This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

←

Archive 3

Archive 4

Archive 5

RevDel criteria far too broad

I feel a need to complain that I think the RevDel policy allows too many uses. I think its use should be restricted ONLY to edits that pose a clear and present threat that the Wikimedia Foundation could be subject to major legal action if an edit were merely reverted. The only thing I can think of that fits this criterion would be information that a government would likely claim was releasing government secrets inappropriately.

Kris Kobach said that the League of Women Voters and the American Civil Liberties Union were communist organizations. If someone had claimed that some other individual (not an organization) were a mass murderer or a pedophile, possibly using more explicit language, I don't think it would still justify deletion without a trace. It reminds me of Nikita Khrushchev's comment when he was ousted as Soviet Premier that under Stalin, "Not even a wet spot would have remained where we had been standing." We should allow the wet spot to remain, unless doing so could threaten the viability of the Wikimedia Foundation.

I'm writing now because an hour ago I got a notice that the article on "Nazism" was changed by User:Black Kite saying, "RD2: Grossly insulting, degrading, or offensive material". User:Black Kite does NOT even appear in Nazism: Revision history. No mere mortal is allowed to review that evaluation. That seems to me to be a blatant violation of the general open nature of Wikipedia. I don't think it should be allowed for material that is "insulting, degrading, or offensive", even if grossly so, because it's not subject to public review. See also my comments on Talk:Nazism#Reverting revisions without allowing others to see what was reverted. Thanks, DavidMCEddy (talk) 15:33, 14 February 2022 (UTC)

What about where a person puts something in when it is so outrageous / nonsense that they certainly know it will be deleted, they put it in only because they want it to be in the edit summary and history? North8000 (talk) 16:18, 14 February 2022 (UTC)

@DavidMCEddy: it's a waste of time even discussing this. It's calling not for a change in RevDel but in WP:OVERSIGHT as it is much more restrictive than the criteria there. Even copyright violations don't get suppressed. And of course it would allow 10 year old children to post their age and possibly become prey. This won't get traction. Doug Weller talk 16:48, 14 February 2022 (UTC)

The revdel criterion exists in response to extensive abuse over the year by attention-seeking trolls and long-term abusers of the most virulent kind who try to permanently enshrine gross abuse in Wikipedia revisions, that has no value to the encyclopedia, and which simply aggrandizes the abuser. It is used on a daily basis to deny recognition to to abuse, and has been completely uncontroversial. I have commonly used it to delete threats, personal abuse and harassment aimed at other Wikipedia editors by WMF-banned users. It can be reviewed by other administrators for appropriateness. The WMF has made it clear that they will not agree to non-administrators having the ability to view deleted content. As for suppression/oversight, that is reserved for certain very narrowly-defined edits, usually involving personal information, that is restricted even from view by administrators. And that is in turn subject to review by arbitrators and other oversighters. Acroterion (talk) 04:22, 15 February 2022 (UTC)

And without question, unsourced or poorly sourced accusations of criminal activity are subject to revision deletion under the BLP policy - it's compulsory. That's not open to any kind of debate. I strongly advise you to read WP:BLP for the reasons why. Acroterion (talk) 04:27, 15 February 2022 (UTC)

Thanks for the reply.

I asked, because I got a notice of a change to Nazism, but the change was hidden so the only thing one could see without admin privileges was that an edit had been reverted. I could not even see who had decided to hide the reversion. (WP:BLP does not apply in this case.)

I complained on the Talk page and was told this is standard Wikipedia policy for "Grossly insulting, degrading, or offensive material" and that the text reverted was "Heil hitler" repeated a dozen times.

That change seems to me to be offensive, stupid, silly. However, I saw no evidence in that article of "extensive abuse over the year by attention-seeking trolls and long-term abusers ... who try to permanently enshrine gross abuse in Wikipedia revisions".

Was the editor in that case making other more grossly offensive edits in other articles, which I couldn't see? And doing so in ways that could not be controlled by blocking certain IP addresses, e.g., as discussed in Wikipedia:Congressional staffer edits?

Thanks, DavidMCEddy (talk) 04:52, 15 February 2022 (UTC)

@DavidMCEddy: you aren't responding to the points raised by us so I assume you no longer want to see RevDel revised the way you suggested. And yes, we can block after the fact, which still means the edits need to be dealt with. Doug Weller talk 08:11, 15 February 2022 (UTC)

Revdel messing with page preview

Hi! Recently the page Sayani Gupta had been vandalized and revdel. Afterwards, the page preview read out the revdel'd content (a repeated death threat), which was fixed after ScottishFinnishRadish made a minor edit to the article which updated the page preview. How should this issue be reported to mediawiki, especially since I can't (nor wish) to link the revdel'd versions that caused the bug. A. C. Santacruz ⁂ Please ping me! 18:54, 18 February 2022 (UTC)

Or, at least, has anyone seen a similar bug after revdel was used? Not sure what caused the issue. I assume it was something caching related, and if I didn't see a quick copyedit to make I would have tried a dummy edit. ScottishFinnishRadish (talk) 18:57, 18 February 2022 (UTC)

Deletion of intervening edits

Between the addition and the removal of an offending material there may come many intervening edits, and if revision deletion is used, then they will get deleted as well.

How much collateral damage is it acceptable to have in the way of intervening edits? There are obviously several factors at play, and the choice of whether to delete or not has commonly been described as a balancing act. The existing guidelines and guides imply that proportionately little such damage is tolerated (WP:CPAA: so long as the infringing text is removed from the public face of the article, it may not need to be removed/deleted permanently; WP:RD: RevisionDelete is mainly intended for simple use and fairly recent material...). However, from the recent discussions (see #RD1, attribution and intervening edits, the links there and the follow-ups) it appears that it has now become commonplace for the removal of small-scale copyvios to lead to the deletion of large chunks of article history.

