the entire article and ask it to evaluate it for only neutrality and balance, paying close attention to attribution.
4. Lastly I'll ask it to evaluate the full article against the inverted pyramid style and suggest a reordering of paragraphs, reporting to me in simple format such as 1 → 3, 3 → 2, 2 →1 for manual re-ordering (so AI doesn't inject hallucinations I then fail to detect).
5. At no point do I trust AI's representation of facts. I have not found it to be accurate enough yet with providing only facts supported by the listed sources and it still has occasional problems with nuance (just as any human editor would).

Again, this is just how I use it and I find it helps. I don't think it makes the review process any faster. But I feel it's better at evaluating NPOV and is certainly a better wordsmith than I am for paraphrasing, re-phrasing, etc.Michael.C.Wright (^Talk/_Reviewer) 17:25, 2 January 2026 (UTC)

Thanks for the feedback.

For 1. Currently, there isn't a way to provide a custom prompt, but I can add support for reviewers to write their own prompt.

For 2. That is certainly possible. I can incorporate this into the AI Assistance Panel so it analyzes only specific, selected aspects of the article.

3. Maybe a checkbox can be included in the confirm review dialog to indicate whether AI was used or not. Should this checkbox always be included, or should it appear only when AI assistance is used?

Your approach would certainly help reduce hallucinations by the AI model. The AI-integrated approach in the script is intended to provide a quick, low-level analysis of the article. When using an AI model manually, you are likely using the latest version of that chatbot. However, one limitation of integrating AI into EzPR 2025 is that it currently uses an older version of Gemini, as newer versions are not allowed for integration using API.

So yes, direct usage of the AI model is generally the better approach. You can consider the AI assistance feature as an additional, optional tool. -- Asked42 (talk) 18:33, 2 January 2026 (UTC)

Update: AI assistance mode has been removed per this discussion. -- Asked42 (talk) 18:21, 9 January 2026 (UTC)

bug about not leaving a note when article was published

please see: Wikinews:Water_cooler/technical#ezpr2_bug_report_it_does_not_add_note_to_article_talk_when_article_is_published thanks Gryllida 04:33, 3 January 2026 (UTC)

Already resolved. -- Asked42 (talk) 18:22, 9 January 2026 (UTC)

Post publication changes review

@Michael.C.Wright, @Gryllida: I am thinking about adding a feature called "Pending Changes Review" or "Post-publication Changes Review." It would appear as an option in the user's action menu. When clicked, it would open a dialog displaying information such as who last checked the article, which editors have made changes since the last review, and so on. There would also be a diff button to compare the last reviewed version with the current version, along with options to either accept or reject the changes, with or without a comment. The reviewer could also choose to log this post-publication review on the talk page.

But before proceeding, I would like to know whether this would be useful or if the current approach is already sufficient. If it is useful, what additional features would you suggest to make post publication edit review more helpful? -- Asked42 (talk) 15:27, 19 February 2026 (UTC)

I like the idea—and thank you again for all of the development work you continue to do for the reviewers!

I think sighting post-pub changes is one of our weaker areas and this may help. If it can help untangle some of the rattiest ones, that'd be even better.

Netumbo Nandi-Ndaitwah sworn in as Namibia's first female president is a good example of what I mean by 'rattiest' or many edits to work through.

I also think having quick and easy access to sighting logs would help. They seem (to me) to be a bit hidden, or hard to find.

I'm happy to test for you.Michael.C.Wright (^Talk/_Reviewer) 15:40, 19 February 2026 (UTC)

I also just remembered that we have the ability to grant temporary reviewer status. You could be granted temp reviewer for testing and developing purposes, unless there are any objections from the community. Options available for the duration include 1, 3, and 6 months. Barring any objections, I think three months would be enough for testing purposes, and it could be extended if necessary.Michael.C.Wright (^Talk/_Reviewer) 16:36, 19 February 2026 (UTC)

Sure. I will try to create an initial plan to achieve that. If temporary rights are required during testing, I will request them. But, there are a few things I am not fully familiar with. For example if an article has, let's say, five pending changes, do you compare only the current version with the last sighted version, or do you review each change separately against the next one? This will help me understand the complexity of this. --Asked42 (talk) 13:48, 20 February 2026 (UTC)

Basically, what I am trying to say is that if you could share a general idea of how you (or any reviewer) usually review and either sight or reject pending changes, that would help me a lot. -- Asked42 (talk) 13:51, 20 February 2026 (UTC)

I do both; check against previous edit and check stable vs latest. Often I need to compare one edit to the previous to know exactly what was changed—especially with articles. When going through the templates recently though, I could short-cut it and only compare the stable version to the latest, pending changes version, because with templates, WN:Archive does not apply, and the content can be changed. I rarely just compare stable to latest pending with articles to ensure content didn't get substantively changed in violation of archive conventions.

