User:ChrisG/Approval mechanism

This is a suggested approval mechanism to meet the need to make Wikipeda a more reliable source and to support Wikipedia 1.0. It is a twist on the namespace workflow mechanism suggested by Magnus Manske here. It is a process firmly based on wiki principles, it uses a simplified version of the model used for featuring an article. Decisions about quality are based on human judgment that each article meets the minimum standards for approval for Wikipedia 1.0 (rather than the standards necessary for a featured article). If it is agreed that standards should be raised for a Wikipedia 2.0 then humans can amend their judgment appropriately during later approval processes.

Since coming up with this idea I discovered this [post] from Erik Moeller which anticipates the basic idea idea by over a year.

Aims

edit

The design aims to:

  • Have no effect on the wiki process for creating and developing articles.
  • Create no fork within Wikipedia
  • Identify specific versions of articles as being stable and hence give the reader the comfort that the flagged version has been reviewed as meeting certain minimum standards.
  • Identify specific versions of articles as suitable for Wikipedia 1.0.
  • Be scaleable: it needs to be able to approve a large percentage of Wikipedia within 3-6 months and then keep going. We have four hundred thousand articles now, and at current growth rates we will hit a million in 2006.
  • Mark versions of each article appropriately depending on the outcome of any approval process for that version. This gives a guide to quality of particular versions in the history.
  • Be simple to understand.
  • Be easy to manage.
  • Be democratic.
  • Avoid requiring any exclusive, 'expert' clubs.
  • Be relatively simple to code.
  • Be difficult to game.
  • Be an evolutionary approach rather than fundamental change in Wikipedia process.
  • Be acceptable to the majority of Wikipedians.

Overview of the process

edit

The process makes use of namespaces to create a workflow process. A copy of the recommended version of an article is moved through the system, with an attached discussion page and votes. The process is best illustrated visually:

 

Detailed process

edit

General

edit
  1. Only one version of a specific article can be in the approval process at one time. This is to prevent confusion and focus attention on a particular version.
  2. No amendment to the version passing through the approval process can take place.
  3. The purpose of the process is to flag versions of an article appropriately after the approval process and to create a collection of articles in a separate namespace which will make up Wikipedia 1.0. The article in the Standard namespace can of course be amended as normal.
  4. Only registered users can approve and reject articles. This is necessary because we may need to ban users from participating in the approval process if they refuse to obey the rules.
  5. Minimum standards for Wikipedia 1.0 will be established in an appropriate article. It is envisaged that Wikipedia 2.0 will have higher standards.
  6. Admins will have the facility to suspend the voting process if they feel gaming is taking place. Policy will have to be developed to support that process.

Initial recommendation

edit
  1. You cannot recommend an article for approval as stable if you have edited this article. This is to prevent gaming.
  2. It only takes a single recommendation to copy a specific version of an article to the Pending Approval namespace. The original version of the article in the Standard namespace is flagged as having a version pending approval.
  3. The talk page is not copied. The new talk page will contain the new discussion about the merits of approving the article. This will be important because it will ensure that people use the appropriate Wikipeda 1.0 criteria to judge the article. People may change their vote after hearing further argument.
  4. Any images used by the page will have to copied and protected to prevent image vandalism. The links within the copied article will have to be amended appropriately.

Pending Approval

edit
  1. In the Pending Approval namespace discussion and voting takes place.
  2. If a person rejects a version they have to give a supporting explanation. This comment should be automatically added to the discussion page for possible rebuttal.
  3. A person can change their vote providing the vote has not closed.
  4. If an article version receives more rejected votes than negative votes within the first day it should be automatically rejected. This rule is to ensure that users do not put forward inappropriate versions of articles.
  5. If a user attempts to sabotage the process by putting forward inappropriate versions of an article, then they will be banned from participating in the approval process for a specified amount of time.
  6. Approval of a stable version requires 80% support and at least 3 votes. A decision is made after one week. It is envisaged that most articles will pass without controversy.
  7. To prevent gaming there will be no fast track procedure. A week is required to ensure that an honest reviewer as a chance to see if something funny is going on.
  8. Approval moves the article to the Stable namespace. If there is an existing version of the article in the Stable namespace it should not be deleted in the process and the two versions should be merged into the same article. The original version of the article in the Standard namespace is flagged as ‘Stable’.
  9. Rejected articles are moved to the Rejected namespace with the new talk page. If there is an existing version of the article in the Rejected namespace it should not be deleted in the process and the two versions should be merged into the same article. The original version of the article in the Standard namespace is flagged as rejected.
  10. The status of the vote and list of votes should be autogenerated in the talk page to make voting and discussion a holistic process.

Wikipedia 1.0 process

edit
  1. The intention of this process is approve articles for distribution on DVD, CD or paper.
  2. A Wikipedia 1.0 group will be set up to approve stable articles for Wikipedia 1.0. This group of reviewers will have the duty of giving the articles a final check before moving the articles to the Wikipedia 1.0 namespace.
  3. The standards by which the Wikipedia 1.0 group apply to this process will be established in an appropriate article. There may be legal and production issues that impinge on their decisions.
  4. This group will have the ability to edit the article to serve the needs of publishing. However, they will have a limited remit:
    • Grammar
    • Spelling
    • Layout
  5. The Wikipedia 1.0 team will continue to use the existing talk page for any required discussion.
  6. Approval by the Wikipedia 1.0 group moves the article to the Wikipedia 1.0 namespace. If there is an existing version of the article in the Wikipedia 1.0 namespace it should not be deleted in the process and the two versions should be merged into the same article. The original version of the article in the stable namespace is flagged as ‘Wikipedia 1.0’.
  7. Rejection by the Wikipedia 1.0 team moves the article to the Unsuitable namespace with the new talk page. If there is an existing version of the article in the Unsuitable namespace it should not be deleted in the process and the two versions should be merged into the same article. The original version of the article in the standard namespace is flagged as ‘Stable but unsuitable’.

