This is a failed proposal. Consensus for its implementation was not established within a reasonable period of time. If you want to revive discussion, please use the talk page or initiate a thread at the village pump.

This page in a nutshell: The version that contributors to an article consider their best version so far is flagged as a consensus version by a reviewer. The default view is not changed. The consensus version is used to see if an article has improved or deteriorated with time

When a certain version of an article satisfies most of it editors, they can request from a reviewer to flag the article as the consensus version. This is kept for reference in the revision history, and subsequent changes of the article can be easily compared to it. Editing and viewing of the article continues as normal for all users of Wikipedia, except there is a link to the consensus version on the top of the article.

Wikipedia is based on consensus. Even for controversial topics editors of contradictory points of view are able to produce remarkable compromises. However, reaching such a balance requires tremendous effort in finding the appropriate language.

Due to special circumstances, such as improvement drives, FA candidacy, mention in popular media, or simply by pure chance articles sometimes receive intense focus by a large number of editors. However, after the push to improve the article, it is often left to develop rather freely. Although obvious vandalism is reverted, the delicate balance that had previously been achieved can easily be lost. It is far from clear that the net result of well-meaning editors adding well-sourced and accurate statements really results in a better article! Therefore, sometimes our featured articles deteriorate and have to be delisted.

What we need is a method to systematically compare a series of changes with previous version, and incorporate improvements while discarding addition whose overall effect is detrimental to the article. This requires the consensus tagging to be a fairly quick process, so that article can be flagged at reasonable intervals.

For editors of an article to request the flagging of an article, they need demonstrate that there is consensus on the talk page. This is very similar to our Featured Article Candidacy discussion, but here there are no set criteria that the article must conform to. The question is only, whether the editors agree that it is so far the best version they have been able to produce.

This process would work as follows:

• An editor creates a section on the talk page with the title like "Proposing consensus version", where she provides a permanent link to the version that is to be flagged.

• Discussion continues for seven days, where all editors all welcome to give their views on that particular revision. When time is up, the discussion is closed, and that section should not be modified. During the discussion editors should indicate whether they are satisfied with the given revision: is it the best version of the article that they have seen so far. In particular, is it better than the previous consensus version (if there is one).

• If an editor is not satisfied with the proposed revision, good reasons should be given as to why it is not acceptable.

• If an editor wishes to discuss a revision that is substantially different, he should start a different thread. Note that this is already an indication that there is no consensus. However, it does make sense to suggest a revision with only trivial fixes (e.g., typos and formatting).

• When the discussion is over and the nominating editor feels that there is consensus for a particular version, she may request that a reviewer looks over the discussion.

Since this discussion is meant as part of the normal growth of an article, these consensus version nominations should not be widely advertised. Although anyone is welcome to comment, unlike other QA processes, it is primarily meant as means for the regular contributors, who are watching the page, to discuss and evaluate the progress of the article.

Very similar to our Featured Article process, there is consensus if most editors agree, and there are no reasonable objections. Examples of reasonable objections are of the form:

• The material "..." gives undue weight to a certain view, is not neutrally formulated, violates our policy on living people, etc.
• Compared to the previous version, the writing has deteriorated, the focus is lost, neutrality and balance is compromised, material based on unreliable sources have been added, etc.

Example of unreasonable objections are:

• The article lacks sources, the writing is terrible, etc.
• Coverage of "..." is missing / too shallow, section "..." is too long, it lacks a criticism section, etc.
• It is biased propaganda written by POV-pushing vandals.

When an editor believes there is consensus in the discussion, he can add a link to the talk page discussion on a Request for Flagging page. Reviewers, who have not edited the page in question, can evaluate the discussion. The role of the reviewers is only to determine consensus, and then flag the revision as indicated by the editors on the talk page. If there is not a consensus, she can attempt to help the editors in finding a way towards one, or if she finds that some objections are not reasonable, she should obviously justify her decision.

The reviewer have an important job in determining consensus, and will seriously impact the encyclopedia. They are elected in a similar way to administrators. Reviewing is a full-time task, and in accordance with the principle of the division of power, reviewers should not be admins. (Changing from admin to reviewer and vice versa requires going through the process again, and the other user right is removed).