I don't believe the system allows us to reject a single change in the middle of a bunch of other unapproved, non-actioned changes. It actually allows it, but takes no action until your next approve above the reject. Then it "rejects" all before and including the rejected one. And I believe contributors can get notified that their edits were reverted. So for that reason also, I will often work through and approve/reject serially or leapfrog only one to the next approved.

It also, for some reason, doesn't self-sight the last rejection. So counter-intuitively, the last rejection action must also be approved/sighted in an extra step. That is easy to forget and leave behind (and would be great if your tool does automatically).

Hopefully that all makes sense and hopefully I accurately described it. I won't be offended if it's now clear as mud and you need more detail for clarity. It may be better to understand my workflow here as leap-frogging the edit history, approving edits in order to reject the previous edits that they've reverted, which is most often the case—me approving counter-vandalism rather than rejecting the vandalism. It's also quicker in the system to approve than reject. There is an extra step when rejecting.

Also after explaining it, I think it will be very beneficial to you to have the reviewer flag temporarily so you can run through one to see how it works, unless there is good enough documentation somewhere for you. The documentation here isn't completely helpful. I've never seen an interface with "accuracy, depth, and readability" for example.Michael.C.Wright (^Talk/_Reviewer) 14:46, 20 February 2026 (UTC)

There are certain things I still find a bit confusing, specifically regarding "rejecting a revision." I will look for more documentation to see if I can find additional info.

After some analysis, I think the main change that could be introduced for post-publication edits would mostly be an improved interface for reviewers. Logic-wise, based on my understanding, there does not seem to be much additional functionality required beyond the current system. -- Asked42 (talk) 13:31, 21 February 2026 (UTC)

┌───────────────────┘
Agreed. If the tool provides a checkbox to auto-sight rejections, I think that alone will be a great benefit.

The rejection process is not clear to me, which is making it difficult for me to explain to you. The opaqueness of the rejection process is likely part of the reason so few reviewers work on the pending changes queue.

At any rate, below I'll try to clarify how I understand the rejection process.

I created Template:Test-flagged-revs to test with. I successfully rejected Change 3 without doing anything with Changes 1 & 2. It rejected all three, as expected.

Rejecting Change 3 also rejects Change 1 & 2

I then rejected Change 4, which at the time was the last, un-sighed change, and it marked it "automatically checked" and I'm not sure why. Normally, with articles, I would also need to sight that last rejection separately. Auto-sighting a rejection isn't an option (that I'm aware of).

I then made test changes 5 - 8 and tried to reject number 7 and I now receive the following internal error:

Internal error as a result of rejecting Change 7

I don't know if this is a bug or my misunderstanding of how the extension is supposed to be used. When I run into an issue like this in an article, I approve/reject changes one-at-a-time.

There is also the fact that the extension is considered by some to be "non-standard and outdated."[1] So maybe it isn't worth the time and should eventually be removed from en.WN. Michael.C.Wright (^Talk/_Reviewer) 15:13, 21 February 2026 (UTC)

i find the review queue hard to handle mostly because i do not have anyone to chat with, when i have a question. if there was a software that allows to generate a table somewhere with list of nicks and their availability, and i could read it without needing to do time zone conversions to my time zone, it would help me personally. Gryllida 22:16, 20 February 2026 (UTC)

i found it hard to follow. would that be for one article or pending review edits queue for one article? Gryllida 22:09, 20 February 2026 (UTC)

The plan was for each article individually, not for a review queue. This is because there is already a queue available for pending changes at Special:PendingChanges. I do not think there is much improvement I could develop for the existing queue, nor do I think a new one is needed beyond the current one. -- Asked42 (talk) 13:27, 21 February 2026 (UTC)

Hi Asked42, I was confused how the tool would improve upon current software, which shows a diff from latest sighted revision. Based on content, I either approve, reject, or make an edit and then self sight it (i.e. that could be formatting change, removing vandalism, or otherwise reverting something that was done after the page was last sighted). That I thought was already as nicely done as possible. How would the new tool improve this, other than display itemised edit history with the UI for approving last revision (currently that is not shown)? Gryllida 18:40, 21 February 2026 (UTC)

@Gryllida: Of course before deciding what new or improved features can be introduced, I need to understand what limitations or problems exist in the current approach, if there are any at all. I have discussed an initial plan in the comment below; you can take a look at that.

