WP Accessibility Day 2026 Preview With Guest Ryan Bracey  🔗 **Source:** [WP Accessibility Day 2026 Preview With Guest Ryan Bracey](https://www.youtube.com/watch?v=z_yg-thde28) > [!note]+ Document Notes > tags: [wordpress-accessibility, wp-accessibility-day, conference-organizing, speaker-selection, wcag-testing, ai-accessibility] > author: Equalize Digital > source: Equalize Digital (Accessibility Craft, episode 182) > reference: https://www.youtube.com/watch?v=z_yg-thde28 > date created: 2026-09-21 09:00 > > > [!info] Generated analysis — not the author's words. Sections below are machine-written > > from the captured source. Only text inside `> [!quote]` blocks is verbatim from the author. > > Do not quote or attribute any other line to the named author. > > # 🏷️ Title: WP Accessibility Day 2026 Preview with Guest Ryan Bracey | Fast Friends Beer Co Vienna Style Lager > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#️-darvo-analysis) - [🎯 Conclusion](#-conclusion) - [🔄 Refactor Appendix](#-refactor-appendix--wp-accessibility-day-2026-preview-with-guest-ryan-bracey) ### 🔗 References - [WP Accessibility Day 2026 preview — Accessibility Craft ep. 182](https://www.youtube.com/watch?v=z_yg-thde28) - [Accessibility Craft show notes and transcript](https://accessibilitycraft.com/) - [Equalize Digital Accessibility Checker](https://equalizedigital.com/accessibility-checker/) - [Second Melody](https://secondmelody.com/) ## 📝 Summary - **What happened.** Accessibility Craft episode 182 previews WordPress Accessibility Day 2026, held online October 7–8, 2026, with Ryan Bracey, the conference's speaker team lead and director of web development and user experience at Second Melody. Bracey walks through the full speaker pipeline: 88 submissions for 24 slots, blind review by a committee of raters, de-anonymization only afterward to check demographic balance, then scheduling in 24-hour UTC coverage. - **Why it matters.** This is a working description of how a volunteer-run accessibility conference makes selection decisions, with numbers attached. It answers, with data from one event, the common objection that merit-first review and speaker diversity trade off against each other: the team reports nearly no post-de-anonymization shuffling was needed this year. - **Key stakeholders.** WP Accessibility Day organizers (Bracey plus two speakers-team members, June on marketing, Joe as a rater); Equalize Digital as producer and sponsor (Amber Hinds, Chris Hinds, Steve Jones, William Patton); 24 selected speakers across 11 countries; a target audience of 3,000 registrants; the American Printing House for the Blind and GitHub as organizations represented on stage. - **Risks and tradeoffs.** Several concrete details in the captured transcript are corrupted by automatic speech recognition — the registration URL, some speaker names, and the gender percentage breakdown. The diversity figures are self-reported by speakers and are not independently audited. The AI-track oversubscription means strong AI submissions were rejected purely on pool size, which is a selection artifact rather than a quality signal. - **Bottom line.** Register free, watch the 16:00 UTC opening session after the keynote, and copy the two-stage review method — blind scoring first, demographic balance check second — for any WordPress team running a CFP. ## 📖 Main Content **The event.** WordPress Accessibility Day 2026 runs October 7–8, 2026, entirely online, and is free to attend. Amber Hinds names a registration goal of 3,000 attendees. This is the event's fifth edition: it started in 2020 and skipped one year. The conference is operated by a 501(c)(3) nonprofit, and donations fund the Zoom Events platform, human (not automated) captioners, and sign language interpretation. The organizers state plainly that donation is optional and attendance is free either way. > [!quote] [00:47:50] > And for anyone who has not yet had a chance, WordPress accessibility day is totally free. You can register. We are a 501c3 nonprofit. > [!quote] [00:47:50] > We have human captioners who come. We have sign language interpretation that we pay for. **The URL is corrupted in the capture.** At [00:47:50] the spoken registration address is transcribed as "wpaccessibility.2026. I think that's a w." The automatic transcription has mangled a domain that contains the year. Treat the exact string as unverified and confirm against the event's own site before publishing it anywhere. Everything else in this note that depends on a spelled-out URL carries the same caveat. **The schedule shape.** The conference runs a continuous 24 hours. Amber gives the US anchor points: it starts at 10:00 a.m. Central, and the final talk runs 9:00–10:00 a.m. Central the next day. The published schedule itself is expressed in UTC, which is the correct choice for a global single-track event — every speaker slot mentioned in the episode is quoted in UTC (16:00, 20:00, 22:00, 13:00, 06:00). Amber pushes back on the assumption that the last slot is a penalty slot. > [!quote] [00:39:00] > and those last couple talks at the end will always have several hundred people on them just like in the beginning. So, it's not bad to be the last talk. **The selection pipeline.