Variant

edit
  • It may be technically easier for the Wikipedia 1.0 process to take place in another wiki built for this purpose. It could copy stable versions from Wikipedia on a daily basis and then prepare them for the needs of production. They could use the same approval mechanism described for approving stable versions, but apply different standards.

Post approval

edit
  1. The decisions of these processes are final. If a version of an article is rejected or determined to be unsuitable then that version cannot be put forward again for approval. Only more modern versions of this article can be put forward for approval.
  2. Likewise if there is a stable version of an article older versions of an article cannot be put forward for approval.
  3. Once a version of the article is flagged as stable another version cannot be put forward until after a set time period (One month?). The initial aim is to provide as many articles as possible with a stable version.

Necessary supporting functionality

edit
  1. The status of a particular version should show up in the history of the an article in the standard namespace. The flag should also provide a link to the voting and discussion process.
  2. Watchlists should show an article is going through the approval process. Clicking a flag should provide a link to the voting and discussion process.
  3. If there is a stable version of a Wikipedia article then there should be a link from the latest version to the last stable version.
  4. The user configuration options should allow a preset of only seeing stable versions of articles, rather than the latest version. Editing should take you to the dynamic, latest wiki version

Nice to have supporting functionality

edit
  1. Searching Wikipedia should give a choice of the latest version or the last stable version.
  2. The system should record successful recommendations by user in percentage terms. The whole process would work far more smoothly if it was easy to identify that a user skillfully selects the most appropriate version of an article for approval. This would reduce needless debate.
  3. Watchlists should be amended so that you can automatically watch certain categories of articles. This would allow reviewers to keep up with articles in what they consider to be their fields.
  4. To inform your decision you should be able to see your vote and comment for a previous version of an article when you vote on a later version.

Implications

edit
  • This process replace the need for peer review.
  • This process could be used for featuring an article. It would just require the ability to recommend an article for featured article standard rather than Wikipedia 1.0. Of course, it might be better to retain the more flexible, manual method which allows quick changes to the article to remove issues.
    • As I am sure other people have stated elsewhere Wikipedia 1.0 should really be Wikipedia 0.1. The only articles deserving Wikipedia 1.0 standard are the featured articles.

Issues

edit

Any issues with this process?

Lack of experts

edit

The lack of experts. One of the most common complaints about Wikipedia is that anyone could edit a page, and get their ignorant opinion put out as fact. JesseW 02:12, 20 Nov 2004 (UTC))

This argument is based on the presumption that the wiki process is fundamentally broken. Encyclopedia Britannica and the media would not be taking Wikipedia so seriously if that was true. The German wikipedia indeed was found to be superior to mainstream encyclopedias in one case study, though we do not have a comparable English study, so we cannot make the same comfortable assumption. Obviously there are issues with some articles especially when an article reaches a high standard or it is controversial; but for 99% of articles the wiki process clearly seems to be working.
This mechanism seeks to stay as true as possible to that wiki process, while putting versions of articles through a quality check. As such it is a minimalistic approach which tries to apply the collective, anarchist, democratic approach that works so well in creating articles to the process of reviewing articles. We do not want to kill the goose that laid the golden egg. :ChrisG 12:19, 20 Nov 2004 (UTC)
Perhaps the best analogy for this approval mechanism would be to compare it to the existing process for featuring an article. That process produces first class articles and uses a democratic voting method. This approval mechanism is far simpler in that it does not allow an article to be altered during the process of voting and the standards will be much lower for Wikipedia 1.0. However, the issues raised about any rejected article will obviously inform the next recommendation. :ChrisG 13:42, 21 Nov 2004 (UTC)

Sock puppets/Groups force versions through to stable status

edit

This process just requires that they convince some of their friends(or sockpuppets) to join up and vote for their article, then it goes to stable. (I don't totally agree with this objection, but it seemed like it should be addressed. JesseW 02:12, 20 Nov 2004 (UTC))

This could of course happen since I have built in immediate approval with 20 votes in the current version. Amending proposal so that an article must spend a week in pending approval. This will give time for a honest reviewer to notice something funny is going on. :ChrisG 11:56, 20 Nov 2004 (UTC)

Cleanup required before voting

edit

Proposing an article puts it straight into voting, which I think is too early. I have my own proposal that freezes the page for cleanup before any acceptance stage. m.e. 09:22, 20 Nov 2004 (UTC)

The simple answer to this is the approval process needs to be scaleable and quick: consider the failure of Nupedia. Trying to clean up an article before voting it will inevitably slow the process. Even more seriously it also creates a fork within wikipedia: the altered version will not be the same version as version in the article history (standard namespace).
Where clean up and editing will probably be necessary is in the production process for Wikipedia 1.0. But that would be acceptable because that would be amending the original article for a specific purpose outside Wikipedia. Indeed it may be simpler to perform that process in another wiki. Proposal amended on this point :ChrisG 12:05, 20 Nov 2004 (UTC)

General comments

edit

Feedback appreciated. If you dislike the process could you add at least one issue to the issue log above.

  • Looks good to me. I suspect people may complain about the lack of experts, (see issues, above). JesseW 02:10, 20 Nov 2004 (UTC)
  • I've added this article to Category:Wikipedia 1.0 - check some of the other articles there, so as to avoid reinventing any more wheels than necessary - David Gerard 01:50, 27 Feb 2005 (UTC)