Wikipedia:Why stable versions

This page is currently inactive and is retained for historical reference.
Either the page is no longer relevant or consensus on its purpose has become unclear. To revive discussion, seek broader input via a forum such as the village pump.

For the latest developments on similar features, please see Wikipedia:Flagged revisions, Meta:Article validation feature and Meta:Reviewed article version.

About how stable versions are intended to work: Wikipedia:Stable versions (obsolete).

Premise

One goal of stable versions is to produce a Wikipedia 1.0 edition. Many people who consider Wikipedia a useful first-cut reference hesitate to fully trust the content of any article, because they believe Wikipedia to be unreliable. Given that our goal is to produce an encyclopedia, this is a major problem: reliable quality is among the most important requirements for an encyclopedia. As a wiki, Wikipedia is subject to sneaky vandalism, and any article may contain information that has not been verified. And, on the principle that a chain is only as strong as its weakest link, problems in some areas of Wikipedia have made all of Wikipedia suspect. Court cases have referenced Wikipedia, reinforcing the need for reliability. As things stand, scholars will place less trust in a Wikipedia article than one in a traditional encyclopedia. Although Wikipedia is used extensively as a reference, in its current state it lacks the respect given to traditional encyclopedias.

Stable versions of articles address this problem by establishing a rigorous process of quality assurance that traditional books and encyclopedias have been doing for a long time. Stable versions will help foster an environment of academic quality for serious researchers and to maintain them [1]. Protected, stable articles have been through quality assurance via meeting high standards and will maintain quality via being static. This process, however, does not disturb Wikipedia's philosophy of openness or its foundation to maintain neutral point of view and continual open improvement since there will always be an editable version.

Wikipedia is both a wiki and an encyclopedia. The community has embraced the former concept, but is sometimes in danger of forgetting the latter. Stable versioning is a critical feature of any encyclopaedia. The current wiki system has proven to be a successful means of generating excellent articles, and there is no reason to replace it. Stable versioning will only complement it. Now that Wikipedia has achieved widespread popularity it is time to address its issues of accuracy.

It can be seen that there is great demand for print and DVD editions of Wikipedia found in the help desk and mailing lists. Although they can be downloaded, stable versions provide a greater validity of the articles as a foundation for serious distribution and to be regarded as a reliable source.

Wikipedia should allow for an environment respectful to experts.

Similar proposals

Here are some similar or almost identical proposals found elsewhere:

Alternative systems

Ideas

Several possible implementations have been suggested; not all of these are mutually exclusive. More comprehensive ideas are on the talk page.

Protection

Stable articles could be protected, with their information not to be changed unless they are republished.
Alternatively, stable articles could be defined by permalinks to particular versions of articles in article space.
Templates in stable articles are to be substed. It is easier to see how this could be done in separate, protected articles than under the permalink approach.
Similarly, images in stable articles would be required to be copied into another namespace such as "Stable images" following the example and be protected. An orphaned image in such a namespace would indicate it is not used and be deleted. (Alternately, images could be "published" in their own right, much as images can now be featured in their own right.)

Location

Stable articles could be located in a separate namespace such as "Stable", "Published", or another appropriate name. For example, the article Mathematics could have a corresponding published article in Published:Mathematics, or Stable:Mathematics where the former is freely editable (including the risk of vandalism) but the latter is not. MediaWiki could then be configured to:
- Automatically protect all pages in the namespace.
- Allow only admins to create new pages in the namespace.
- Include "stable version" as a tab link above the editable version, as it now does with talk pages.
Another alternative would be to use subpages rather than a different namespace. For example, the article Mathematics could have a corresponding published article in Mathematics/Published or Mathematics/Stable
Another alternative would be to place the completed version in a non-wiki URL. For example, the stable version of the Mathematics article could be placed in http://en.wikipedia.org/publication/Mathematics instead of http://en.wikipedia.org/wiki/Published:Mathematics. They would no longer be wiki and should be appropriately named something else.
A revival of Nupedia or a separate site or fork could host the stable versions.
If there is only one Wikipedia certification process, there need be only one location. However, if there are multiple certification groups (e.g. individuals, teams, leagues) working within the MediaWiki software, there should be a mechanism for distinguishing which groups certified which versions of articles, which may imply multiple locations or a different architecture (such as flags on particular versions or tables denoting certification).

