Conformance Options

CSUN Hybrid Meeting

Q1 2023

Morning Slidedeck

1

Meeting Logistics

• Morning Goal: Update our 2 conformance options and try them out
• Agenda
• CEPC
• In order to encourage wide participation:
• Be conscious of how often you are queuing to speak
• Be concise
• Be respectful
• q+ to say, q+ to reply
• Sandbox
• Some hard conversations: If needed, chairs will call a 15-minutes break

2

Quick Background Summary of How We Got Here

3

Scope

• Items are the smallest testable unit. They may be interactive components such as a drop down menu, a link, or a media player. They may also be units of content such as a word, a phrase, a label or error message, an icon, or an image.
• Views include all content visually and programmatically available without a substantive change. Conceptually, views correspond to the definition of a web page as used in WCAG 2.X, but are not restricted to content meeting that definition. For example, a view could be considered a "screen" in a mobile app or a layer of web content – such as a modal.
• User processes are a series of user actions, and the distinct interactive views and items that support the actions, where each action is required in order to complete an activity. A user process may include a subset of items in a view or a group of views.
• The aggregate is the combination of items, views, and user processes that collectively comprise the site, set of web pages, web app, etc.

4

Overall Structure - From First Public Working Draft (FPWD)

• Functional Needs - Statements that describe a specific gap in one’s ability, or a specific mismatch between ability and the designed environment or context.
• Guidelines - High-level, plain-language versions of content.
• Guidelines support one or more Functional Needs
• Guidelines have informative information in the “How To”
• Outcomes - written as testable criteria in plain language
• Outcomes support one or more Guidelines (ideally many outcomes to 1 guideline, but likely not possible)
• Outcomes have an “And” between them (must meet all the outcomes)
• Methods - technology specific ways to meet Outcomes (ideally many to 1)
• Each method is sufficient
• Methods include scope tested (component, view, user processes, aggregate)
• Methods have an “Or” between them (choose the appropriate method) to support:
• Different technologies,
• Different scope
• Tests are scored individually and aggregated for a rating of the Outcome
• Methods should be broken apart when:
• Anytime an “Or” occurs
• Different technologies
• Processes can have Critical Errors where the functional need isn’t met. A site or app can’t pass with critical errors.

5

Conformance from First Public Working Draft

Testing, Scoring, and Conformance

• Individual objects can be scored by pass/fail.
• The aggregate of the individual objects can result in pass based on the percentage that passed. This allows for passing without perfection.
• Tests without individual objects can be scored by an adjective rating or rubric.
• Essential processes are evaluated with critical errors. Any critical error is a fail.
• Outcomes are rated by scale of 1-5
• Overall score is an average of the ratings. Bronze is no critical errors and 3.5 score

6

Feedback Received

304 Github Issues received - common themes:

• People liked being able to pass without perfection. They thought it would improve accessibility by making it more achievable. Some did not want any failures to pass.
• People liked critical errors idea.
• People thought it was too complex. One said to treat it like bug severity.
• People wanted to count the failures, not the passes. Counting was too time-consuming.
• People wanted only objective tests
• It’s too hard to test everything and scoring depends on testing everything

7

Overall Structure - From TPAC

• Functional Needs - Statements that describe a specific gap in one’s ability, or a specific mismatch between ability and the designed environment or context. (Normative)
• Guidelines - High-level, plain-language versions of content. (Normative)
• Guidelines support one or more Functional Needs
• Outcomes - written as testable criteria (Normative)
• Outcomes support one or more Guidelines (ideally many to 1, but likely not possible)
• Outcomes have an “And” between them (must meet all the outcomes)
• Methods - technology specific ways to meet Outcomes (ideally many to 1) (Informative)
• Each method is sufficient
• Methods include scope tested (component, view, user processes, aggregate)
• Methods have an “Or” between them (choose the appropriate method) to support:
• Different technologies,
• Different scope
• Test Sets - Unique combination of tests that is sufficient to meet the Outcome (Informative)
• Test Sets may include different types of tests
• Test Sets will include tests that pass, better, and exemplary