At this subpage there are brief statistics on the number of revisions deleted in the 1,000 most recent instances of WP:RD1 use spanning the last 23 days. The median number is 2, so most deletions likely don't affect intervening edits. However, there are still over 200 instances where 10 or more revisions were deleted, 20 cases with more than 50, and 7 deletions that each removed more than 100 revisions.

A large number of deleted revisions doesn't always equate to a large number of deleted intervening edits (for example, the contributor of the offending text may then fiddle with it using many edits in quick succession). Still, as evident from the list, there are articles where the deletion has affected almost the entire history, sometimes going back a decade.

Is this really OK? Why is it so common? Where exactly do we draw the line? – Uanfala (talk) 02:56, 18 February 2022 (UTC)

I suggest moving this to #RD1, attribution and intervening edits, maybe as a new subsection. I see this as a continuation of that discussion, and readers would not have to jump around. Thank you for keeping it separate from #"Attribution" in RD1 and the RfC. Flatscan (talk) 05:28, 18 February 2022 (UTC)

I infer a recent development from it has now become commonplace. I sampled 500 revision deletions starting 2012-01-01 (Ctrl-F rd1: 88), 2017-01-01 (59), and 2022-01-01 (82). Aside from many "Orphaned non-free file(s) deleted per F5" appearing between 2012 and 2017 and compressing the date range, they look similar with varied revision counts. Flatscan (talk) 05:28, 18 February 2022 (UTC)

I haven't looked in detail, but your 2012 sample does indeed show usage comparable to today's. I only assumed the present state would be the result of a process of change because I imagined that when the guidelines were written they would have been representative of common practice at the time. – Uanfala (talk) 13:59, 23 February 2022 (UTC)

Vulnerable talkpage

I am becoming increasing uncomfortable with Talk:List of country calling codes. I have no idea why, but it attracts a continuous stream of IPs who think it is a place to get help with Facebook and similar. People regularly post emails and whatsapp numbers there. These are clearly not very tech literate people, and the concentration of such people and their information in one location seems exploitable. I have come to feel that there is a good case to simply revdel the entire talkpage history per 4.Oversightable information, or simply per IAR given the situation. Further, while drastic, I think it be justified to indefinitely semi the talkpage, as the relevant talkpage activity is far far outweighed by the stream of emails and phone numbers. I would be interested in the thoughts of others. Best, CMD (talk) 01:54, 29 April 2022 (UTC)

"Attribution" in RD1

Attribution does not require blame

I created WP:Attribution does not require blame as an essay. The licensing details are not in dispute here, but they have been raised regularly in similar discussions.

Blame is a user's precise contribution, the individual diff in the page history.

The Wikimedia Foundation's wmf:Terms of Use § 7. Licensing of Content b.iii, specifies a list of authors as a valid attribution method.
A list of authors does not include page content and cannot provide blame.
Therefore, the licensing requirements do not include blame.

Flatscan (talk) 05:24, 26 January 2022 (UTC)

This was confirmed by WMF Legal around the same time: Special:Diff/1068782406. Flatscan (talk) 05:35, 16 February 2022 (UTC)

Uanfala edited the essay to be more favorable to their interpretation, and I reverted. Special:PermanentLink/1078110743 is the version before their change. Flatscan (talk) 04:29, 18 April 2022 (UTC)

Meaning of "attribution" in RD1

As RD1 is the copyright criterion, its "attribution" should be interpreted consistent with copyright policy pages, including wmf:Terms of Use § 7. Licensing of Content, WP:Copyrights § Re-use of text (policy), and WP:Copying within Wikipedia (guideline). They use it in the context of the licensing requirements.
A recurring opinion is something along the lines of "Licensing attribution is self-evident, so the mention must mean something more." When WT:Revision deletion/Archive 1#Community consultation was being considered in September–October 2009, the attribution requirement was not widely understood among editors. WP:Copying within Wikipedia would be promoted to guideline around a month later (WT:Copying within Wikipedia/Archive 1#Guideline proposal). WT:Revision deletion/Archive 1#Some questions (also October 2009) and WT:Revision deletion/Archive 4#RD1 wording (2014) clarify that licensing is the sole concern.
Copyright cleanup practice considered a list of authors as sufficient when WP:Selective deletion was used, prior to revision deletion. If RD1 were intended to restrict deletions and keep more revisions visible, there has been no change to reflect that.

Flatscan (talk) 05:26, 26 January 2022 (UTC)

I have always understood that bit to be referring to the attribution required for licensing purposes, which is just the name of the editor. I think we should really get rid of that phrase. The only way revdel could pose a problem for licensing purposes is if someone hid the username for an edit which wasn't reverted, and it's very unlikely that RD1 would be used to hide a username. And if there really is a licensing problem with using revdel then it would apply to all the criteria and not just RD1. Hut 8.5 17:49, 26 January 2022 (UTC)

