User Generated Content

From Silver

User Generated Content

Draft as of 29 July 2021

Note for AGWG/Silver Reviewers: This draft proposal is intended for upcoming WCAG 3 Working Draft publication. Various portions of this document are intended to go in different locations of WCAG 3 documents, including the Conformance Section, the Glossary, the guideline on alternative text, and potentially also the WCAG 3 Explainer. Those intended locations are generally called out in each section.

Editor's Note: Note: Exactly where to place this Editor's Note in the next WCAG 3 Working Draft is yet to be determined. The Status of this Document and the Review Questions Section are two possible locations.

We seek feedback on the User Generated Content Conformance category newly introduced in this Working Draft. Please provide comment on any or all of its aspects including:

  • The Glossary Definition of User Generated Content;
  • The proposed User Generated Content Conformance section;
  • The provided example outcome and methods for User Generated Content alternative text;
  • The introduction of the new User Generated Content Conformance concept itself.

We further seek comment on what WCAG should say (or not say) in certain circumstances including:

  • Whether and how to account for any entity's use of social media platforms when that platform is utilized as part of that entity's official on line presence;
  • Whether and how to account for content contributed on some social media platform by an official of some entity where the contribution relates in some way to that entity (such as when an employee of an entertainment company replies to a post by a fan on an unofficial fan page for an artist who works for that entertainment company).

End of Editor's Note

Glossary Definition

Note: the text below will go into the glossary at https://rawgit.com/w3c/silver/User_Generated/guidelines/index.html#dfn-user-generated-content

User Generated Content is content that is created by site visitors, either by using the authoring environment provided by the owner or author, or imported from content otherwise contributed to the author (for example, sent via email). Any content provided by employees, contractors, or authorized volunteers of the publisher is not User Generated Content.

User Generated Content (Conformance Section)

Note: This section is intended for the Conformance Section of WCAG 3 at https://rawgit.com/w3c/silver/User_Generated/guidelines/index.html#user-generated-content Portions of the explanatory text may also be used elsewhere such as in the Explainer. Note: See also github issue #450.

Web content publishers may include content provided by the users of their digital products. We refer to such content as User Generated Content.

Examples of User Generated Content include:

  • social media postings and comments,
  • uploaded photographs, or
  • uploaded videos or other multimedia.

User Generated Content is provided for publication by visitors where the content platform specifically welcomes and encourages it. Any content provided by employees, contractors, or authorized volunteers of the publisher is not User Generated Content. The purpose of the User Generated Content Conformance is to allow WCAG 3 outcomes and methods to require additional or different steps to improve the accessibility of User Generated Content.

An important part of WCAG Conformance is the specific guidance that is associated with individual WCAG 3 guidelines and outcomes. Not all WCAG 3 guidelines will have unique outcomes and testing for User Generated Content. Unless User Generated Content requirements are specified in a particular guideline, that guideline applies as written whether or not the content is User Generated.

The web content publisher should identify all locations of User Generated Content (such as commentary on hosted content, product descriptions for consumer to consumer for sale listings, and restaurant reviews) and perform standard accessibility evaluation analysis for each. If there are no accessibility issues, the User Generated Content is fully conforming.

Steps to Conform

If accessibility issues are identified, or if the website author wants to proactively address potential accessibility issues that might arise from User Generated Content , then all of the following should be indicated alongside the User Generated Content or in an Accessibility Statement published on the site or product:

  1. Clearly identify where User Generated Content can be found on the publisher's digital product (perhaps by id href);
  2. Clearly identify the steps taken to encourage accessibility in User Generated Content such as prompting the user for ALT text for their uploaded images before they are accepted and the disallowal of text attributes except as they are part of semantic markup such as strong, headings, etc., as enumerated in Guideline Outcomes;
  3. Utilize AI tools as they become available and feasible to use, to search User Generated Content and help repair accessibility problems.

