User:Andrewa/The third draft regarding avoiding primary topic

Much of this essay is obsoleted by User:Andrewa/Primary Topic RfC.
Sections of it are still of interest, but aren't likely to be further updated.

This is an essay.

It contains the advice or opinions of one or more Wikipedia contributors. This page is not an encyclopedia article, nor is it one of Wikipedia's policies or guidelines, as it has not been thoroughly vetted by the community. Some essays represent widespread norms; others only represent minority viewpoints.

In that all pages belong to the whole project, any user may edit this one. But it's generally more helpful (and polite) to discuss the proposed change on its talk page first.

This page in a nutshell: The concept of primary topic (or P T) has a significant downside.
The supposed advantages of it are partly illusion, partly obsolete, and partly both.
A gentle phase-out is both possible and desirable.

The principles

One: All readers should be able to get to the topic they want. This is not negotiable.
Two: Readers should get to the topic they want with a minimum number of mouse clicks.

It is also of course desirable that page titles should make things as easy as possible for editors, making wikilinking easy and natural, correction of mislinkings quick and easy, and wasting as little time as possible. These also have an effect on readers, obviously; The most productive and least hassled editor base will produce the best encyclopedia for them to read. But the readers come first.

The downside of Primary Topic

Ambiguous names hinder readers

It is generally assumed that it is to the advantage of readers if the article they want is at the base name. This is true in some cases, but in others the very opposite is the case.

There are certainly some scenarios in which having the wanted article at the base name will save one mouse click. But there are also scenarios in which it will:

Require one extra mouse click.
Require two extra mouse clicks.
Make it unlikely that the article will be found at all.

This is so counter-intuitive that many just seem to assume that the scenarios that save a mouse click are more likely. No evidence has ever been offered that this is the case. It would be interesting to see it.

Page moves break links

Every page move breaks any incoming external links. It is to the readers' advantage if the number of page moves is minimised.

Articles at ambiguous base names are vulnerable to moves whenever the primary topic changes or is seen to have changed.

Deciding the Primary Topic is subjective

Perception of primary topic is heavily dependent on the background of the reader (and/or the editor).

This perception is often so strong that Wikipedians often claim, in good faith, that it makes Wikipedia look silly (or even POV) to disambiguate their own pet meaning and prefer another.

Certainly, Wikipedia looks silly to a surfer or a sailor if the physics article is at the base name wave, and equally silly to a physicist if the article on ocean waves is there. But if the DAB is at the ambiguous name, we look scholarly. We may be seen as stuffy and pedantic, but at least we're not seen as stupid.

Probably more important, we're not seen as POV. Disputed names such as Macedonia should always be disambiguated. But if we disambiguate only disputed names, even this smells strongly of POV to those involved and defending their cause. Disambiguating all ambiguous names avoids this bad smell.

Ambiguous names hinder editors

An editor who mislinks to a DAB gets a warning. An editor who mislinks to an article at an ambiguous name gets no warning.

An editor who wishes to create an article at an ambiguous name goes through a circuitous process of deciding what the primary topic might be, and then possibly defending that decision. If ambiguous names are avoided, the process is greatly simplified.

The advantages of Primary Topic

Common names do not look funny

This is the big one. Wikipedians are a cowardly lot. They think that if they disambiguate common terms like mathematics, people will laugh at them at parties. It would perhaps be more helpful and scholarly to acknowledge these other meanings up front, but the lowest common denominator is a lot safer. Fortunately there are workarounds that allow us to leave mathematics right where it is, to little if any disadvantage. Not worth arguing about.

Some readers who do not know that a term is ambiguous have no problem

That is, some readers. If there are any physicists who don't know about surfing, then they'll get to the right article when they go to or link to wave, assuming that it's the main and normal meaning and that everyone else thinks the same way. But surfers who know no physics will get to the wrong page, similarly assuming that their favourite meaning is the best one. Similarly with valve, a chemical engineer will get the right page, while a guitarist who calls a vacuum tube a valve and calls the other sort of valve a tap will be astonished.