Thank you to Flatscan for his clear and thorough statement of the position. I endorse his view entirely.—S Marshall T/C 02:02, 28 January 2022 (UTC)
I linked this discussion from WP:Administrators' noticeboard#Revdel on Himachal Pradesh. I had mentioned my intent there before creating these sub-sections. Flatscan (talk) 05:20, 31 January 2022 (UTC)
Noting that I was explicitly told that WMF Legal's perspective is that attribution in form of a list of names in the pages history is suitable. Moneytrees🏝️^{Talk/CCI guide} 05:41, 31 January 2022 (UTC)
Like Hut 8.5, I've always found that sentence (If redacting a revision would remove any contributor's attribution, this criterion cannot be used.) to be strange and confusing in the context of RD1, and never really applicable to the way the redactions are performed. DanCherek (talk) 16:35, 31 January 2022 (UTC)
Highly theoretical idea for creating a case in which redacting a revision would actually remove contributors' attribution: Revision 1 is fine, revision 2 contains a copyvio, revision 3 adds desirable text copied from a compatibly licensed source together with in-text attribution, revision 4 removes the in-text-attribution and the copyvio ("cleanup"). ~ ToBeFree (talk) 21:02, 31 January 2022 (UTC)
I think it's clear by now that the word "attribution" should be dropped: it's ambiguous in this context and evidently a source of confusion and disagreement. However, the underlying principle – don't use revdel for small-scale infractions if that's going to disrupt the page history – should still be clearly stated, possibly in a way that applies to all RD criteria. There's already some exposition in the sections Wikipedia:Revision deletion#Misuse and Wikipedia:Revision deletion#Large-scale use, but that does appear to at least partly presuppose the context of revdel use on discussion pages. There's still good reason why this will need to be explicitly mentioned within the text for RD1: the principle is far more relevant there than for other criteria. Copyvios on average take longer to detect than say, gross incivility or offensive material, and so there's a higher chance of intervening edits accruing in the meantime; also, the need to preserve visible page histories is much stronger for articles (where copyvios tend to occur) than for other pages (where other revdel criteria are more likely to be used). – Uanfala (talk) 15:06, 10 February 2022 (UTC)
I split this from your section to focus on the "attribution" wording. Flatscan (talk) 05:22, 15 February 2022 (UTC)

What is your definition of "attribution"? Is it individual diffs (blame)? Flatscan (talk) 05:26, 18 February 2022 (UTC)
I plan to create a RfC to propose removing the clause and sentence that mention "attribution". Flatscan (talk) 05:22, 15 February 2022 (UTC)
And replace it with what? – Uanfala (talk) 14:38, 15 February 2022 (UTC)
No replacement, just removal. Flatscan (talk) 05:29, 16 February 2022 (UTC)

Historical RD1 samples

I took a few samples of revision deletions from last month, 5 years ago, and 10 years ago. I chose 10 revisions as a cutoff to approximate deletions that were likely to include non-infringing contributors. My findings are consistent with my interpretation of policy and understanding of practice above. To state it differently: assuming my samples are representative, the restrictive interpretation has been violated almost every day for over 10 years.

Revision deletion samples, `limit=500`
Starting	Through	Ctrl-F rd1	Notes
2012-01-01	2012-01-20 12:26	88	RD1 with >=10 revisions most days
2017-01-01	2017-01-03 14:58	59	RD1 with >=10 revisions each day; +many "Orphaned non-free file(s) deleted per F5"
2022-01-01	2022-01-04 02:03	82	RD1 with >=10 revisions each day, except 2022-01-04 which had only 2 hours

I started with 2012-01-01 because it is just over 10 years ago. Revision deletion was enabled for admins in May 2010. I suspect that there was some lag for procedures to develop and administrators to onboard. For example, ((Copyvio-revdel)) was moved to Template space in November 2010. Flatscan (talk) 05:47, 23 February 2022 (UTC)

RfC: Remove "attribution" clauses from RD1

Consensus for 2A. There's a consensus that the current language is confusing and not required. Both Option 1 and 2A had lots of support but since 2A also helps assuage concerns about attribution, that's the language I find the most consensus for. Galobtter (pingó mió) 02:03, 16 May 2022 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Should the clause and sentence that mention "attribution" be removed from the RD1 criterion? Flatscan (talk) 05:21, 17 February 2022 (UTC)

Option 1

Blatant violations of the copyright policy ~~that can be redacted without removing attribution to non-infringing contributors~~. ~~If redacting a revision would remove any contributor's attribution, this criterion cannot be used.~~ Best practices for copyrighted text removal can be found at WP:Copyright problems and should take precedence over this criterion.

Option 2A – 2 proposed by MLauba and added by Flatscan at 05:24, 26 February 2022 (UTC); 2A revised by Flatscan at 05:18, 27 February 2022 (UTC)

Blatant violations of the copyright policy ~~that can be redacted without removing attribution to non-infringing contributors~~. ~~If redacting a revision would remove any contributor's attribution, this criterion cannot be used.~~ Best practices for copyrighted text removal can be found at WP:Copyright problems and should take precedence over this criterion. Username must not be hidden under RD1.

Rationale (RD1)

This change is a simplification that will have no material impact on policy or practice. The text being considered is:

Overly complicated In practice, it boils down to "Deselect the Delete editor's username or IP option so the username remains visible." When RD1 was drafted in September–October 2009, revision deletion and the attribution requirement were not widely understood. Since then, administrators have gained years of hands-on experience with revision deletion, and WP:Copying within Wikipedia was promoted to a guideline and continues to be cited often.
Redundant The Notes on use and Changing visibility settings sections advise against hiding usernames and mention the license/copyright considerations. Admins are aware that usernames should not be hidden unless specifically needed. Note that the requirements apply to all revision deletion, not only RD1.
Misinterpreted Readers are invited to speculate, "It must mean something more, otherwise why is it there?"
1. "Wikipedia's licensing requires that individual diffs must be visible." This interpretation was debunked by WMF Legal a few weeks ago. Also see #Attribution does not require blame above. #RfC on Change RD1-wording, a similar proposal from 2017, was opposed on this basis.
2. "'Attribution' means something other than licensing attribution." As RD1 is the copyright criterion, its "attribution" should be interpreted consistent with the copyright policy pages, which all use it in the context of the licensing requirements. Also see #Meaning of "attribution" in RD1 above.

