Wikipedia:Pure wiki deletion system
|This is a failed proposal.
|Discussion of this proposal last took place at this Village Pump Proposals thread.|
On this page, users are attempting to develop a specific, serious, detailed proposal for a more transparent, less bureaucratic, and more consensus-driven Wikipedia deletion system. Edit this page only if you are modifying or adding to the proposal, or improving its wording; otherwise:
- Go to the talk page to discuss the proposal or to suggest how it can be improved
- Go to meta:Talk:Strengths and weaknesses of the current deletion system to discuss the wider question of whether such reform is necessary or desirable
- 1 Background
- 2 Aim
- 3 General strategy
- 4 Details
- 4.1 What kinds of articles would PWD be applicable to?
- 4.2 Key changes
- 4.3 Code changes needed
- 4.4 Policy, guideline changes needed
- 4.5 Common objections and responses
- 4.5.1 Increase in Vandalism / Edit wars
- 4.5.2 Harder to notice / check if legit / more work for us
- 4.5.3 Sockpuppets / Inclusionists break the system
- 4.5.4 Material is never actually deleted
- 4.5.5 It's unnecessary because admins already provide copies of deleted pages upon request
- 4.5.6 Additional complaints
- 5 How to help
- 6 See also
The instigators of this proposal perceive several serious problems with the current system for deleting Wikipedia articles (see meta:Strengths and weaknesses of the current deletion system). They believe that problems arise because the centralized, bureaucratic system currently in use does not mesh well with the highly decentralized and devolved mechanisms of Wikipedia. Stated simply, they consider that we should have more faith in the wiki system.
The rapid growth and high quality of Wikipedia are testament to the efficiency and effectiveness of the decentralized, reversible, transparent "wiki" process. All article-related decisions in Wikipedia are made using this process, with one exception — deletion. It is unclear why this is the case.
To enable deletion decisions to be made using the same decentralized, reversible, transparent "wiki" process that is used successfully for every other article-related decision on Wikipedia.
The central pillar of the proposal is one small change in the software:
- Links to blank articles will appear the same as links to non-existent articles.
Blank pages will appear to all intents and purposes as if they didn't exist, except there will be a link to their history on the page.
Anybody will be able to delete any page, simply by clearing all the text from it. Similarly, anybody will be able to revive any page by entering new text into it or by reverting to an earlier version from before it was deleted.
A page that has just been deleted will be specially marked on the "recent changes" page and on users' watch-lists; links to it will stand out as an "edit link" (i.e. a redlink).
The current deletion feature will remain for the purposes of deleting pages when it is necessary to remove old versions of the page, such as copyright infringements. There would be a 'Special:Log/blanked' page logging deletion/blankings (as there is now for moves, deletions, account creation, etc.) which users could track to spot vandalism; this would be necessary both to avoid negating the utility of existing things like Special:Shortpages, and to allow users to track for vandalism by deletion of less-frequented pages.
What kinds of articles would PWD be applicable to?
- G1 (patent nonsense), G2 (test pages), G3 (pure vandalism and blatant hoaxes), G4 (recreation of a page that was deleted per a deletion discussion), G5 (creations by a banned or blocked user(s)), G8 (pages dependent on a non-existent or deleted page) although this CSD should probably be abolished, G11 (unambiguous advertising or promotion);
- A1 (no context), A2 (foreign language articles that exist on another Wikimedia project), A3 (no content), A5 (transwikied articles), A7 (no indication of importance (individuals, animals, organizations, web content), A9 (no indication of importance (musical recordings)), A10 (recently created article that duplicates an existing topic);
- R2 (redirects from the article namespace), R3 (implausible typos);
- C1 (unpopulated categories);
- U2 (nonexistent user);
- T2 (misrepresentation of policy), T3 (duplication and hardcoded instances); and
- P2 (underpopulated portal).
- It is expected that PWD would be inapplicable for:
- G6 (technical deletion), G7 (author requests deletion) although some authors might be content to just blank their content, G9 (office actions), G10 (pages that disparage or threaten their subject), G12 (unambiguous copyright infringement);
- C2 (renaming or merging) which should be handled through redirects;
- U1 (user request) although some authors might be content to just blank their content; U3 (non-free galleries); and
- P1 (any portal that would be subject to speedy deletion as an article).
- Blanked pages will not disappear from users' watch-lists or contribution histories. When a user blanks a page, this will stand out on "recent changes" and users' watch-lists as an edit link.
- Under the current deletion scheme, only administrators can examine the history of deleted articles and restore old versions. Under the new system, just like with blanked pages right now, anybody will be able to examine the full history of blanked articles, restore old versions, or write completely new articles.
- When a user follows a link to a blanked page, the page will behave the same way that deleted pages do now, except there will be an additional message, reading:
- "A former version of this article was deleted by [USER NAME/IP] on [DATE]. The reason given for deletion was [EDIT SUMMARY]. You may view the article's history, edit the last version, or type a new article into the white space below."
- There will be parser functions, #ifblank and #ifnotblank, to determine whether a page is blank or not.
Code changes needed
- The blanking or unblanking of a page would be logged, and listed on watchlists and recent changes like the other logs(the move log, the protection log, etc.)
- Blanked pages will not show up in searches using Special:Search, random pages using Special:Randompage, or the list of all pages using Special:Allpages.
- The URL prefix to the page will be listed in robots.txt, and the page will emit "noarchive", "noindex" and "nofollow" tags, to prevent caching in search engines.
- This is to prevent the engines from even indexing a blank page, which might otherwise have quite a high rank due to links.