Interlinking

Interlinking between developing articles in the wiki and final revisions in the new system should be maintained.

Interface

There is a debate as to whether stable versions or wiki versions should be shown by default. The advantage in showing wiki versions by default is that it encourages the editing which has made Wikipedia successful. The choice between the stable or wiki version as the default is, in some degree, a choice of whether Wikipedia shall be distribution-centric or production-centric.
A possibly hybrid solution would be that if there is a stable version, it shall be the default with its wiki version hidden behind (available either on a tab or through the nav bar). If there is no stable version, then the wiki version is the default. This might make particular sense if the bar is set high for stable versions, constituting articles that are essentially viewed as complete. Although improvements can still continue to be freely made, designation of a stable version might constitute a consensus that further changes would not add any significant value, and a guideline could encourage Wikipedians focus on other articles as enough time has already spent on this one.
Reader/editor mode: stable versions by default, but even without logging in, someone could indicate a preference for seeing wiki versions.
Different URLs: if stable articles are stored in something other than the /wiki/ folder, people would choose which version they would browse by using the appropriate URL, and people who visit [language].wikipedia.org would be offered a choice of which version of the Main Page to visit. Articles which don't have a stable version could be mirrored in /stable/ or /published/ with a large prominent notice that it is a mirror of the editable version and thus potentially unreliable (or see point below).
Stable versions may need a completely new UI look, readily distinguishable from wiki pages.

Credibility

Encourage to identify editors or reviewers, particularly experts, with names and addresses to be accountable for their work.
User:DavidLevinson/Future, proposal for version certification process by individuals, clubs, or leagues.
- Wikitech-l: Feature Proposal: Certification (Wed Oct 30 12:02:28 UTC 2002) [2]
WikiProject Wikicite has a project idea of dedicating a separate datasource of full citation information (perhaps in a separate namespace) for every citation in the article with integration into the wiki syntax (see also Wikipedia talk:Forum for Encyclopedic Standards#Software assisted citation.3F).

Certification

There need not be any one single WP-wide process for certifying articles. The team at Good Articles is already using a process (acceptable to them, but maybe not to everyone) for certifying articles. The Featured Articles process is more rigorous. Different certification systems, used by different groups, may exist.

One possible certification process is

A version is arbitrarily nominated quoting the oldid to be voted upon and to reach consensus with.
Each version can be rated, marked, voted to be validated articles. These validated articles become candidates for a stable version. [3] Reviewing for stable versions and article validation are separate processes. [4].
Stable versions will be reintroduced to the working article Wikipedia namespace to prevent forking.

Demonstration

test site
old test site, by Magnus Manske

Objections

This proposed policy is a major structural change since Wikipedia's foundation and so common objections are likely to arise though there is also major support. Here are some succinct common objections (bold titles) and replies (text):

Anti-wiki

It is anti-wiki (because it is not wiki).

Wikis are useful but alone do not solve everything. Stable versions are warranted because we are trying to build an encyclopaedia, and stability is an important element of an encyclopaedia. However, the developmental versions are still wikis. This process maintains the spirit of Wikipedia while avoiding so many of the problems we currently face. The discontent about the un-wikiness is a sign of a problem we have: a focus on the supposed needs of the community rather than the needs of the final end user. An encyclopaedia or other reference isn't anything without an audience, and making the process work to the benefit of the end user, rather than ourselves and our wiki-feelings, ought to be the focus here.

Freezing

Freezing is bad.