** The call for speakers opens around April. The speakers team is three people: Bracey as lead plus two others. Submissions this year totaled 88 — Bracey first recalls 102, then checks and corrects to 88 on air. Against a 24-slot roster including the keynote, that's roughly 3.7 applicants per slot; Bracey rounds it to "about four times as many applicants as we do have slots." > [!quote] [00:16:48] > It was actually 88. So, we got 88 submissions this year, which is pretty good to fill our 24 slot roster minus our keynote. That's 23 speakers. Anonymization goes further than stripping the name field. A person outside the reviewing team reads each submission and redacts identifying details inside the description text — personal names, company names — replacing them with the literal word "redacted." > [!quote] [00:19:38] > So it is very anonymized. It's not just that we don't see your name. It's like we don't know what company they're from. Reviewers score on a 1–5 scale. The rating committee is deliberately broad because raters have known biases: Bracey, Amber, and Joe gravitate to technical talks; marketing-team reviewers score case studies, personal stories, and physical-space accessibility talks higher. Spreading the panel across those preferences is the mechanism that keeps the program from collapsing into one genre. Only after scoring does the team de-anonymize and review the selected set for demographic balance. This year, Bracey reports, the merit-ranked list needed almost no adjustment. > [!quote] [00:23:49] > just based on the merits of the talks, which I think counteracts some of the arguments you hear sometimes around being intentional about having diverse speakers and it should just be about the merit. **The diversity figures.** For 24 speakers including the keynote: 11 countries represented, 3 first-time speakers, 17% part of the LGBTQ community, 17% non-white identifying, 46% living with a disability, and 1 speaker identifying as living in poverty. The gender breakdown is garbled in the capture — the transcript reads "if three first-time speakers of our speakers identify as male, 50% identify as female percent identify as non-binary," which has clearly dropped at least two numbers. Only the 50% female figure survives intact. > [!quote] [00:22:14] > So this year of our 24 speakers including our keynote we have representation from 11 different countries. Three of our speakers have never spoken before at all. > [!quote] [00:22:14] > 17% of our speakers are part of the LGBTQ community. 17% are non-white identifying, 46% of our speakers identified as living with a disability. The 46% disability figure is the one Chris singles out as the meaningful number for an accessibility conference, and both hosts connect it to the disability-rights principle of nothing about us without us. Amber adds a counterintuitive operational detail: at WP Accessibility Day the balancing problem runs the opposite direction from most tech conferences. In at least one prior year the top-scoring anonymous submissions were all from women, and the team went back through the pool looking for men. > [!quote] [00:25:11] > is so interesting because it's like the opposite of most tech conferences are like trying really hard to get more female speakers. **The AI oversubscription effect.** The organizers signaled interest in AI topics ahead of the CFP, and AI became the largest submission category. That raised the internal competition bar inside that one track. > [!quote] [00:18:56] > It's almost like if you picked that, you had a lower chance of getting selected. Chris Hinds submitted an AI talk — accessibility best practices inside AI workflows, prompts and guardrails for accessible output, pitched at beginners with enough depth for intermediates new to AI — and was not selected despite a 5/5 score from Amber. Two Equalize Digital staff, Steve Jones and William Patton, were selected through the blind process. Amber says she correctly guessed Steve's submission from internal context but had no idea which was William's, which she offers as evidence the redaction works. **The first perfect-score talk.** For the first time in the event's history, every reviewer scored one submission 5/5: "Accessibility is not enough: what we learned building the website for The Dot Experience," presented by Melody and Carl, the director of strategy and creative technical director at Field Trip, an agency Bracey places in Kentucky. They built the website for The Dot Experience, a museum for blind people and people with vision disabilities, being built by the American Printing House for the Blind. Danielle Burton, a blind screen reader user connected through APH, joins as a co-speaker and guided the build. The talk runs at 16:00 UTC, directly after the keynote. > [!quote] [00:28:40] > they are in the 16 UTC time slot. So, they will be on directly following our keynote speaker. The talk's argument is the one that matters most for practitioners: passing automated checkers and satisfying WCAG success criteria is a floor, not the goal. > [!quote] [00:30:06] > And that was their whole crux of we know what is technically compliant, but what's actually compliant for people Amber makes the same distinction from the host chair, naming WCAG explicitly. > [!quote] [00:30:06] > for the fact that building a really good accessible experience is not just about compliance or like passing, you know, the checkers or a checklist of any sort like web content accessibility guidelines. **The keynote.