- Currently, redlinks are not indexed by search engines because all links to them point to http://en.wikipedia.org/w/ type addresses, rather than http://en.wikipedia.org/wiki/ and all /w/ addresses are nofollow'ed.
These code changes have been proposed at MediaZilla:3843.
Policy, guideline changes needed
- A list of Criteria for Full Deletion, stating what sorts of content must be sent to AfD, so old revisions can be hidden from non-admins.
- Clarifying the dispute resolution process in regards to deletion (i.e. how to deal with deletion wars). Suggestions:
- Blankings should always be subject to the 3RR. However, both parties are encouraged to list the page on AfD for discussion before this becomes relevant.
- A certain standard for edit summaries should be declared, and blankings with edit summaries not meeting that standard should be revertable as simple vandalism.
Common objections and responses
Increase in Vandalism / Edit wars
- blanking is already possible. lame editors cause edit wars, not policy
- Articles under dispute may become battlegrounds for edit wars. The frequency and scale of such edit wars might pose a significant problem.
- After first revert, move to talk page and/or list the article on AfD for discussion.
- Contested articles should never be deleted in this way in the first place, but discussed or AfD'd.
- This is already the case because anyone can edit articles, anyway.
- We would need protection used more often.
- Recreated pages would be similar in number to current AfD process. (no-increase)
- Not any more likely to require protection from high-speed edit war than current process.
- Anyone would be able to delete (an article!)
- This is already the case. Reckless deletion is still vandalism, be it a vowel or an article.
Harder to notice / check if legit / more work for us
- This relies on someone watching the page to notice that it's been deleted.
- Already the case for preventing vandalism.
- The proposal also includes a new Special: page showing blankings and unblankings.
- Most anti-vandal tools already have features for easily detecting page blankings.
- It would be hard to tell legitimate blanking from vandalism.
- No more so than legitimate edits from vandalism, most would be obvious.
- In many cases, it would be easy to tell legitimate deletions from vandalism simply by reading the article title, as is true today.
- Encourage well written and linked edit summaries.
- Don't we lose a central discussions point for all deletions.
- AfD, or similar, ain't going away. There will always be contested deletions.
- We don't have a centralized discussion points for other kinds of edits. We don't see a need for centralized discussion of deletion any more than other edits.
- A bot could be written to transclude discussions on talk pages.
- This makes it more difficult to see what a deletion is about.
- Blanked pages would include something like: "You may view the article's history, edit the last version, or"; the "edit the last version" link would provide a direct link to the deleted text.
- Perhaps the edit summary of the blanking edit should be included on blank pages.
- This makes it more difficult to track deletions and proposed deletions.
- A new Special:Log/blanking proposed to list all blanked pages chronologically.
- Deletions would appear clearly as red-links in Recent Changes
- A spuriously deleted article would be much easier to check for and revert, and would be much less offensive or dangerous than the sort of article content vandalism we see today.
- Due to AfD's size issues, this would actually make deletion information more readily available by allowing users to see all the proposed deletions in a simple, searchable list.
Sockpuppets / Inclusionists break the system
- Inclusionists can send every attempted PWDS to AfD. What's the benefit?
- The same inclusionists could contest all speedy deletions. This doesn't happen.
- At absolute worst, there will be the same number of articles in AfD as we have right now.
- Pages with sockpuppets support cannot be deleted.
- Why? If contested, pages supported by sockpuppets would go to AfD, as they do currently.
Material is never actually deleted
- The PWDS does not remove slanderous pages from the history. Someone could even link to them and make them look like part of Wikipedia.
- It is also true that much horrible, possibly illegal content is currently left in the history, as with nearly all cases of vandalism to existing pages. We have WP:OVERSIGHT for this.
- This is wrong. PWDS does not replace AfD, it provides an additional avenue. The "Delete" button given to administrators could still function in the same way as before, effectively making every administrator an oversighter.
It's unnecessary because admins already provide copies of deleted pages upon request
- It's a pain in the neck to have to request a page from an admin, and for the admin to provide it; therefore, in most situations, people won't make the request, and will do without.
- Generally, admins will only provide the most recent version of the page, but there may be important versions in the history. The user can't see them, and therefore won't know to ask for them.
- This proposal assumes that the devs will implement it.
- The proposal does not assume anyone will implement it. It simply lays out what a possible alternative would look like. Supporters of the proposal may choose to implement it, but no assumption of work by the devs is made.
- This still sounds like total mayhem.
- Proposed deletion is a similar but time-delayed "anyone can delete" system that has been working smoothly for some time now.
- When the idea of an encyclopedia that "anyone could edit" was proposed, everyone but a very few thought it would result in chaos and anarchy, too. Wikipedia works because of its open nature that relies on consensus. Based on the overwhelming success of this model for content edits, we think deletion should be handled the same way.
- This will just cause confusion.
- Confusion could come from two sources; the newness of the procedures, and the actual procedures themselves. The newness would wear off; if this is a worthwhile idea, it should not be rejected simply because it is new. If the claim is that the procedures themselves would cause confusion, it is necessary to specify how. Many possible specific claims of confusion are discussed above. If you have another one, please add it to the list.
- "because the wiki process is supposed to aid fast building, not fast destruction."
- The wiki process is supposed to support fast editing. The time saved by reducing the load on the deletion process would free many editing hours daily for building. Furthermore, any inappropriate destruction is and would be dealt with as vandalism, and has little to do with this proposal. We don't have a voting process every time we remove one paragraph from the current version of a page; why should this be different just because a page only has one paragraph on it?
How to help
- Help refine this proposal.
- Check the talk page for further discussion.
- Create and test development models of Experimental Deletion on real pages.