편집 검토의 개선/편집 검토를 위한 새로운 필터
|Edit Review Improvements (ERI)|
New Filters for Edit Review is a suite of improvements designed to help edit reviewers be more efficient and effective on pages like Recent Changes, Watchlist and Related Changes.
An additional goal of the project is to help new contributors, who require a more supportive edit-review process according to research. The new tools do this by enabling reviewers to identify, for the first time, edits by new volunteers who are making errors but are, nonetheless, acting in good faith.
The New Filters for Edit Review introduce an easier yet more powerful filtering interface as well as a whole list of new filters and other tools, including live page updating, user-defined highlighting, the ability to save filter sets for re-use, and predictive filters powered by machine learning. This page summarizes the new tools and other benefits the project provides. The documentation pages can help you learn how to most effectively use these new capabilities. The New Filters were created by the Global Collaboration Team.
- 1 What’s new about the ‘New Filters’?
- 1.1 New filter groups
- 1.1.1 Get predictions about edit quality
- 1.1.2 Get predictions about whether users are in good faith
- 1.1.3 Search by experience level
- 1.1.4 Search by new change types
- 1.1.5 Get (only) the latest
- 1.1.6 Watch your Watchlist (from Recent Changes)
- 1.1.7 See the ‘Unseen’ (on Watchlist)
- 1.1.8 Search by namespaces and particular edits
- 1.2 Power tools
- 1.3 All-new interface
- 1.1 New filter groups
- 2 Project status
- 3 See also
What’s new about the ‘New Filters’?
The following items describe key additions and enhancements of the New Filters for Edit Review project.
New filter groups
Get predictions about edit quality
Available on certain Wikimedia wikis only
The “Contribution quality” filters offer predictions about edits. Are edits likely to to be of high quality or to have problems? With those filters, it is easier to search for good edits, to thank or help productive contributors. The filters can also detect bad-quality edits and help patrollers to improve them.
ORES, a machine learning service, makes the quality predictions based on statistics. ORES has been trained on a large set of edits, previously scored by human editors. These quality predictions filters are available only on wikis, where the ORES “damaging” test is enabled.
Get predictions about whether users are in good faith
Available on certain Wikimedia wikis only
The “User intent” filters offer predictions about whether edits were made in good faith or not. With those filters, it is easier to search for edits made in good-faith, by good-faith users. By enabling detection of bad-faith edits, filters assist reviewers looking for vandalism.
Like the Quality filters above, these predictions are made by the machine-learning service ORES, based on statistics. These intent predictions filters are available only on wikis, where the ORES “damaging” test is enabled.
Search by experience level
Three new filters let you search by the experience level of editors who are making changes. These filters can be useful in a number of ways. For example, research shows that new editors are particularly vulnerable to rejection. The Newcomers filter enables reviewers to treat new users with the care they require.
These are the three new experience levels:
- Newcomers - Registered editors who have fewer than 10 edits or 4 days of activity.
- Learners - Registered editors whose experience falls between "Newcomers" and "Experienced users." (corresponds to autoconfirmed status on English Wikipedia).
- Experienced users - Registered editors with more than 500 edits and 30 days of activity. (corresponds to extended confirmed status on English Wikipedia).
The filters return the edits only of users who are currently logged in.
Search by new change types
In the past, reviewers had the ability to hide Category changes and Wikidata edits only.
The New Filters add three additional “Types of change” filter:
- Page edits lets you filter for changes to content and discussion pages.
- Page creations lets you use Recent Changes to patrol new pages.
- Logged actions lets you isolate—or exclude—things like account creations, page deletions and administrative actions generally (the events tracked on Special:Log).
Get (only) the latest
The “최신판” filters let you choose to see only the most recent edit to any given page. All prior changes on that page will be omitted.
Watch your Watchlist (from Recent Changes)
Use the “주시 중인 문서” filters group on Recent Changes to isolate or highlight changes made on pages you watch.
See the ‘Unseen’ (on Watchlist)
Beta feature, on Watchlist only
On Watchlist, use “확인한 변경사항” and “미확인한 변경사항” filters (in the “주시문서 목록 활동” group) to identify changes to pages you have or haven’t visited since the changes occurred.
Search by namespaces and particular edits
Search by namespace
Filter namespaces to narrow your search.
Search by tags
Filter tags used on your wiki to find particular edits.
Search by username
The Users filter menu enables you to exclude or include edits by particular users.
Make results more meaningful with Highlighting
Highlighting lets reviewers apply color intelligently to emphasize edits that match desired criteria. Used correctly, highlighting adds meaning to Recent Changes search results and can help you find what you’re looking for more quickly.
Bookmarking: save and reuse your settings
The New Filters provide a lot of new functionality. To help you manage the wealth of options conveniently, we’ve provided the ability to save sets of filters, by name, for later use. Just click the bookmark icon to save your current settings. You’ll find all your saved settings in the “Saved filters” menu, where you can easily switch among them.
Manage preference right on the page
With bookmarks, you can declare a set of saved filters as your default. Those settings will load automatically every time you visit the page. This means you no longer need to visit the Preferences page to set your preferred behavior.
We plan to remove all redundant Recent Changes and Watchlist preferences, consolidating the remaining options on one tab.
A much-requested feature, “Live update” lets you use Recent Changes for near-real-time monitoring. In Live Update mode, the Recent Changes page automatically adds new changes to the top of the search results every three seconds.
Easier to use
To make the sophisticated functionality offered by the New Filters as friendly as possible, the filtering interface has been completely re-imagined. Functions are grouped logically and explained clearly. An "사용 중인 필터" area shows at a glance what your settings are. And a smart help system provides in-context guidance.
In a number of ways, the New Filters are organized using a different logic from the old ones. For example, by putting related filters into logical groups, the New Filters enable users to intelligently broaden or narrow their searches. Similarly, most previous Recent Changes filters offered only a simple option to hide a certain property (X). By augmenting those old filters so that they now let users show or hide both X and its opposite (notX), the current tools provide more control. For example, previously, one could show Wikidata edits or hide Wikidata edits, but one could not show only Wikidata edits.
On Recent Changes (and Related Changes)
Since May 2017, the New Filters have been available as a beta feature on Recent Changes (and Related Changes). In the months since then, Global Collaboration Team has received and acted on numerous feature suggestions and bug reports from community members and others, and the features have been through multiple rounds of user testing. They have, in general, received a very positive reception.
In September and October 2017, the principal New Filters features listed above graduate out of beta on Recent Changes only, becoming standard parts of MediaWiki. Users who want to can opt out of the improvements in their preferences.
Meanwhile, the Global Collaboration Team will continue to add new features to the beta and make fixes, so please give us your suggestions!
In September of 2017, subscribers to the New Filters beta will get the New Filters on Watchlist.