8

How To Document

Guidelines Document

Overall Structure - From Current Editor’s Draft

​

• Guidelines - High-level, plain-language versions of content. (Informative)
• Guidelines support one or more Functional Needs
• Outcomes - written as testable criteria (Normative)
• Outcomes support one or more Guidelines (ideally many to 1, but likely not possible)
• Outcomes have an “And” between them (must meet all the outcomes)
• Assertions - A formal claim of fact, attributed to a person or organization�
• Methods - technology specific ways to meet Outcomes (ideally many to 1)
• Each method is sufficient
• Methods include scope tested (component, view, user processes, aggregate)
• Methods have an “Or” between them (choose the appropriate method) to support:
• Different technologies,
• Different scope
• Test Sets - Unique combination of tests that is sufficient to meet the Outcome
• Test Sets may include different types of tests
• Test Sets will include tests that pass, better, and exemplary

9

Normative

Informative

Methods should be broken apart when:

• They have a different test set
• Anytime an “Or” occurs
• Different technologies
• Different test scope

Guidelines List (Placeholder)

• 2022 project to analyze all the WCAG 2.2 success criteria
• Functional need
• Scope (site, view, component, process)
• Test Type
• Outcomes or subguidelines?
• Results: 23 guidelines sorted by topic to identify needed expertise

10

Current State

• Editor’s Draft with updates
• Addition of Placeholder Guidelines
• Update to introduction

11

Conformance Option 1:

Prerequisite/Required/Optimized (PRO)

Read Ahead for Option 1

12

Single Page Summary

• Prerequisites and Sectors
• Required Methods
• Optimized Methods
• Levels:
• Prerequisite - Progress only, way to quickly determine gaps
• Bronze - All required baseline outcomes
• Silver - Bronze and at least 50% of optimized methods
• Gold - Bronze and 75% of optimized methods

13

Key Concept: Prerequisite Methods

• Tests or criteria that would be recommended to implementers before assessing conformance.
• Includes:
• Automated tests (where applicable), and
• Criteria that need to be met before assessing conformance.
• Intention:
• Help implementers prepare for conformance testing
• Assess readiness for an accessibility project.
• Tool by implementers to assess tools or platforms they intend to use (Assessing a new video player widget by the prerequisite would ensure they can confirm it supports captions)

​

14

Key Concept: Sector

Intended to represent the potential for industry-specific accessibility criteria.

Once an organization has achieved Bronze conformance, they may need or want to focus on requirements that specifically address the needs of their user base.

These needs are not in conflict with other accessibility requirements, but are meant to help provide focus when deploying resources. Implementers can use sectors to build a roadmap for Silver or Gold conformance, create plans, and assess strategy.

Current Sectors Proposed:

• Education / Training
• Employment
• Health Care
• eCommerce
• Entertainment
• Utilities / Banking
• Content Creation

​

Possible Uses:

• Outcomes could be tagged with sectors. What is required at Silver and Gold could change.
• Optimized methods could be tagged with sectors and list of recommended methods could be generated

15

Example 1: Alternatives to audio or video content that identify or serve the same purpose as the original content

Prerequisites

• Platform being tested has the ability to:
• Generate captions
• Display subtitles/captions when provided (i.e. can I upload an SRT file)
• Attach a transcript
• Support multiple audio tracks
• Other supporting features for bronze conformance

16

Required Methods

• Captions
• Audio description
• Auto-generated captions and sign language content
• Transcript with:
• Mentioned background sounds/music
• Differentiation between speakers

Optimized Methods

• Sign language interpretation
• Extended audio descriptions
• CART transcription for live events
• Transcript with:
• Identified speakers
• Identified background sounds/music
• Pre-recorded: Transcript is edited for accuracy/clarity
• Previous methods, but during live events and/or for user-generated content

Example 2: Non-text content requires a text alternative.