Editor's Note: Appropriate scoring is yet to be provided. Once provided, fully conforming content will score as fully conformant. It remains to be determined how to score User Generated Content that has accessibility issues; and to define what minimum threshold scores might be acceptable; and what critical errors might prevent a conformance assertion. We expect WCAG 3 to provide this guidance within individual guidelines and outcomes and to support testing for conformance. The working group is looking at alternative requirements to apply to User Generated Content guideline by guideline, and is seeking feedback on what would serve as reasonable requirements on how to best support accessibility in User Generated Content with known (or anticipated) accessibility issues. See the new example Outcome in Text Alternatives End of Editor's Note

Specific guideline recommendation

Note: An important part of WCAG Conformance is the specific guidance that is associated with individual WCAG 3 guidelines and outcomes. Not all WCAG 3 guidelines will have unique outcomes and testing for User Generated Content. Unless User Generated Content requirements are specified in a particular guideline, that guideline applies as written whether or not the content is User Generated. This gives us the flexibility to provide specific User Generated Content guidance where appropriate without forcing it to be present in each and every guideline whether or not it's actually needed in that guideline. The specific Guideline example below further develops the guideline for text alternatives published with the WCAG 3 First Public Working Draft (FPWD) to provide specifications for User Generated Content. Since the Outcomes and Scoring models for WCAG3 are currently being revised, this is the only example we recommend for now. Additional examples and recommendations should be expected in future Working Group drafts.

Text alternatives for User Generated Content

Individuals uploading photos to social media may not always provide good quality alternative text, if they provide any at all. We therefore score the quality of User Generated text alternative content differently from publisher provided text alternative content for conformance purposes as described below. The key to that scoring is the work the web site does to prompt their users for alternative text and provide guidance on creating appropriate alternative text. Also key is the extent to which the web site is taking advantage of the latest in computer vision and machine learning tools to automatically create, and repair User Generated alternative text.

Note: precise scoring of these is pending an update to the WCAG 3 Scoring proposal.

Outcome: Text alternative available for user generated content

Note: This section is intended for the Outcomes and Methods tabs in the Text Alternatives guideline: https://rawgit.com/w3c/silver/User_Generated/outcomes/text-alternative-available-UCG.html and is based on the recommendation of the Silver Text Alternatives Subgroup: https://docs.google.com/document/d/1LCLBGChfDTClgLG8f-GJSs2QfhX2wdkeS99IhwIBqPE/edit#

If the content is user generated content, then a text alternative mechanism for non-text content is available via user agents and assistive technologies. This mechanism allows users who are unable to perceive and / or understand the non-text content to determine its meaning. The publisher of the user generated content:

  • Provides functionality for adding non-text content, then users are able to modify programmatically associated text alternatives for non-text content.
  • Provides functionality to check if Generic strings (e.g. "image") and irrelevant strings (e.g. the file name, file format) are not used as text alternatives
  • If the repair attempt occurs during an authoring session, users have the opportunity to accept, modify, or reject the repair attempt prior to insertion of the text alternative into the content.
  • If the repair attempt occurs after an authoring session has ended, the repaired text alternatives are indicated during subsequent authoring sessions (if any) and users have the opportunity to accept, modify, or reject the repair strings prior to insertion in the content.
  • Provides the functionality for adding non-text content, when users enter programmatically associated text alternatives for non-text content AND the text alternatives are automatically saved and suggested by the authoring tool, if the same non-text content is reused.
  • Provides the functionality for adding non-text content, when users enter programmatically associated text alternatives for non-text content AND the author has the option to edit or delete the saved text alternatives.

The above references are adapted from the Authoring Tool Accessibility Guidelines (ATAG) 2.0 B.2.3 Assist authors with managing alternative content for non-text content.

Note: How the specific details of this outcome will be managed and what is required vs what is recommended still needs discussion. This is an example of how an outcome could be used with user generated content.

This outcome relates to guideline Text alternatives.

Methods:

  • Functional Images (HTML, PDF, ePub)
  • Decorative images (HTML, PDF, ePub)
  • Informative images (HTML)
  • Author control of text alternatives (ATAG)