However, I agree that the tool may not be able to introduce anything revolutionary or fundamentally different from the current system. If implemented, most of the changes would likely be related to interface and UI improvements rather than underlying logic or functionality. Would you like to suggest anything related to post-publication changes, or is there anything that could be improved within the scope of EzPR on that matter? -- Asked42 (talk) 18:50, 21 February 2026 (UTC)

Main limitation of this software. From my perspective. Is the lack of motivation of reviewers to work together. For two years I have been requesting real time communication. Nobody was interested. I am not sure what would help for this. Some software should request that if a reviewer opened wiki, then they should respond to a note 'I will be available for ... (20 minutes or 2 hours or whatever)' and forces that user to join a real time communication thing and stay there for that duration and every time there is a new message, destroy their computer speakers and screen with the most annoying notification possible. This was a feature in w:MSN chat in 2001, it popped up to top of all windows and rang. That was not an audio call, it was just 'buzz that user' button that popped up the text chat on top of their everything they have open, even if it was a video game. (I revised an article an hour ago, the reviewer was available for next hour or perhaps hour and a half, and spent time doing template documentation improvement and not the review of the changes that I made, for example; if there was a real time chat thing, a communication could have helped with further revisions and/or publishing.) Gryllida 20:30, 21 February 2026 (UTC)

"Main limitation of this software. From my perspective. Is the lack of motivation of reviewers to work together." - Well, EzPR does include a method for collaborative review, also called partial review. You can read more about it on the documentation page. But whether reviewers choose to collaborate is entirely up to them; the gadget itself cannot decide or influence that. (@Gryllida)

Regarding the software you suggested for live communication, that is an interesting idea, but unfortunately it is outside the scope of EzPR, or any gadget for that matter. What you are proposing involves using another platform, which is not something that can be implemented through a user script. I think this might be better discussed elsewhere, perhaps at the Water Cooler, as it is not directly related to this gadgets. -- Asked42 (talk) 06:21, 22 February 2026 (UTC)

i mean to have real time communication not partial review, although i can see how that feature would be awesome if somehow one reviewer did want to resume work of another :)

even push notifications would help, phab:T113125 -- Gryllida 06:47, 22 February 2026 (UTC)

meta:Community_Wishlist/W498 -- Gryllida 06:56, 22 February 2026 (UTC)

Thank you for the detailed explanation. It was very helpful in giving me a clearer idea. As I understand it, the following could potentially be done with this new feature (please correct me if I am wrong):

1. The tool would log all edits made after the last stable version of the article.
2. The reviewer could choose to compare each revision with its previous one.
3. The reviewer could serially mark each revision as stable (from oldest to newest), which would update the log accordingly.
4. The reviewer could mark any specific revision as the current version of the article: essentially meaning that all revisions after the selected one would be reverted.
5. If the reviewer does not want to go through all the edits one by one, they can simply accept the current version or revert it to the last stable version (as I understand, this can already be done).
6. If reverted, the reviewer could choose to automatically sight the last edit or keep it unsighted.
7. The reviewer could also choose to log/add the sighting actions on the talk page (I am not sure whether this should be the default or optional).

If there is anything I have missed, or if you would like to suggest improvements, please feel free to let me know. -- Asked42 (talk) 17:11, 21 February 2026 (UTC)

This looks good. There is also a separate log associated with FlaggedRevs; Special:Log/review. That's the log I mentioned "having quick and easy access to sighting logs would help." That may also pertain to your #1 above.

If the tool could provide an expandable—but collapsed by default—field containing the review logs, that might also be useful when working through a long list of changes that has been partially worked through already. Currently the logs aren't displayed in the normal Edit History and one has to view them separately; in a different browser tab, new browser page, etc.

For reference; the article "Explosion occurs near railway offices in Athens, Greece" currently has a list of edits to be sighted as well as previously approved or rejected edits along with a review log. You can see edit summaries in that log that aren't visible in the Edit History, such as "downcase" and "No content change."Michael.C.Wright (^Talk/_Reviewer) 17:44, 21 February 2026 (UTC)

Thanks for the suggestion and that specific example. I will integrate the review logs into the tool, along with the revision logs. However, since it seems there are not many significant issues with the current system, I think it would be better to first determine whether there is actually a problem that needs to be solved before investing resources into further development. The original plan when I proposed this was to make the pending changes review process more structured and easier to follow. -- Asked42 (talk) 18:56, 21 February 2026 (UTC)

it would be better to first determine whether there is actually a problem that needs to be solved before investing resources into further development.