There is more to stable versions than just freezing articles. It must be accompanied by quality check. WikiProject Spoken Wikipedia can benefit from static articles for their audio to exactly match a worthwhile article. It would give the Wikipedia:WikiReader project a much easier job in choosing a good version to be printed.

It is appreciated that wikis provide articles to be rapidly improved. Stable versions serve a different purpose for the ending stage of the article lifecycle process; to finalise the article for wide and official distribution.

Dynamism

It undermines dynamism.

Dynamism by itself can waste resources, and stability aids dynamism. Certain readers consider stability to be paramount over dynamism as stability is steadfast of its purpose and where dynamism can be instable. The current model is in flux at all times and, sadly, there is some evidence that once an article gets a certain quality it can start to lose that quality slowly. The new model could be compared to an arcade game. You can always play the game, but the game always tracks the high score. A lower score wouldn't erase the amazing record. Even if an article degrades, there is still a "good" version. An editor can at a later time step in and try to "out-do" the high-score. Another analogy is that Wikipedians often find themselves pushing the boulder halfway up the hill, and turning around only to have it roll most of the way back. These are not just simple vandalism that can be easily reverted but big ugly dump-edits that require major effort to clean up, fact-check, and integrate. One side benefit of a Stable Versions process is that it puts chocks behind the boulder every so often. Disruptions by well-meaning contributors are minimised since stable versions can be dedicated to maintaining high-quality articles at all times whereas the developmental wiki versions encourage riskier edits of boldness and experimentation.

Immediacy

Editors will become less interested because their edits are not immediate.

The perceived loss of instant gratification is important to take into account but stable versions will not replace the existing wiki system that has served us well. Stable versions need to exist but wikis will still remain the most prominent. Stable versions are secondary to wikis and are only to complement. For a solution of minimal disruption, people would hardly feel any change in Wikipedia has taken place because they would only just see an extra tab called "stable" if there is a stable version available. A stable version can be created and once it has been published, it can be forgotten about to focus on the wiki version. There may well be a separate view for final end users or readers that stable versions have more emphasis.

Users of candidate articles for a stable version might have more incentives to correct minor mistakes or to rewrite badly structured parts of an article if their changes are granted some kind of "protection", provided they make it into "stable".

Openness

It undermines openness because Wikipedia is an encyclopedia in which anyone can edit.

Stable versions are very much editable but not in the sense of wikis. Stable versions are atomic and to edit it is to replace it. The quick effect of editing wikis only applies to wikis: wiki means "quick", in Hawaiian. The word "edit" does not necessarily imply the effect is immediate. It is also suggested that openness (including the community) is a means to an end: to build a free encyclopaedia of high quality and wide distribution. [5]

Harder to fix

It is harder to fix small mistakes.

It is indeed harder and thus more importance is needed to be placed on publication of stable versions. They are to be relied upon after all. We have built articles with wikis, but without responsibility and accountability. The authors' reputation is put on the line just like in other non-wiki publications. Only important and transparent addenda and errata can update the article to avoid frivolous changes, because stable versions are not wikis. Otherwise, it can be republished again.

Founding principles

Our founding principles are eroded.

As the name implies, Wikipedia has a dual property of being a wiki, yet it is also an encyclopaedia. The wiki part has been well embraced favourably by its contributors. Let's not forget the encyclopaedia part, in the sense of being a more closed nature, is needed for the final end user. Wikipedia was founded from Nupedia which has a lot in common with stable versions.

Decision

Who decides it is good enough?

Consensus would be the most important component and that takes the bulk of the work in deciding. Experts would be very useful if we can come up with a proper means to identify them. The final responsibility and decision ultimately falls upon officials such as Administrators or Bureaucrats. Stable versions may fuel the intensity of edit wars, POV conflicts and rejection of "good" changes for the benefit of reaching a quality article of NPOV more quickly. Further decisions involve Wikipedia's dispute resolution processes.