We've always done it this way

Yes. And the mind boggles at why anyone who accepts this as a valid argument would want to edit Wikipedia at all.

Other

Um, are there any others? Suggestions welcome! See the talk page.

How readers find articles

A little analysis.

By searching using the Wikipedia search facility
By external search engines such as Google web search and DuckDuckGo
By Wikilinks from within Wikipedia
By Wikilinks from other language versions and sister projects
By incoming external links from other websites
By following links given to them in emails. SMS messages and similar
By typing in a URL given in a published source such as a textbook or course reading list
By following shortcuts they have themselves created

and possibly by other means but those are the main ones.

These can be grouped as follows:

External search engines.
Internal Wikipedia links.
Sister project links.
Incoming external links.

Incoming external links

We have no way of fixing or even finding these. All we can do is provide redirects or hatnotes or both whenever we move a page.

Sister project links

In theory we can find and fix these, and Wikidata helps to avoid breaking them in the first place. But we make no other effort to fix them, and are not likely to do so.

Internal Wikipedia links

Having New York State at the base name New York created many thousands of mislinkings, now fixed. There are likely to be mislinkings pointing to every article at an ambiguous base name, created on a regular basis. In the case of links from edit summaries, past page versions in page histories, and archives of talk and other discussion pages, these mislinkings will never be fixed.

Even Wikipedia search gives little help when the required article is at an ambiguous name, unless a redirect already exists from an unambiguous name. The person must guess that the ambiguous name is the article they want... and yes, they often do this. Or, they may guess what (disambiguation) means, and get there in only one extra mouse click. If the article was at an unambiguous name, they'd get there at the first mouse click, most of the time.

A person who wrongly guesses that the base name is the article they want, on the other hand, often needs two extra mouse clicks to get first to the DAB following the hatnote, and then to the article they want.

External search engines

We have little control over these. Some may list redirects, some not. Some give the first few words of the lead, some do not.

If the article is at an an unambiguous name it will be easily found using an external search engine. And if not, maybe... or maybe not.

The way forward

There are three issues. They are loosely related but any one or two of them can be actioned independently.

Ambiguous names

The Big Bang approach, to move all articles to unambiguous names and all DABs to their base names, is not the way to go. Definitely not now, and probably not ever.

Instead:

All new articles should be at unambiguous names.
All new DABs should be at their base names.
All article moves should be to unambiguous names.

This minimises the damage of page moves, while leaving articles such as mathematics and London at their existing ambiguous names, probably indefinitely. It achieves nearly all of the benefits of the Big Bang, with relatively little effort and trouble.

See User:Andrewa/Primary Topic RfC.

Redirects from unambiguous names

Where existing articles are at ambiguous names, there should always be a redirect from an unambiguous name, such as already exists at London (England) (since 2006). This is an independent change, and could probably be boldly done, even without requiring discussion. On the other hand, best not to unnecessarily risk ruffling feathers. It would not be helpful to have a flurry of redirects for discussion entries.

Disambiguator of DABs

A more explicit disambiguator, such as index of articles, would be helpful to those who are not familiar with Wikipedia jargon. But such a change cannot be done boldly. It requires at least an RfC with good participation and consensus. In the meantime, perhaps redirects from for example London (index of articles) (redirecting to London (disambiguation) in that particular case) or from a similar, more generally understood title could be boldly created. (Or maybe not. Again, it might just ruffle feathers. Another RfC? Just to make sure these aren't all RfDd?)

This is a bigger job than it might appear, owing to the number of hatnotes and hatnote templates that refer to the existing disambiguation disambiguator. Many of these templates are transcluded, so (assuming the RfC is supported) rather than modifying the existing templates, new templates should be created, and the existing ones deprecated.

Where they exist, redirects from the existing pages with the disambiguation disambiguator should be preserved, but no new ones should be created.