Flatscan (talk) 05:21, 17 February 2022 (UTC)

Poll (RD1)

Support removal per Flatscan's rationale and my comment in the section above. That part has always seemed out-of-place and confusing to me because the redaction of copyright violations only involves hiding the revision content itself, not the contributor's username. DanCherek (talk) 06:11, 17 February 2022 (UTC)
Addendum: I support option 1 only and oppose option 2 per my comment below: Edit summaries are occasionally hidden under RD1. This occurs most commonly if someone creates (or replaces) a page with copyvio and they don't use an edit summary, so the edit summary defaults to "←Created page with 'blah blah copyvio...". That kind of revision deletion doesn't happen too often, but it shouldn't be prohibited. This is a recent example of a valid copyright-related redaction of an edit summary. I don't believe it would be an improvement to label the redaction of copyvio in an edit summary as a policy violation. DanCherek (talk) 17:16, 26 February 2022 (UTC) Now that it has been revised, I support either option 1 or 2A. DanCherek (talk) 05:26, 27 February 2022 (UTC)
Support removal this clause is pointless and misleading. All that is required for attribution purposes is the list of users who contributed to the page (not the content of the edits), and it's extremely unlikely that RD1 would be used to hide a username. Having this clause in RD1 and RD1 only implies that RD1 is somehow special in this regard, which isn't true - any licensing issues would apply equally to all the criteria. There are other sections of the policy which advise on things like large scale use. Hut 8.5 08:45, 17 February 2022 (UTC)

I don't think option 2 is a good idea because edit summaries sometimes contain copyvio, but it's better than the current state. Hut 8.5 17:18, 26 February 2022 (UTC)

Also support 2A. Hut 8.5 10:40, 27 February 2022 (UTC)

Support removal. We have both editor consensus and confirmation from WMF Legal that revisions can in all cases " ... be redacted without removing attribution to non-infringing contributors" and that redacting a revision does not " ... remove any contributor's attribution ...", so that sentence-and-a-half is both outdated and actively misleading. Justlettersandnumbers (talk) 12:02, 17 February 2022 (UTC)
Oppose per comments in the preceding discussion. Attribution is necessary not just for compliance with the particular version of the Creative Commons licence we're using, it's essential for any meaningful work here (that's the whole point of having page histories!). But even if "attribution" is understood in the narrow sense, then revision deletions, when there are intervening edits, can make proper attribution impossible. If you reuse content and opt to provide a list of contributors, then you're no longer able to do that because you don't know who those contributors are any more: you're forced to give credit to all users who have made edits in the deleted section of the history, even for edits that have been reverted as vandalism. However, I would support a thoughtful rewrite of that bit of the policy, but simply removing its only provision against misuse – that should be a clear no-no. – Uanfala (talk) 14:24, 17 February 2022 (UTC)
It's hardly the "only provision against misuse". #Large-scale use remains policy regardless of this RfC's outcome. DanCherek (talk) 14:40, 17 February 2022 (UTC)
Both #Large-scale use and #Misuse are relevant, but they're written in a way that addresses the context of use of the other RD criteria, so a reasonable person may conclude they don't apply to RD1. I don't believe these sections would have been written the way they were if RD1 didn't have this major limitation built in. If we're going to remove mention of this limitation then we're also going to have to rewrite those sections. – Uanfala (talk) 23:46, 17 February 2022 (UTC)
Oppose - removal. Support a rewrite that properly states the probable original intent: "if redaction removes any non-infringing contributions RD1 may not be used" It's an appropriate segue to mentioning WP:Copyright problems as the superseding text.--John Cline (talk) 10:48, 22 February 2022 (UTC)
If I find egregious copyvio in a single revision and remove it, I often take some time to manually tidy the article in the same edit (e.g., fixing malformed citations or curly quotes, removing excess whitespace). With your suggested change, I wouldn't even be able to request revision deletion of that one infringing revision because it would hide my own minor fixes. That seems far more extreme than what anyone else has advocated on this page or in the linked discussions. DanCherek (talk) 14:12, 22 February 2022 (UTC)
Considering the original language for RD1: "Blatant copyright violations. This does not include revisions on the same page that contain non-violating content that were posted in good faith by users not associated with the copyright violator." was certainly " advocated on this page".[1] And when it was changed, it was changed to the current language [2] per this RfC, now a linked discussion itself (and quite relevant, since the current language came directly out of it). Together, these show that my position isn't extreme at all. And better than advocacy on this page, or a relevant linked discussion, I'm going to ask MRG to comment here and believe that she will. Best regards.--John Cline (talk) 03:18, 23 February 2022 (UTC)
Please provide specific quotes/diffs that support your position. The foundation of your argument is unclear.
The current wording was originally proposed by User:MLauba. Note this clarification: A null edit crediting the contributors after the redaction would however be acceptable, wherever practical. This method is covered by WP:Copying within Wikipedia#List of authors, although it was not yet a guideline at the time. I don't see an issue in either of MRG's comments: Special:Diff/318470703, Special:Diff/318846041. I read them as concern that admins may hide usernames, which does remove attribution.

RD1 was discussed in 2014 by MLauba and MRG: WT:Revision deletion/Archive 4#RD1 wording, which affirmed that a list of authors is sufficient.
MLauba: To meet the attribution requirements, "X, Y and Z wrote this article" is sufficient.