Nominated articles are mandatorily placed on a reviewing phase akin to software beta testing called Stable Review Version. It is open to the community as to how the article is validated taking as long as possible. Once it has been decided, it becomes a Stable Release Version.

Featured articles

Featured Articles is adequate.

It is well known that featured articles lose their status. Making a stable version is useful to achieve a milestone to avoid wasting resources. Of course, articles will continue to have the option to be forever improved, and again marked with further milestones.

Furthermore, an insufficient number of articles are featured. There are many articles that are good and should be stable, even if they are not as interesting as featured articles.

Featured articles have standards that are too high. Stable versions are applicable to article with lower standards such as Wikipedia:Good articles.

Accurate as others

Despite having high-profile cases of inaccuracy, on the whole, Wikipedia is very accurate and all tests conducted on it by independent reviewers have shown that it is about as accurate as other encyclopaedias.

The likely reason that these articles remain accurate as it is, is due to the exhaustive effort in patrolling changes where the effort could be put to more productive use. Stable versions are about more than just being reliable at protecting against vandalism and other less obvious inaccurate content. The issue with the current system is that even if an article is trusted and reliable, it's simply impossible to expand and develop that article while keeping it completely trusted and reliable at each instant. All new content has to be refined and refactored, corrections have to be examined and accepted or reverted by interested parties, and so on. The Stable versions mechanism isn't a simple defensive measure like protection; it's a relatively low-cost way of creating a second space of articles that, at all times, maintains an exceptionally high standard of quality. Currently, although all articles tend to improve in the long run, you can't reasonably expect that every article will be as good or better tomorrow at 9pm than it is today at 9pm. The ability to take high-quality snapshots is important, for one thing, for successfully moving toward a print version of Wikipedia, but also to create an additional resource for people who have a specific requirement for a high degree of quality, such as researchers. It also allows us to counter some of the oldest arguments against Wikipedia: any temporary problems manifested in the working versions can be attributed to the fact that they are "just a work in progress", and the critic should look at the stable version before they start complaining.

Need to change

Is it really needed? There is no major threat that we would have to change.

The single most important factor in this proposal is the receipt of endorsement from Wikipedia that the article is accurate, otherwise it is no different to the wiki articles we currently have. People do not have full faith in Wikipedia. The loss is the potential for Wikipedia to be more reliable prior to implementation of this proposal and that many good contributors have already turned away because of what they perceive as an environment unfriendly to producing quality articles. The wiki will continue as it is where anyone can be free to be bold, alleviating the pressure to stable versions that address all those problems critics have been complaining about. Schools, for example, could make use of stable core topics, known to be vetted for quality and accuracy, in a way that some feel they cannot now make use of Wikipedia. And, with all due respect, this is far more important for well defined, scholarly topics like Napoleon I of France or Famine than it is of an article about a non-core topics like a Pokemon character or new topics like Spysheriff.

Growth

Any user reading the published version instead of the active version will be one (or probably 0.5 or less, but still something) less user to see (and therefore have a chance of editing) the current version, so to the degree this suggestion is a success and widely used, it will slow down Wikipedia's growth.

Wikipedia, particularly the English one, currently has a most phenomenal growth in the world and so has great affordibility to take risks slowing its growth. This proposal believes it can maintain or better that growth because Wikipedia would have been made so much more attractive to a new market due to the higher quality. Quantity of growth is important but the growth with and in quality needs to be taken into account as well as the improved efficiency on maintaining articles. If the model of having the wikis to be the default view rather than the stable versions is pursued, this minimal disruption while at the very least having stable versions to exist is reasonable to have.

A high traffic given to stable versions rather than the wiki versions is still beneficial to Wikipedia. When users see the Stable Review Version as the default, they can improve it by "testing" it, reviewing it for accuracies and corrections which is a different task to wiki editing but an improvement of Wikipedia in quality nonetheless.

