Page MenuHomePhabricator

Expand accessibility portion of Readership Gaps Taxonomy
Closed, ResolvedPublic


Though the readership gaps taxonomy is largely complete, we identified that one area in which it could be improved was the accessibility component of readership gaps. This will entail the following:

  • Summarize literature on Wikipedia + accessibility (this will likely be less surveys that have been run and more focused on smaller qualitative studies)
  • Identify resources that describe accessibility challenges for Wikipedia
  • Describe possible metrics for accessibility + readership

Event Timeline

Summary of conversation with Ted Drake of Intuit:

  • High-level: Wikipedia should be the most accessible site on the internet.
  • Thoughts from a recent conversation with an individual who uses a screen-reader and their view of Wikipedia:
    • Too much information usually -- this could be improved by building in better landmarks for navigation. Not just sections but also reaching disambiguation pages, related content, edit history, etc. -- i.e. any grouping of content that consistently shows up across articles
    • Actually goes directly to Wikipedia very little now -- mostly uses Google Home to access the content because it greatly reduces the overhead by surfacing the content most relevant to the particular need
  • Readability is one of the largest accessibility challenges but is not generally something that can be automated
  • Don't try to create a separate "accessible" version of Wikipedia -- exception being simplified versions of the content such as with Simple English Wikipedia.
  • It could be hard to automatically surface accessibility issues to editors but plugins such as WAVE on Chrome are quite good.
  • Edit suggestions that focus on accessibility are a good idea
  • Suggestions for improvement to Wikipedia (this is out of the scope of this work but worth recording and considering for measurement of gaps):
    • Improve the reader view for Wikipedia (stripped down version of a page that browsers can generate)
    • Improve landmarks -- i.e. consistent hierarchy to pages that screen readers can use to simplify navigation. This goes beyond sections and also includes common use-cases for readers -- e.g., perhaps the primary infobox, link to disambiguation pages, etc.
    • For individuals with short-term memory loss, making it easier to remember what they have (not) looked at can be quite helpful. This can in part be done through browser history but, for instance, links can be styled more obviously when they have been visited:
    • Make sure that sections have labels for screen readers
    • For tables that are used for layout as opposed to actual data (e.g., the WikiProject Accessibility navigation box), add role=none so screenreaders know that they are not actually data tables.
  • Research / measuring impact:
    • Identifying user needs through surveys, usability testing, etc. are good ways to identify barriers
    • No great way of identifying users with accessibility challenges e.g., via usage logs. For privacy reasons, 3rd-party extensions etc. aren't visible to a website. Exception is for apps where you can detect if someone for instance is using a screen-reader (e.g., to add additional buttons to facilitate navigation) but this is going to be a very imperfect proxy for measurement of whether you are reaching a given population of people
    • Otherwise pay attention to simple indicators like # of steps / time to get to complete a process
    • We should consider submitting work to Web4All if we continue to do work in this space
  • Resources:

Unrelated to discussion with Ted, but some resources that I have thusfar identified:

@Isaac thanks for capturing your conversation. very informative.

  • +1 for submitting to Web4All if the work becomes bigger. Even if we don't submit, we can propose a panel on the broader topic if relevant.
  • My read of what you have here is that "readability" can be the main focus. curious to hear your thoughts (in the future discussions, no rush).

Weekly update: no progress but next steps will be to work Ted's thoughts into the existing taxonomy and identify remaining gaps.

Weekly update: no progress.

Weekly update: no progress. Though I did note that there is a new, relevant user group whose page I should go through for relevant information:

Weekly update: added draft of accessibility section to Readers taxonomy on Overleaf.

tl;dr: I'll incorporate some of the below into the literature and metrics sections around accessibility + readership.

Did some additional academic search to make sure I didn't miss anything obvious there.

From W4A, I read through:

Google Scholar:

I didn't include the specific papers, but math formulas are particularly difficult when it comes to accessibility. See Extension:Math for some details.

Regarding detecting accessibility issues that might affect the readability of content, I went through the resources listed in WikiProject Accessibility's Navigation Template and in particular this Dos and don'ts list is a good place to start. Other language's instantiations of WikiProject Accessibility might have additional suggestions. In particular, a few items and related templates that I identified that might make them more easily actionable:


Gaps write-ups from Overleaf copied below. Still some iteration likely but at this stage, I would consider this task complete. @leila let me know if you concur.


The Disabilities gap reflects how individual disabilities might affect one's ability to access the knowledge within Wikipedia. While individuals who are blind might be the most salient example, disabilities fall into many categories: cognitive, developmental, intellectual, mental, physical, or sensory disabilities. There is scant literature on the degree to which individuals who read Wikipedia have various disabilities beyond anecdotal evidence---e.g., an interview with Graham Pierce, an individual who is blind and a prolific editor, tips for improving accessibility, a survey from 2008 that indicated that Wikipedia was a popular site for individuals who use screen readers. A number of projects focus on improving the accessibility of Wikimedia sites though, such as WikiProject Accessibility, WikiProject Usability, WikiBlind User Group, and Para-Wikimedians User Group.

While there is little data on readers with disabilities, we know about many of the barriers that they can face so initial metrics might focus on measuring the accessibility of a given project---e.g., proportion of images with captions---as opposed to levels of readership in communities of people with disabilities. In general, providing a variety of ways to access content---e.g., text, images, video, audio---while reducing barriers to understanding---e.g., good color contrast, high readability---is a good approach to improving the accessibility for readers of all abilities. For example, Wikispeech provides text-to-speech for Wikipedia articles and VideoWiki provides a tool for collaboratively editing videos from images and wikitext.


The Disabilities gap reflects how individual disabilities might affect one's ability to access and contribute to the knowledge within Wikipedia. While individuals who are blind might be the most salient example, disabilities fall into many categories: cognitive, developmental, intellectual, mental, physical, or sensory disabilities. There is scant evidence beyond the anecdotal of the degree to which individuals with various disabilities contribute to Wikipedia (see WikiProject Accessibility Userboxes, this interview with Graham Pierce, and evaluation of editing with a screenreader in 2008). Various groups have also been established for individuals with disabilities such as the WikiBlind User Group and Para-Wikimedians User Group.


Additionally, content might be structured in such a way that it is inaccessible to readers with various cognitive, developmental, intellectual, mental, physical, or sensory disabilities---e.g., blindness, dyslexia. For English Wikipedia, the Accessibility component of the Manual of Style was written to address these issues and the Accessibility Dos and Don'ts provides a quick overview of some of the more prominent accessibility gaps in content. In particular, the following barriers are identified (to which we add how they might be better tracked):

@Isaac yup. it's ready to be resolved. thanks! :)