Reading/Web/Desktop Improvements/Frequently asked questions/fi
Why is the width of the content limited? Why is there so much white space?
Our main reason for limiting the width of the content is to improve the readability of all of the amazing content on our wikis. Reading text in an efficient manner is crucial to the large majority of all reading and editing use cases across our projects. While there are several factors that affect readability – e.g. font size, contrast, font, and line length – we have decided to focus on line length initially. Formative line length research on reading of printed texts recommends line lengths between 45 and 90 characters per line (cpl). Recent research on reading of website text focuses primarily within the range of 35 – 100 cpl, with most recommendations falling towards the smaller end of that range. However currently without any width limitation on article content readers might find themselves with line lengths far above the recommended range. A 2005 study summarizes the latest research well: "short line lengths are easier to read", and furthermore, regarding learning and information retention, "subjects reading the narrow paragraphs had better retention than those reading the wide paragraphs".
Lastly, while it’s always important for us to do our own research and form our own conclusions, we think it’s worth noting the overwhelming amount of major websites that have similar limitations on content width. For example: academic journals like Nature, news websites like The New York Times, government and intergovernmental websites like the UN, academic documents like LaTeX, and word processors like Google Docs and Etherpad. Those examples, combined with the extensive research, gives us confidence in this decision.
In short, limiting the width of the content allows for better readability, less eye strain, and better retention of the information itself.
But what about all of the white space!?
We have heard from around 30 editors (particularly people with large screens) who are frustrated by all of the white space created on the sides of the page, though some of them agree that the width limitation is better for reading. There seem to be two main causes of this frustration:
- The white space feels like wasted space
- The white space is bright and distracting
Our goal is to create the best reading experience we can, not to fill every pixel of the screen with content. And in this case less is actually more — people are able to read more easily with shorter line lengths, and focus more easily without the distraction of sidebars or other elements. If the best layout is one that includes white space that is okay — there is nothing inherently wrong with white space.
Additionally, as the project proceeds, we are hoping to begin utilizing some of this space for other functionality. We have started experimenting with making the sidebar sticky to the left side of the page (link to prototype). Further along in the project we plan to experiment with putting a table of contents and/or page tools next to the content. Also, as , limiting the content width gives us new options for content layout, such as a right-hand column dedicated to infoboxes and images.
Why can’t readers just make their browser windows smaller?
Several people have pushed back saying: if people want the content to be more narrow they can make their browser window smaller, or click the “Mobile view” link at the bottom of the page. As mentioned above: since we know that the majority of people come to read articles we should optimise the layout around that use case. We only have one chance to make a first impression and we should aim to give people a great experience as soon as they arrive, without them having to make adjustments.
Tables and other templates don’t fit within the limited width, isn’t that bad?
We have received several reports of tables with long horizontal scroll bars, or templates that expand past the limited width. We’d like to point out that a large percentage of our users, who don’t have large screens and are accessing Wikipedia from their laptops, already had issues with tables and templates even before the change. We should work to make sure that all of our content is as responsive as possible to accommodate all visitors.
Why don’t we just make it a setting?
One of the best parts of the MediaWiki interface is how configurable it is. And while we could make a setting for content width we wonder if it might be beneficial to encourage a common experience that is shared between editors and readers. This could potentially be helpful to editors when making decisions about page layouts (note: 1024px is mentioned as a minimum size to consider in the English Wikipedia Manual of Style, though that’s not quite the same thing). Currently an editor might be editing a page at a width of 1500px, while a reader reads it at a width of 1200px. By implementing a max-width we don’t remove this discrepancy completely (because there would still be variation below the max-width, for people with narrower screens), however we would be greatly limiting the range of variation.
That said, we are not inherently against configurability. If you would like to continue using the new version of the Vector skin without the limited width, you can use a local user script or gadget to do so. We can recommend this one.
How did we decide on 960px for the width?
Please review this page to learn more about how we made this decision: Reading/Web/Desktop_Improvements/First_Prototype_Feedback_Report#Introducing_a_max-width
What about my wiki?
When will these changes be available on the largest wikis?
We hope to see the changes set as default on all wikis later in the year. Each community is welcome to join the early adopters.
Are the improvements to be implemented on sister projects and on non-Latin script wikis?
Yes. We have already made a list of early adopter wikis which represents various sizes and scripts. We also wanted to ensure that at least one non-Wikipedia project is selected.