Prerequisites

• All non-text elements have one or more of the following:
• Alt attribute
• Label
• Title attribute
• Other means of providing an associated text alternative
• Method to indicate they are decorative (where applicable)
• Automated checks are run to assess non-text content (where applicable)

​

Required Methods

• If an automated check is possible, results meet requirements:
• Non-text elements that are decorative are identified as such
• Non-text elements that are informative have a text alternative

17

Required Methods (cont)

• Manual Testing for non-text elements:
• Is there an associated text alternative or label for the element?
• Is the associated text alternative for the element accurate?
• This is not an assessment of quality, the text alternative should provide the user with the necessary information or point to where it can be found
• If the non-text element is decorative, is it properly identified?

Optimized Methods

• Is the associated text alternative informative to the user?
• Usage of at least one advanced method for a complex image or non-text element:
• Extended descriptions
• Figures
• Tactile graphics
• Audio visualization for graphs

How it will overall be evaluated (Levels)

Different methods are required for each level

• All Bronze methods (all are required)
• Silver = Bronze + Silver Required Methods + 50% of optimized methods
• Gold = Bronze + Silver + Gold Required Methods + 75% optimized methods

18

Questions/Issues for Consideration

• In this model, would methods need to be normative?
• Does having required methods along with a percentage address equity challenges?
• How would technologies that do not have a method or are proprietary be addressed?
• Is this simple enough to be acceptable to stakeholders?
• Are sectors required? Educational sites have administrative sections, would the administrative sections have to meet educational requirements?

19

Conformance Option 2

23-2

Read Ahead for Option 2

20

Single Page Summary

• Testability should not dictate level.
• Quantifiable Outcomes, Qualitative Outcomes, and Organization Practices.
• Quantifiable and Qualitative are of equal importance. They just differ in testability.
• Stacked Outcomes allow fewer outcomes and bring attention to Qualitative Outcomes.
• Levels and Requirements
• Bronze - The minimum level of conformance. Nothing below this. All outcomes required for bronze must be testable (assertions are testable) and technically possible. All outcomes marked as bronze must be passed.
• Silver - more accessible. Bronze plus % of additional outcomes (see details in linked section)
• Gold - exceptional efforts. Bronze plus Silver plus larger % of additional outcomes (see details in linked section)
• Matrix Model shows how new AT & new Test can cause Qualitative to become Quantifiable.

​

21

Key Concepts

1. Only Testable Outcomes can be required - since an author has to be able to reliably know if/when they have succeeded.
2. Assertions of a Procedure or Organizational Practice are testable.
3. Qualitative Outcomes are as important and should be considered in parallel with (at the same time as) Quantifiable (Testable) Outcomes.
4. A key component of this Model is to maximize awareness of and incentives to implement Qualitative Outcomes.
5. Qualitative Outcomes are qualitative only until some means for objective measurement is found. Then they become Quantifiable Outcomes.

22

Key Concepts - cont’d

• “Organizational Practices” is introduced as a type of (assertable) provision
• Procedures that are carried out at an organization wide basis
• Auto-Testability is not a level or sub-level - or special set of any kind
• “Stacking of Outcomes” == the inclusion of different levels of accessibility in one outcome (pass, better, exemplary)
• Requirements and encouragements are intermixed
• Complete process must be accessible for any part to be accessible
• Guidelines should be written so that they gracefully transform as mainstream and assistive technologies advance (with subsequent WCAG 3 releases)
• From Qualitative to Quantifiable (as it becomes Testable or tool becomes test)
• From Author manual to Author automatic (as auto gets better than average human)
• From Author to User Agent (covers all sites including non-accessible sites)

23

Examples of Stacking with a) Quantifiable and b) Qualitative Outcomes:

Example for Quantifiable outcomes, of implementation with increasing quality:

• Contrast
• Contrast for running text has a contrast of at least 4.5:1 with any part of the text background behind it unless it is large text in which case the contrast is at least 3:1
• It is pass if the contrast is 4.5:1 with 3:1 for large text
• It is better if the contrast is 6:1 with 4.5:1 for large text
• It is exemplary if the contrast is 7:1 with 5:1 for large text

​

Example for Qualitative Outcomes, of implementation with increasing quality:

• Plain Language
• Content includes a plain language summary of the content
• Pass - Assertion that a version of the W3C plain language guidelines were used to create summary.
• Better - Assertion that a version of the W3C plain language guidelines were used to create the whole component/page
• Exemplary - Assertion that a version of the W3C plain language guidelines were used to create content with multiple levels of plain language

