Due to lack of mechanisms for true collaboration growth is incredibly hard for even third party wikis. Below are some encountered issues.
Concrete problems:
- Assumption 1 -> anyone can create an article on their first try. This is a very core wiki principle and flaw. Nobody can produce a good page in whatever medium (be it a book or letter or whatever) on their very first try. Yet the interface persuades a user to do this, relying on the principle that it will be improved overtime. Except that sometimes before one even gets that chance the article may be deleted. Wasted work...
- Assumption 2 - Just because a non-existent page is reached it should be created - page creation workflow is non-existent as far as default mediawiki is concerned. Consider this, the link to improve or create pages is called ?action=edit. The only way to reach said link is either by knowing it, or reaching it accidentally.
- Assumption 3 - As long as it has content , anything created in the main namespace is deemed a page. One can create a page with a single character, a word, gibberish ("aszxdcareqweqdas"). No sensible restrictions on what can be saved as a page.
- Before the advent of internet, an application would be basically awful and poor quality if it didn't have built-in documentation. Nowadays, interfaces are designed to assume that users are mind-readers. They simply choose the lazy way of sending the user to a separate internet page that may contain help. For wikis, this is worse, one can reach a "documentation" page that has been vandalized and be even more confused.
- There is much duplication of effort - the simple fact of the matter is that volunteers are inherently unreliable. This means that they can work in haphazard ways, and can individually (un)knowingly disrupt each others work, or agree to do something but not do it. This is made worse by the fact that there is no way to get an overview of what processes, pages, or places need most help. In some cases this leads to decline rather than growth.
Some possible solutions:
- Basic tutorial - it would be easy to either guide the user through a compulsory basic tutorial, and perhaps a sort of micro-quiz before starting to edit. While this could discourage many, the ones that remain may actually produce better content. An alternative is a personal draft ( created with the help of the tutorial) invisible to others until the user publishes it. The draft would be automatically deleted if not published within a week or so, maybe after a copy is emailed to the author.
- Entry point for creating a page - One possibility is that redlinks on search pages only appear after considerable users search for the same keyword. The most important issue to solve here is the creation of a sensible entry point for creating a page (not search or 404). See how wikia for instance created it : community.wikia.com . A more wikimedia-centric approach was discussed here Article Creation Workflow/Design , https://phabricator.wikimedia.org/T29311).
- Definition of a content page - Some basic mechanisms or restrictions for creating a page should be introduced, e.g. minimum character count, minimum number or links, a reference or a heading or whatever. Content consisting of one character or word is not a page, it is a scribble. Mediawiki even allow(ed?) pages with no content.
- Online help / in-context help - Help for most frequently used dialogs, and even a general description of the interface itself, along with a glossary and definitions of terms used by it. Visualeditor has some basic help. Wikitext editor focuses on wikitext, and doesn't provide any help for many of its dialogs or even describe what some terms mean.
- General Queue -This will not likely be achieved by this team. It simply requires a general Queue, e.g. What articles are reasonable, what articles need more work, which have low coverage, which could benefit from simple fixes and are good for beginners, etc. The reliance on categories for this kind of thing is purely misguided. Currently, editors fly blind, randomly improving things without a sense of progress and sometimes duplicating work or reverting someone else's efforts. See https://phabricator.wikimedia.org/T120742, https://phabricator.wikimedia.org/T91655.