MRG: Attribution is fully satisfied by the list

Your stance does not conform with actual practice. I posted above in a new subsection, #Historical RD1 samples. Quoting myself: assuming my samples are representative, the restrictive interpretation has been violated almost every day for over 10 years. The first page (default 50 actions) of Special:Log/Moonriddengirl (permanent link) includes multiple many-revision RD1 deletions that contradict your position.

Your notification of MRG feels non-neutral, which approaches WP:Canvassing#Campaigning.

If we continue, we should move this to a new subsection below #Discussion (RD1). Flatscan (talk) 05:50, 23 February 2022 (UTC)
I had completely forgotten that the language that trips up people these days was my own in the first place. Re-reading the old debates again, I can confirm that back in the day the verbiage "that can be redacted without removing attribution to non-infringing contributors. If redacting a revision would remove any contributor's attribution, this criterion cannot be used" came from a lack of understanding of the RevDel interface. The policy was written before the tool was finalized and the concerns we (MRG, myself and the then-active WP:CP admins) had were mooted once the interface went live in its final version. To be clear, I support the simplified language proposed here. MLauba ^(Talk) 18:03, 23 February 2022 (UTC)
Your rebuttal is compelling Flatscan. I think your comments about my message to MRG, however, are a bit off the mark; discussing it further would be off topic here so let's not. Regarding the foundation of my argument being unclear, I stated that the "probable original intent" (where attribution is now used) would be "non-infringing contributions" Linking to the "original language" clearly shows that the "original intent" was in fact "non-infringing contributions" then called: "non-violating content". Without debating the rightness or wrongness of that position, I did, and do think it was stated with sufficient clarity. And while the proposal may well carry the day, it will do so with me in dissent: I remain opposed. Best regards.--John Cline (talk) 00:21, 24 February 2022 (UTC)
Fair enough, I have been focused on the longstanding text. I will point out for other readers that the original wording was replaced (you provided the diff above) shortly before promotion to policy – months before revision deletion was enabled for admins – and never restored. Flatscan (talk) 05:29, 24 February 2022 (UTC)
Support removal, option 2A. The confusing language appears to have been my own, written before the RevDel interface was available and we all saw how easy it was to separate redacting the revision text without touching the edit summary nor the editor name. If we wanted to be more prescriptive, we could add something like Only the redaction of revision text is permissible under this criterion before pointing to the WP:CP guidance. MLauba ^(Talk) 18:26, 23 February 2022 (UTC)
Support removal. This clause was for a time when we had less understanding and judgement of how attribution works. We currently accept a list of contributors without necessarily seeing the diffs; admins accept revdels like the ones that are done at CCI and CP on a regular, where they can affect over half of the history. There is no attribution issue in the eyes of the people that regularly work to fix such issues with how RD1 is completed. Sennecaster (Chat) 18:54, 23 February 2022 (UTC)
Strong Oppose The assumption on your part is incorrect in the case of revision deletes. CC-BY-SA 3.0 states under section 4(c) The credit required by this Section 4(c) may be implemented in any reasonable manner; provided, however, that in the case of a Adaptation or Collection, at a minimum such credit will appear, if a credit for all contributing authors of the Adaptation or Collection appears, then as part of these credits and in a manner at least as prominent as the credits for the other contributing authors (emphasis my own). Each subsequent change in revision history falls squarely in the license's definition of "Adaptation". If a user contributes a great deal to a page, but in the same edit modifies a single line of a section that happens to contain a copyvio, it would be a violation of the license keep the contribution, but hide the facts of who added said content. --Elephanthunter (talk) 00:29, 25 February 2022 (UTC)
Your independent reading of the CC-BY-SA license invents a problem not recognized by wmf:Terms of Use § 7. Licensing of Content, WP:Copyrights § Re-use of text (policy), or WP:Copying within Wikipedia (guideline), nor by the perspective shared by WMF Legal a few weeks ago. WP:CWW has quoted that portion of the license since it was a draft in 2009. Revision deletion (usernames visible in the same place in the page history) affects credit less than the superseded WP:Selective deletion (list of authors in an edit summary or as a separate page) and copying between pages (list of authors or an additional hop to the source page's history). In case you mean credit for the deleted revisions specifically, my quick take is that they are hidden and not being "Distribute[d]". Regarding "hide the facts of who added said content", do you mean username (rarely hidden), the individual diff ("blame", not required by the license), or something else? Flatscan (talk) 05:23, 25 February 2022 (UTC)