** Ed Summers, head of accessibility at GitHub, delivers "The Renaissance of Achievement for People with Disabilities." Summers is blind, lives in North Carolina, and works with a guide dog named Loki, a German Shepherd who appears in his speaker profile photo and is described in the photo's alt text. The keynote is the one slot the organizers curate directly rather than selecting from the CFP; prior keynotes were Lainey Feingold on accessibility law two years ago and, last year, a speaker on German accessibility law (the name is corrupted in the capture as "Vidily Freriedman" — most likely Vitaly Friedman). June, who leads the marketing team, recommended Summers after seeing a viral post about an airline passenger objecting to his guide dog and another passenger who swapped seats and befriended him. Bracey had no shared network with Summers and cold-messaged him on LinkedIn. > [!quote] [00:32:45] > So I cold reached out to him on LinkedIn. Just sent him a message saying who I am. The keynote's thesis is a historical pattern argument: every new computing wave arrives inaccessible and disabled users spend years catching up. > [!quote] [00:33:26] > He dives into how with the emergence of every new technology, people with disabilities, especially users who are visually impaired, are typically left behind and have to catch up. Bracey walks the examples: graphical interfaces weren't designed for people who can't see them, and it took years before markup became usable with assistive technology; then the Flash era. > [!quote] [00:34:11] > So he said there was about 10 years there where the internet was just mostly unusable for people. Smartphones repeated the cycle. AI, in Summers's framing, breaks it. > [!quote] [00:34:11] > He then dives a little bit into AI, what the emergence of AI means, and how this one is unique in the field because it's a textbased technology. It is very friendly for people with visual disabilities. That is the "renaissance" claim: a text-native interface is the first major platform shift that arrives already compatible with screen readers, and Summers describes using AI tools to close gaps left by the graphical-interface era. **The rest of the lineup mentioned on air.** Kevin Andrews, with a higher-education and WordPress background, presents "You didn't fix accessibility, you automated the problem: AI in the WordPress ecosystem." Simon Miner, a past WordPress Accessibility Meetup speaker, presents "AI, accessibility, and WordPress: deciding what to trust, test, and do yourself," tagged intermediate, which Amber argues is beginner-accessible in practice. India Anderson speaks at 22:00 UTC on "From plugin to product experience: accessibility lessons from improving WooCommerce extension onboarding" — Amber notes Accessibility Checker itself has no onboarding wizard and she's watching for transferable lessons. A speaker whose first name is transcribed as "Santa" presents at 06:00 UTC on October 8 on who owns accessibility after a WordPress site launches, a question Amber frames around maintenance contracts and unannounced marketing-department edits. A speaker transcribed as "Jin Harry" closes the schedule with "Real world remediation pushback"; she DM'd Chris on LinkedIn seeking examples of clients rejecting accessibility recommendations, and Chris had none to offer because Equalize Digital's inbound clients are already convinced. Two Equalize Digital staff are on the program. William Patton, senior plugin developer, speaks at 20:00 UTC on "WCAG and CI/CD: catch accessibility bugs like any other bug," rated advanced. Steve Jones, CTO, speaks at 13:00 UTC on "When does page zoom become an accessibility failure?" with Amber MCing. > [!quote] [00:42:59] > So, shout out to our own senior plug-in developer William Patton who will be speaking at 20 UTC WCAG and CI/CD catch accessibility bugs like any other bug. It is an advanced talk. Talks are deliberately spread across audience levels and job roles — beginner through expert, technical, business, design, and personal-narrative, including a talk on neurodivergence. > [!quote] [00:45:05] > So no matter where you are in your journey of accessibility, if you've maybe never even heard about it until today, there's a talk for you. **Technical specifics actually discussed.** WCAG is named repeatedly, and the success-criterion framing appears when Amber distinguishes "did I pass this success criterion" from real usability. Page zoom as an accessibility failure mode maps to WCAG 2.2 success criteria 1.4.4 Resize Text and 1.4.10 Reflow, though the episode doesn't cite numbers. CI/CD accessibility testing is named as a talk topic but not demonstrated. Screen reader use is described through Danielle Burton's role and Ed Summers's work. Alt text appears as a practice example — Summers wrote his own alt text describing Loki in his speaker photo. Automated checkers are discussed as a floor. The Accessibility Checker plugin is described in the sponsor read: real-time scans on save, dashboard reports, a front-end view, unlimited post and page scanning in the free tier, and bulk site scanning in Pro. **No ARIA attributes, roles, or patterns are discussed anywhere in this episode**, and no specific WordPress core accessibility ticket, component, or release is named. Anyone expecting core-team specifics from this source will not find them. **Contested and uncertain claims.