The problems with the system I think are inherent to how it was designed and implemented and not something easily fixed. The primary problems, as I understand them are:

1. Rejections themselves require an additional approval
2. There doesn't seem to be an option to self-sight a rejection if it is the last in the list of pending changes
3. Edit summaries associated with Review logs are only included in the page's edit history if a rejection is involved.

I think #2 could be fixed with a JS interface, which could alleviate #1. Beyond that, my hopes for the tool were things like optional drop-down choices for edit summaries, logs readily available if needed, and the option of pushing automated talk page notices to make approvals/rejections more transparent to all contributors.

I think these features could help improve the process of working through pending changes, and might even make it more consistent between/among the reviewers with boilerplate summaries and talk-page notices that are PaG-informed. I'm less hopeful that they'd get more reviewers to actually do the work.Michael.C.Wright (^Talk/_Reviewer) 13:40, 22 February 2026 (UTC)

I understand. I will prepare an initial version and then you can test it. Later we can make further refinements. -- Asked42 (talk) 16:46, 22 February 2026 (UTC)

Maybe could help to notify user(s) involved that their edit(s) were or were not accepted and a comment; and show a timer of how many hour passed after article was published. Gryllida 21:37, 22 February 2026 (UTC)

By "notify" i mean actually leave them a talk page message and not only an Echo notification as then that could become a space for discussion. Gryllida 21:38, 22 February 2026 (UTC)

This notification feature is now available as well. Thanks for the suggestion. please check the message below. -- Asked42 (talk) 14:39, 26 February 2026 (UTC)

@Michael.C.Wright, @Gryllida: The feature for reviewing post-publication changes (PPCR) is now available. There are a few minor tweaks that may still need to be made, but it is ready for testing.

Please visit any published article that has pending changes. In the p-menu (the same place where you find the "Review" option), you will see a new option called "Post publication Review." Click on it to open the interface.

At the moment I am unable to fully test the sighting related functionality myself. I have performed some basic dummy tests but I cannot properly test actions related to sighting. If you could test it and check whether everything is working correctly, that would be very helpful.

Documentation for the interface is available here. If there is anything you would like to see added or improved in the interface please feel free to suggest it. -- Asked42 (talk) 14:38, 26 February 2026 (UTC)

It looks great! I did just tested it at Explosion occurs near railway offices in Athens, Greece. I thought I was approving to the most recent, unsighted change but it only approved the most recent and didn't approve the previous two (the archival edits), which is interesting and possibly handy (and could also have been something I did, not a bug). Without the tool, I am unaware that later edits could be approved without approving or rejecting previous edits.

The talk output may need to be addressed: Talk:Explosion occurs near railway offices in Athens, Greece#Post Publication Review Logs.

I am unfortunately stepping away from my computer now for most of the day and won't be available until possibly this evening, Pacific time.Michael.C.Wright (^Talk/_Reviewer) 15:49, 26 February 2026 (UTC)

Thanks for testing it out. Just to clarify, I assume you mean that if an article has three unsighted edits, then approving the most recent one should automatically sight the previous two as well. If that is the case, I will make the necessary change to handle that accordingly. -- Asked42 (talk) 16:34, 26 February 2026 (UTC)

Sorry, I should have been more clear. I was in a hurry to get out the door.

I seem to recall there being at least two options: approve/reject edits one at a time, or approve the latest version as-is. I 'assumed' by approving the last change, it would approve all previous as well. But it didn't. Which is interesting.

I should have checked to see if the archive template was visible when not logged in. It was a previous edit that was still pending (I think I manually sighted the two previous edits pertaining to the "archive" of that article).

At any rate, I'll give it another test and kick the tires a bit more.Michael.C.Wright (^Talk/_Reviewer) 23:34, 26 February 2026 (UTC)

I have made changes to automatically sight all previous edits if the current revision is approved. The related templates for logging on the talk page have also been created. -- Asked42 (talk) 17:14, 26 February 2026 (UTC)

Suggest automatically adding signature at end of these 'Post Publication Review Logs' entries on article talk, in case someone wants to 'Reply' to that message. Gryllida (talk) 21:38, 26 February 2026 (UTC)

Done: Thanks for pointing that out. --Asked42 (talk) 13:28, 27 February 2026 (UTC)

Three new features for EzPR25

Hello, @Michael.C.Wright, @Gryllida three new features have been added to EzPR2025:

1) Dynamic Marking: In the Sentence Marking (SM) section, there is now a new button called "Dynamic Marking". If enabled/clicked, the text will be highlighted using the inline comment gadget instead of the {{highlight}} template. The highlight will be attached to an inline comment/issue. If Dynamic Marking is turned off, the current method using the template will continue to be used. (You can learn more about the Dynamic Marking gadget; it can also work independently of EzPR2025.)