MediaWiki:Wikimedia-copyrightwarning, the system message behind the text shown next to the "Publish" button when editing a page, says "You agree that a hyperlink or URL is sufficient attribution under the Creative Commons license." Interestingly, the Reply Tool's message (MediaWiki:Wikimedia-discussiontools-replywidget-terms-click) lacks this part. I have now asked WMF Legal whether adding it there as well would be beneficial. Anyway, as linked to by Flatscan above, the TOU say this as well (§7.b.i). ~ ToBeFree (talk) 19:32, 2 March 2022 (UTC)
Oppose If attribution is not preserved, then there is no point to the wiki in the first place, since we need attribution in order to make sure we don't violate the CC license. Also, attribution is important so we can reference changes to a page later, which is often helpful in resolving disputes. 2601:647:5800:1A1F:B59F:66D4:2C8D:EA30 (talk) 00:13, 26 February 2022 (UTC)
Attribution is preserved. This is clear in the proposal, and the newly-added Option 2 2A is even more precise. Flatscan (talk) 05:28, 26 February 2022 (UTC) – updated 2 to 2A Flatscan (talk) 05:23, 27 February 2022 (UTC)
I added a new Option 2 and relabeled the original proposal as Option 1. ~~I believe that they are close enough to count previous supports toward Option 2 also.~~ Flatscan (talk) 05:25, 26 February 2022 (UTC) – struck, being discussed below. Flatscan (talk) 05:48, 26 February 2022 (UTC)
Support removal, Option 1 or 2A, as proposer. I prefer 1 because I think the added sentence is unnecessary, but 2A is acceptable. Flatscan (talk) 05:27, 26 February 2022 (UTC) – updated 2 to 2A Flatscan (talk) 05:23, 27 February 2022 (UTC)
Support Option 2. I would've supported Option 1, but Option 2 is clearly the better of the two options (and both better than the current text). Option 2 makes it extremely clear what is and is not allowed to be revdel'd under RD1. –MJL ‐Talk‐^☖ 17:13, 26 February 2022 (UTC)
Would you confirm if you support the revised Option 2A? Flatscan (talk) 05:23, 27 February 2022 (UTC)
@Flatscan: I preferred the old text, but I'm still fine with it. –MJL ‐Talk‐^☖ 19:29, 2 March 2022 (UTC)
I revised Option 2 to 2A. Flatscan (talk) 05:22, 27 February 2022 (UTC)
Support removal Option 2A is also fine. It is clear that RD1 as used regularly is not removing copyright attribution as the username would not fall under RD1. That makes this sentence unnecessary and likely confusing. I prefer straight removal over option 2 since that too feels unnecessary since I can't see an admin comfortable to work with revision deletion removing the name of a user under RD1. --Trialpears (talk) 09:59, 2 March 2022 (UTC)
Support 2A. Although all the proposed texts are a clarification and an improvement, 2A is the very best. It accurately reflects the advice we've received from the Foundation, it fully complies with the terms of use, and it reduces the workload on sysops to clean up copyright violations.—S Marshall T/C 11:17, 2 March 2022 (UTC)
I Support the revised option 2A. — Diannaa (talk) 20:24, 3 March 2022 (UTC)
Strong support either option 1 or option 2A. I've been confused by this passage before and never seen a situation in which it has changed the course of an action taken based on a genuine issue. Attribution does not require blame, indeed. Many of the oppose reasons are simply incorrect. We should be much clearer that copyvios can be revdelled on sight. — Bilorv (talk) 09:08, 10 March 2022 (UTC)
Weak Oppose I'm not entirely convinced that the removal of this extra language would be a net positive here. If someone could assuage my doubts that attribution requirements cannot be violated as a result of this modification of CFRD1, then I would change my mind. As it stands, my understanding of the revdel tool is limited by not having access to the tool itself. Bearing that in mind, I understand it to be the case that revision deletions taking place over a long series of edits or for particularly old edits carry the risk of "collateral damage" so to speak, i.e. removal of attribution for content placed during intermediary edits. If an author's only contribution(s) fall into the range of this deletion, then wouldn't the attribution be destroyed? They would not appear in a potential list of authors compiled from the page history. Iff that's the case, then please interpret this as an oppose !vote. AlexEng^(TALK) 22:11, 14 March 2022 (UTC)
Hi AlexEng! I hope an example can clear some things up. If you look at this history of this draft, there is a little bit of "collateral damage" in terms of ability to see who wrote what – you can't see what IP 193.135.216.61 changed, and you can't technically see what Ingenuity changed (though the edit summary makes that obvious). However, the revision deletion did not remove attribution for any contributor, because the usernames have not been hidden, just the revision contents, and a list of contributors' usernames is sufficient, as explained above. While there are reasons not to use revision deletion on an excessive number of revisions, the rationale for my !vote is that these reasons are not actually related to the sentence under discussion. DanCherek (talk) 23:08, 14 March 2022 (UTC)
Thanks, DanCherek. Isn't there also an option to delete a revision along with the author information? Or does the tool retain authorship in all cases? AlexEng^(TALK) 23:28, 14 March 2022 (UTC)
Yes, administrators can choose to hide usernames too. This is typically done under the RD2 criterion when someone signs up with a vile and flagrantly offensive username. Because hiding the username does indeed remove attribution, Option 2A was proposed above to explicitly disallow doing that under RD1, to address some editors' concerns. DanCherek (talk) 23:36, 14 March 2022 (UTC)

Discussion (RD1)

Notifications: ((rfc|policy)) (diff), WP:Village pump (policy) (diff), WT:Copyright problems (diff), and WT:WikiProject Copyright Cleanup (diff). I did not list at WP:Centralized discussion. Flatscan (talk) 05:34, 17 February 2022 (UTC)
- The RfC expired on 19 March 2022 (diff 1, diff 2) Flatscan (talk) 04:33, 20 March 2022 (UTC)
WP:Centralized discussion (diff) --John Cline (talk) 07:58, 23 February 2022 (UTC)
- Archived on 11 March 2022 by Barkeep49 (diff 1, diff 2) Flatscan (talk) 04:30, 16 March 2022 (UTC)
Mentioned in WP:Administrators' newsletter/2022/3 (diff) Flatscan (talk) 05:36, 2 March 2022 (UTC)
The problem I have always had with If redacting a revision would remove any contributor's attribution, this criterion cannot be used is that it is not clear to me when "redacting a revision...removes any contributor's attribution". Removing the username does, yes, but most RD1 deletions are of revision text and I am never certain when doing so would remove any contributor's attribution. Jo-Jo Eumerus (talk) 11:24, 2 March 2022 (UTC)
If this is a request for clarification, the short answer is practically never, as attribution in the page text is not a recommended method. Also see previous comments by DanCherek and ToBeFree and background from MLauba. Flatscan (talk) 05:31, 5 March 2022 (UTC)

Option 2 (RD1)