On which wikis are these changes turned on by default?
Currently, these are:
|Latin script Wikipedias||Non-Latin script wikis||Sisarprojektit|
We are open to add more wikis to this list!
How can this be deployed on my home Wikimedia wiki?
If you are interested to see the Desktop Improvements as default on your wiki,
- ask your community and reach the consensus,
- contact SGrabarczuk (WMF), email: sgrabarczuk wikimedia.org if you need support.
How can I enable it on my own (third-party) wiki?
$wgVectorDefaultSkinVersion = '2'; $wgVectorDefaultSkinVersionForExistingAccounts = '2'; $wgVectorDefaultSkinVersionForNewAccounts = '2'; $wgVectorIsSearchInHeader = true; $wgVectorLanguageInHeader = true; $wgVectorUseWvuiSearch = true;
We are glad to learn that you appreciate our improvements!
What is the scope of the project?
Will Monobook or Timeless be affected?
No. These changes will be applied to Vector only. Vector has been the default interface on Wikimedia wikis since 2010. No other skins will be affected, including Monobook, Timeless, Minerva or Modern.
No. We will not change anything that's within the light gray article content area (except for the table of contents):
How can I suggest improvements?
How do you work with the communities?
- Prior to the deployments:
- We performed user research, and reviewed gadgets and user scripts. See the Repository page for more details.
- We have been reaching out to various wikis asking to join the early adopters.
- We had a roundtable discussion at Wikimania in 2019 (see the outcomes).
- We have run two prototype testing rounds. Editors could gain an understanding of our ideas, and share what they appreciate or find confusing.
- Shortly after the deployment of each feature improvement, we collect usage data via API for each early adopter wiki.
- We run A/B tests for logged-in users. A half of them can experience the changed interface, and a half does not see any difference. Next, we compare the statistics. In the case of negative results, we improve the change or roll it back.
- For logged-out, we compare before and after. Unfortunately, we are not able to perform A/B tests on logged-out users.
- We watch the project talk page. We also engage in discussions on individual wikis regularly.
- Our contracted ambassadors make it easier for us to work with a few communities more closely, react more quickly and effectively.
Other technical questions
How can I disable it?
This is only available to authorized users. It's possible to turn the improvements on and off within user preferences. We have also provided an opt-out button in the left sidebar (accessible on each page): Vaihda vanhaan ulkoasuun
Showing the setting to enable/disable Legacy Vector (the interface without our improvements)
Showing the opt-out button in the left sidebar
We will not remove the opt-out link. Legacy Vector will continue to be available through that link, similar to other skins that have been default in the past, such as Monobook.
How can I report a bug?
Check the following page to see if your bug is a know issue Reading/Web/Desktop Improvements/Breaking changes.
Why not make a new skin? What will happen to Legacy Vector?
It would be an excellent idea to make a new skin, but in the case of Wikimedia skins, it's easier to change an existing one than to create a new one from scratch. There are various reasons:
- it would be too complex to make the existing extensions, gadgets, and user scripts compatible with yet another skin, and too costly to maintain their compatibility,
- it would be too challenging to build and maintain yet another skin (as a total replacement is not an option),
- it would be less likely for the communities to collaborate effectively in the process of building a new skin.
Technically, Desktop Improvements are similar to previous features or projects such as Page Previews or VisualEditor. The only difference is that this time, there will be more of them. Vector documentation should remain relevant.
We will keep and maintain the Legacy Vector. There is no intention of its removal.
Why not use beta features only?
Beta features are available for registered users only, and the improvements are intended to serve our readers and unregistered users as well. Therefore, using beta features only would give us feedback from a very specific type of user that is not representative of our entire base of users. And moreover, we wish to receive the readers' and anonymous users' feedback from the earliest deployments.
What are the feature's success metrics?
Increase utility among our existing audiences, proxied by:
- Increase searches per session by 5% over the course of the project
- Increase language switching per project by 5% over the course of the project
- Increase in positive and welcoming sentiments towards the site (via surveys and user testing)
- Increase in sentiments of trust and credibility (measured via surveys and user testing)
As we define the changes we want to make with more specificity, we will expand and iterate on this list.
- Calculating Line Length: an arithmetic approach by Ernesto Peña, PhD (Visible Language Journal)
- Computer text line lengths affect reading and learning by Peter Orton, Ph.D. IBM Center for Advanced Learning