2) Easier Sentence Selection: For Sentence Marking, you no longer need to copy and paste text into the fields. Just open the SM section and select the text directly from the article. A "+" icon will appear next to the selected text. Click it, enter the issue, and done. (Know more about this)

3) New Post-Publication Tools Page: After publishing an article, visit the Post Publication Tools page.

You may already be familiar with Lead Manager, which is available there. A new tool called QuickData has also been added. QuickData allows reviewers to quickly create a Wikidata item for the article and add the sitelink automatically. Some statements are pre-filled to help create the item faster. (This tool is still very new and may not have been widely tested yet. Suggestions and improvements are welcome.)

If you try any of these features in your upcoming reviews, please let me know. I would also appreciate feedback on whether any changes or improvements are needed, or if there are tools you think are unnecessary and could be removed. -- Asked42 (talk) 08:17, 14 March 2026 (UTC)

This may feel overwhelming. Let me know if there is any confusion or if a simpler explanation is needed. -- Asked42 (talk) 08:18, 14 March 2026 (UTC)

Thank you for the update. For point 1, I assume if a user wants to see these inline comments, they will need to enable a gadget. Is this correct? For point 3 I did not use Wikidata, I do not know what to do with the QuickData (or why or what for). Gryllida (talk) 08:23, 14 March 2026 (UTC)

Currently, Dynamic Marking is enabled for all users. In my opinion, it should be a default gadget so that it is enabled by default and users do not have to enable it manually. If users have to enable it manually, the main purpose would be defeated. (many non tech savvy users may feel confused about enabling it manually.)

Regarding Wikidata, almost all Wikimedia sister projects create a Wikidata item for their articles or pages. This is a common practice across our sister projects. Wikidata allows translated articles to be linked through sitelinks between different language editions of Wikinews. There may also be other uses that I am not fully aware of. Many reviewers already maintain Wikidata items for published articles, so QuickData will help them with this task.

In my opinion, creating a Wikidata item for a published article is a solid step to include in the reviewers' workflow. -- Asked42 (talk) 13:24, 14 March 2026 (UTC)

Love the idea of QuickData and Dynamic Marking! I'll test as soon as I can do a review.

Thank you again for the continued tool support and development. Making the review and publish processes easier will hopefully equate to more active reviewers.

That post-pub tools page looks great too. The layout looks modern, clean, and easy to understand. 🙏Michael.C.Wright (^Talk/_Reviewer) 12:59, 14 March 2026 (UTC)

I tested QuickData earlier and it worked great. I used it to create d:Q138679940

This may be too difficult or overly complicated and is not a priority or a huge need; could the script search for a sisters link and if found, place the title of the link in the main subject (P921) property? For example: d:Q132777472 for Gunman enters Pennsylvania hospital ICU, kills officer before police fatally shoot him. This will associate the sister article with our article.

I'm not sure if we should also be doing the mirror of that; adding our article to the main subject of the sister article...I might ask around about that.Michael.C.Wright (^Talk/_Reviewer) 23:40, 15 March 2026 (UTC)

Just to make sure I understand correctly, you are suggesting:

• Scan the article for "Sister links" section, if available.
• If found, prefill the "Main subject(s)" field on the QuickData form.

Is that all, or is there anything I missed? --Asked42 (talk) 14:06, 18 March 2026 (UTC)

Yeah, that was it. Thanks again. I also like the button at the top of an article to put it under review. 👍 Lots of good stuff to help reviewers.Michael.C.Wright (^Talk/_Reviewer) 14:11, 18 March 2026 (UTC)

Done: If the mentioned item in the sister link template is available on Wikidata, it will be linked as a main subject. -- Asked42 (talk) 14:37, 19 March 2026 (UTC)

need 'nowiki' or template-link (tl) in drop-down option

In the Style review criteria, in the drop-down menu of Criteria specific comment, "Please use the {{source}} template properly.", the mention of the template should either be enclosed in the HTML 'nowiki' or using {{tl|source}} to avoid transclusion. I ran into this here for an example.

Thanks again for the tools and your support of them!Michael.C.Wright (^Talk/_Reviewer) 09:09, 6 April 2026 (UTC)