Easily outdated

It is not worth pursuing because they are easily outdated. Our energy can be spent better in places other than in certifying articles.

Articles may well remain less up-to-date than the current wiki version. As long as they have properly been certified of their accuracy, this aspect still makes stable versions useful. The more properly they are certified with great quality assurance and control, the higher the trust readers will have with Wikipedia, and the more the certification is considered well spent. Wikipedia is expected to continue for many, many years to come and eventually we will have a large collection of stable versions to be realised in the long-term rather than the short-term.

An article can always be improved, but it has to take a temporary rest at some point for Wikipedia 1.0. It is suggested that there is an Average End Quality, where it is most appropriate for stable versioning. Stable versions do not replace the wiki versions and stable versioning is not an obligation that every Wikipedian must fulfill, so the wiki editing can continue to improve articles and used as it is independently and in parallel of stable versioning. Wikipedia 1.0 goals for print and audio distributions would be much easier to reach.

Not every single article should have a stable version but core topics or useful ones that can lead to print or published editions of Wikipedia. There may well be many articles which do change rapidly because there is new information or it has not been fully developed. However, there are also articles that naturally do not change substantially to be ready for stable versioning. This is a trade-off to consider when choosing a worhtwhile stable version. Even if a stable version is readily outdated, this stable version is still very much useful especially with amendments (e.g. version 1.1). Another stable version (e.g. version 2.0) can be published for the same article down the track. All stable versions (i.e. version 1.0, 2.0, 3.0, etc.) are all useful in parallel unless they are deprecated.

Silly idea

This is a silly, stupid idea.

The possible general outcomes could be that it would be rejected, succeed, or become inactive. Wikipedia:Baseline revision has a very similar but simple concept to this proposal and now it has become historical. The idea may or may not be too premature as of 2005/2006. This proposal is essentially a research where it has not been done before to prove this idea to be a failure or a success. This proposal is to directly address Wikipedia:Pushing to 1.0. A practical, experimental implementation that is easily cleaned up is undertaken without disrupting Wikipedia as much as possible.

There are so many versions of an article to choose from and there are new ones that keep on coming. Is it necessary to only rely on the very latest version when it could be subjected to vandalism or that it has not been verified? This stable versions idea is like picking a suitable version by hand from so many versions to choose from which they can be good or bad. The ones picked by hand is significantly more trustworthy that one that is automatically given the latest one.

Smaller, less "interesting" articles

Articles such as George W. Bush may benefit from some form of "triage". However, the smaller articles, which only have a single, none, or perhaps only a few infrequent editors, will suffer a form of "stability gulag" in which new changes are not brought forward until somebody notices them and verifies them. This will cause a (comparatively) small subset of the encyclopedia to have higher quality, while the much larger articlespace deteriorates in squalor.

Standards

Stable articles need to meet a number of standards. The following are currently some of the criteria.

Implementation

This proposal would not have integrity unless it is proven to be practical. The current implementation is an experiment. Nearly all of the work is located in the Stable namespace and be protected. If for any reason the experimental implementation does not work out, all changes can be undone by simply deleting all articles that begin with Stable: and Talk:Stable: for minimal disruption of Wikipedia to make a point.

Individuals certify with a link to User:Username/CertifyStableTrue or User:Username/CertifyStableFalse that are embedded within {{Certstabletrue|Username}} or {{Certstablefalse|Username}}, respectively. This is useful to track the certification history of an individual by Special:Whatlinkshere/User:Username/CertifyStableTrue and Special:Whatlinkshere/User:Username/CertifyStableFalse.

External links

Bug 3044 on Bugzilla of MediaWiki
Information Quality Discussions in Wikipedia, examination of the information quality aspects of Wikipedia by University of Illinois
Wikipedia to restrict access to encyclopaedia (18 December 2005, Financial Times), "to introduce a more traditional fixed version of its contents in an effort to increase its reliability"