Given some of the misgivings in the poll above, we could also consider rewording the criterion to become extremely specific, eg:
Blatant violations of the copyright policy Best practices for copyrighted text removal can be found at WP:Copyright problems and should take precedence over this criterion. Only revision text may be hidden under RD1.
This would eliminate all concerns about RD1 being used to tamper with attribution data (or edit summaries, for that matter). MLauba ^(Talk) 09:02, 25 February 2022 (UTC)
Thanks for proposing this again, it's a great idea. I added it as a new Option 2. Flatscan (talk) 05:25, 26 February 2022 (UTC)
Edit summaries are occasionally hidden under RD1. This occurs most commonly if someone creates (or replaces) a page with copyvio and they don't use an edit summary, so the edit summary defaults to "←Created page with 'blah blah copyvio...". That kind of revision deletion doesn't happen too often, but it shouldn't be prohibited. DanCherek (talk) 05:39, 26 February 2022 (UTC)
Good catch! It had crossed my mind, but I hoped it wouldn't be an issue. Flatscan (talk) 05:48, 26 February 2022 (UTC)
I should have reverted for more discussion as soon as I saw this objection. Forbidding existing practice was not my intention. I tried to address it by revising Option 2 to 2A. Flatscan (talk) 05:22, 27 February 2022 (UTC)
Thanks! DanCherek (talk) 05:26, 27 February 2022 (UTC)
I don't think this is an improvement: copyvios can occur outside of revision text, and restricting RD1 to revision text will only look like a good idea if there are intervening edits with substantive contributions, but if they are such intervening edits, then revdel ideally won't be used in the first place. – Uanfala (talk) 21:11, 26 February 2022 (UTC)
In my opinion, the only way to clarify revision deletion that might allay the valid concerns of others would require the addition of a new criterion for "complex copyright violations" (perhaps RD1.1) which should only be used when the copyright owner has filed a complaint or the infringement has been repeatedly reinstated from the page's editing history. This is because RD1 is for "blatant copyright violations" which is considerably different, and they each require different methods of handling, particularly when free-content edits overlie the infringement which definitely precludes the use of RD1 (consider CSD-G12 for an example of how the two are handled). The original RD1 language was likely an effort to prevent commingling the two types of infringement under one criterion but half-steps (short of creating a separate criterion) have failed and the current proposal will only make matters worse.--John Cline (talk) 22:56, 26 February 2022 (UTC)
An RD1.1 as described is hardly necessary. OTRS or Office actions have all the tools necessary to handle content holder complaints, and in fact are the only venues for content holders to address their complaints in the first place - if someone pops up on a talk page and claims that he owns inserted content, they will be directed to OTRS to clear it up. As for addressing repeat reinsertion, admins still have the rest of their toolset at their disposal, protection and if needed blocking. As to your speculations about the original RD1 language, well I wrote it, and Flatscan has kindly linked to the discussions around that further up here, and there's no need to speculate (spoiler alert, no, the original language was to avoid losing attribution, at a time we didn't know for sure how RD worked). MLauba ^(Talk) 10:55, 27 February 2022 (UTC)
Thank you for your reply MLauba. To clarify: if RD1 is for Blatant violations of the copyright policy, and: it is subordinate to the "best practices" outlined at Wp:Copyright problems where blatant copyright violations are described, and speedy deletion is prescribed when such "blatant infringement" is found to exist, under what circumstances would one be justified in disregarding the prescribed CSD-G12 deletion to perform RD1 revision deletion instead? I submit that no justifiable circumstances exist and, therefore, RD1 could never actually be used as written. While I am not proposing this now, because I do not wish to burden the RfC in progress with options developed after its start, I do suggest that the simplest way to resolve this paradox, in my opinion, would be to rename RD1: Gross violations of copyright policy. I further believe that it should delineate therein: two specifically narrow provisions where revision deletion could be used: 1.) Uncomplicated violations, and 2.) Complicated violations — which would potentially render as:
Gross violations of copyright policy
This criterion may be used for:
* Uncomplicated violations where removal of the infringing content by revision deletion does not affect the free content that would otherwise exist had the violation never occurred.
* Complicated violations after an investigation has been conducted, the article has been cleaned, and revision deletion has been subsequently requested.
Best practices for copyrighted text removal can be found at WP:Copyright problems and should take precedence over this criterion.
This is consistent with RD1's original language,[3] which is not the text written by you[4] and published without proper attribution[5] so the spoiler doesn't quite apply. For the record, I suppose it would be good if FT2 commented regarding the intent of RD1's original language. Best regards.--John Cline (talk) 08:19, 28 February 2022 (UTC)
For the sake of clarity, the nuance between G12 and RD1 is that G12 is a blunt tool used to delete an article wholesale (usually close to its creation) whereas RD1 is a fine tool meant to excise copyvios introduced later down the road, and, in general, where reverting to the last known good version would lose valid contributions outside of the copyvio content. It's a way to deal with the "fruit of a poisoned tree" problem that doesn't penalize the other contributors. As an aside, not sure where you are going with the "published without proper attribution", but I would have been grossly overreaching in directly implementing verbiage I proposed myself after the conclusion of a community discussion I was part of. I'll gladly own up to the fact that the current version is badly written and confusing, but I don't see how its adoption would have been improper. Happy to submit this under review should you feel otherwise, though. MLauba ^(Talk) 13:18, 28 February 2022 (UTC)
I'm good; thanks for your reply. Cheers.--John Cline (talk) 14:41, 28 February 2022 (UTC)
I revised Option 2 to 2A. Flatscan (talk) 05:22, 27 February 2022 (UTC)

Closing (RD1)