�

24

Examples of Outcomes with MIXED (Quantifiable & Qualitative) Stacking

Example of mixing Quantifiable and Qualitative outcomes, with increasing quality implementation measures:

• Text Equivalents for visual elements
• All information presented in visual form is available in e-text
• Pass is All information presented in visual form is available in e-text
• It is better if the text equivalents are of high quality as described at https://w3.org/wai/textequiv
• (maintained by W3C as a stable reference point for guidance for writing ‘good’ descriptions.)
• It is exemplary if there are both short and complete descriptions of the content of visual elements that contain a lot of information such as graphs or illustrations.

​

Example of mixing Quantifiable and Qualitative outcomes, with increasing quality implementation measures:

• Visual Equivalents for audio or audio-visual
• All speech and important sounds in pre-recorded and live Audio-visual presentations is presentable by general user agents in synchronized e-text.
• Pass is as written above
• It is better if such e-text is visible within .5 seconds of the audio
• It is exemplary if it is positioned and identified to match sound source

25

How will overall conformance be evaluated (Levels) 1/2

• Bronze - the minimum level of conformance. (nothing below this) - conformance claims by URL (or set of URLs), but must include any URLs required to complete process (task completion)
• All items in bronze must be passed for conformance.
• All items in bronze must be Testable
• (question of whether assertions or similar mechanism can be considered at bronze)
• All items need to be technically possible today
• All items need to be able to be applied to all types of web content or scoped to only apply where they should
• All pages/items/views in a process must pass Bronze for any page/item/view in that process to pass

​

• Silver - more accessible (incentivize and reward organizations for going beyond what is requirable) - points-based incentivization system (could be points or percentages or other scoring mechanism, to encourage greater accessibility)
• Must pass Bronze requirements
• Also any or all of the following:
• Require some additional testable requirements; and
• Require some percentage of (stand alone) qualitative / procedural; and
• Require some percentage of Better / Exemplary; and
• Require some specific organisation practices.
• Silver point incentives are largely embedded in layers on the Bronze requirements (stacked provisions).

26

How will overall conformance be evaluated (Levels) 2/2

​

• Gold - exceptional efforts - but note that even Gold will not necessarily be accessible to all types, degrees, and combinations of disability. We don't even know how to do this yet for all types of content.
• Must pass Bronze requirements
• Require additional organisation practices, and higher percentages from Silver
• It is possible that few organizations will pursue a gold level of conformance so this is an opportunity to carve out space for those organizations which might act as role models in this space - and be recognized in an award like manner.

​

Note 1 : There is no level below Bronze but there can be guidance below bronze. There is no “auto-testable” level or sub-level or set.

Note 2 : Qualitative items may become testable with advancing technologies or a new test. (And may shift its level.)

27

Matrix Model:

If it’s testable, it moves left (to quantifiable); if you require it, it moves up

28

Questions/Issues for Consideration at CSUN

• RETROFIT/PROACTIVITY PROBLEM: How might we incentivize “early and often”, since “accessibility later” leaves people cut off out of the gate, costs more, and “later” also usually results in fewer qualitative measures and org practices?

​

• TRADEOFF CONUNDRUMS:
• Is it better to have more accessibility in a level but fewer orgs adopting that level, OR is it better to have easier levels (less accessibility) and more organization reaching that level. Which helps more?
• If doing a % = success. Is it better to have fewer more fundamental provisions OR a larger of more diverse provisions where authors could all choose easier non-fundamental provisions (that apply across disabilities).
• If all people are equally important then aren’t all provisions equally important?

​

• SECTOR/ACTIVITY SPECIFICITY: Where should we (and should we at all) have sector or “content-type” specific guidelines.
• Examples
• Education/training videos with simultaneous audio and visual content that is essential to understanding would require extended audio descriptions.
• Images that have complex information in them that are used in education/training must have long descriptions that provide all essential information information
• Guidelines may have more importance in particular sectors/activities than others, so the absence of a guideline may exclude more end-users if we don’t consider the sector/activities.

29