@Uanfala and John Cline: Will you accept me closing this discussion, or do you insist that I file at WP:Closure requests? Removal is supported by a significant majority. (The other opposers contributed only a few times and have low activity.) Flatscan (talk) 04:33, 20 March 2022 (UTC)

An uninvolved editor should close it. DanCherek (talk) 05:00, 20 March 2022 (UTC)

Thank you for asking me this, I appreciate the consideration and answer in two parts. First, I neither object to your closing the discussion nor will I challenge the outcome on that basis. I do, nevertheless, reserve the right to challenge the closure on merits that otherwise might come to be or to participate in a challenge that another might file. The second part is simply to say that while I do not object, I also, in this case, do not advise. In my opinion, the outcome is too important, the controversy: too unsettled, and the potential negativity: too costly for you to volunteer in this role. Time isn't short and nothing urgent is needful of haste. Ultimately, the decision is yours and no ill, either way, will come out of me. Best regards.--John Cline (talk) 10:07, 20 March 2022 (UTC)

I'd also like to suggest filing this for 3rd party closure. Let's not give rise to another round of discussion now or in the future due to procedural grounds. MLauba ^(Talk) 13:51, 20 March 2022 (UTC)S

Okay, I will list there with a link to #RfC: Remove "attribution" clauses from RD1 with no message beyond ((Initiated)) and my ~~~~ signature. Are there any other suggestions? Flatscan (talk) 04:22, 21 March 2022 (UTC)

Listed as WP:Closure requests#Wikipedia talk:Revision deletion#RfC: Remove "attribution" clauses from RD1 (diff) Flatscan (talk) 04:26, 23 March 2022 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Should REVDEL be mentioned as a possible remedy for DEADNAMING?

Your feedback would be appreciated at this discussion regarding WP:DEADNAMING and WP:REVDEL at Wikipedia talk:Manual of Style/Biography. Thanks, Mathglot (talk) 16:40, 21 June 2022 (UTC)

One or two lines on the time of edits

Is there some reason on the page history of a random Wikipedia page as to why some gets one line and others get two (see the recent history of this page for an example). Iggy (Swan) (Contribs) 22:15, 22 June 2022 (UTC)

One line (and a light gray strikethrough) indicates that the edit has been revision deleted, while two lines (and a black strikethrough) indicates that the edit has been suppressed or oversighted. DanCherek (talk) 22:20, 22 June 2022 (UTC)

As illustrated at WP:OS#Table. Mathglot (talk) 04:49, 23 June 2022 (UTC)

"where all changes will be reverted"

This policy currently says that one of the acceptable scenarios to RD a username but not the edit's content is where all changes will be reverted. Is that accurate? To me, that does not seem compliant with CC BY-SA's attribution requirement. CC BY-SA does not make a distinction between hosting the current version of a page and the old version of a page. We need to maintain attribution both for the current version of an article and for all previous versions. So merely reverting a substantive edit would not be enough to make it okay to RD the username but not the edit's content. The WMF would still be violating that user's copyright by hosting that old revision; and if someone chooses to restore that revision or incorporate it into some other page, they would be violating that user's copyright as well. -- Tamzin^{[cetacean needed]} (she|they|xe) 09:56, 3 October 2022 (UTC)

Surely if the only person entitled to claim the right to be attributed requests revision deletion of their username or IP address, they are asking to no longer be attributed? Ϣere SpielChequers 10:01, 3 October 2022 (UTC)

@WereSpielChequers: Self-requested username removals are covered as a separate exception. I'm talking about cases where an admin revdels someone's username (typically for being abusive), but not the content of a substantive edit they made. Which is rare—since accounts with abusive usernames usually make edits that either need to be revdelled too and/or fall below the threshold of originality—but it does happen, e.g. Special:Diff/1113763870 (CC Oshwah). -- Tamzin^{[cetacean needed]} (she|they|xe) 10:06, 3 October 2022 (UTC)

My sympathy is stretched thin by such pillocks. But if there is an occasion where someone wants to exercise their right to attribution and goes through the rename process to some name that meets our policies, I suspect that a null edit from their new account could resolve the issue. Don't we do something similar when importing content with a compatible licence? Ϣere SpielChequers 11:29, 3 October 2022 (UTC)

As far as CC BY-SA is concerned, they're entitled to attribution under the name they request. We, as a community, can of course say "We don't want to host your work if you're going to request attribution under some horrible name", but there's no "offensive pseudonym" exception to the license. I'm not really concerned here for the feelings of LTAs, and more concerned for the possibility that some LTA gets their username deleted and sees an opportunity to escalate from regular trolling to copyright trolling. In the worst-case scenario, such a person could sue an editor who innocently restores their content, as that restoration would now violate the attribution requirement.

I think the core of that is well-enough settled, which is why we have a general prohibition on redacting usernames but not edit content. My question is whether reverting an edit really suffices to mollify those attribution concerns. And my answer to my own question is that it doesn't. The way I see it, if you need to revdel a username, and the content of the edit is above the threshold of originality, you must both revert and revdel the content, or otherwise the result is a copyright violation. In a case where the content doesn't need to be reverted, well, we have to pick either hosting both content and username, or neither, but we don't get to pick and choose. -- Tamzin^{[cetacean needed]} (she|they|xe) 11:45, 3 October 2022 (UTC)

I agree with Tamzin's analysis: the edit must be reverted and the text of all intervening revisions should be revision deleted to prevent restoration. The items in the sentence could be expanded to a bulleted list to accommodate the added length. Flatscan (talk) 05:32, 15 November 2022 (UTC)

I made the proposed changes in two edits: Special:Diff/1134877408 and Special:Diff/1134877458. Flatscan (talk) 05:36, 21 January 2023 (UTC)