** The merit-versus-diversity argument at [00:23:49] is a single-year observation presented as a general counterexample; it's suggestive, not conclusive. The speaker demographic figures are self-reported. The claim that AI is inherently friendlier to visually impaired users because it's text-based is Bracey's summary of Summers's argument, not a direct claim, and it sets aside the accessibility of the AI product interfaces themselves. The Field Trip agency location ("Kentucky, I believe") and Simon Miner's repeat-speaker status are both stated with explicit uncertainty by the speakers. **The beverage frame.** The show's format pairs each episode with a craft beverage; this one is a Vienna-style lager from Fast Friends Beer Co of Austin, Texas, 5.2% ABV, sold in a one-pint can, chosen for an Oktoberfest theme. All three rate it two thumbs up. It occupies roughly the first thirteen minutes. It's the show's established structure, not padding, and it's noted here once. ## 🗒️ Notes #### ⏱️ [00:00:47](https://www.youtube.com/watch?v=z_yg-thde28?t=47) Episode framing and housekeeping - Episode 182 of Accessibility Craft; show notes and full transcript published on the show site. - Hosts Amber Hinds and Chris Hinds; guest Ryan Bracey. - Hosts ask for likes, subscribes, comments, and reviews, tying reach directly to accessibility awareness. #### ⏱️ [00:01:33](https://www.youtube.com/watch?v=z_yg-thde28?t=93) Ryan Bracey's background - Director of web development and user experience at Second Melody, an agency in New Jersey. - Runs a department covering website strategy, design, build, and launch; works in-house alongside copywriters and brand designers. - Path into the community: web design work sparked accessibility interest → WordPress Accessibility Meetup (run by Equalize Digital) → WP Accessibility Day volunteer → organizer → speaker team lead for roughly three years. - Names imposter syndrome unprompted when praised. #### ⏱️ [00:02:59](https://www.youtube.com/watch?v=z_yg-thde28?t=179) What the speaker team lead does - CFP opens around April; submissions reviewed anonymously to produce unbiased rankings. - After selection, de-anonymize and build a schedule balanced across demographics, race, gender, and ability. - Ongoing work: email support through submission and onboarding, answering questions, recording sessions with speakers. #### ⏱️ [00:04:34](https://www.youtube.com/watch?v=z_yg-thde28?t=274) Beverage segment - Fast Friends Beer Co Vienna-style lager, Austin, Texas, 5.2% ABV, one-pint can, chosen for an Oktoberfest theme. - Discussion of the can's fall color palette and argyle-in-the-shape-of-Texas motif; Bracey, a designer, approves. - Tasting notes: toastier than expected, fuller-bodied than expected for a lager, muted citrus nose, moderate carbonation, amber color, head dissipates fast. - All three rate it two thumbs up. #### ⏱️ [00:07:32](https://www.youtube.com/watch?v=z_yg-thde28?t=452) The merch store as an accessibility experiment - shop.equalizedigital.com sells Accessibility Craft and "make WordPress accessible" swag. - Lifetime profit roughly $20. > [!quote] [00:07:32] > It was really more of just a experiment to see how accessible Woo Commerce is out of the box. #### ⏱️ [00:13:47](https://www.youtube.com/watch?v=z_yg-thde28?t=827) Conference dates and registration goal - October 7–8, 2026, online. > [!quote] [00:13:47] > So, talking about hanging out on October 7th through 8th, we are all going to be hanging out online, the three of us here and let's call it 3,000 of our other accessibility friends if we hit our registration goal #### ⏱️ [00:15:57](https://www.youtube.com/watch?v=z_yg-thde28?t=957) Sponsor read — Accessibility Checker - Real-time scans on each save; reports in the WordPress dashboard; a front-end view for hard-to-locate issues. - Free tier scans unlimited posts and pages; Pro adds bulk site scanning from 10 to 10,000 pages. - Coupon code "accessibilitycraft" for 10% off any plan. #### ⏱️ [00:16:48](https://www.youtube.com/watch?v=z_yg-thde28?t=1008) Submission and slot numbers - Speakers team: Bracey plus two others, one of the smaller teams. - 88 submissions (corrected live from a first guess of 102) for 24 slots including keynote. - Roughly four applicants per slot; many strong talks cut. #### ⏱️ [00:17:31](https://www.youtube.com/watch?v=z_yg-thde28?t=1051) Chris's rejected AI talk - Topic: accounting for accessibility best practices in AI workflows and skill sets. - Substance: which prompts produce accessible output, what guardrails to give the model, feeding context for consistent results. - Aimed at beginners with enough depth for intermediate and advanced practitioners new to AI. #### ⏱️ [00:18:56](https://www.youtube.com/watch?v=z_yg-thde28?t=1136) AI track oversubscription - Organizers signaled interest in AI topics; AI drew the most submissions of any category. - Selecting the AI category therefore lowered an individual submission's odds. - Chris scored well in blind review but lost on pool size; Amber gave his submission 5/5. #### ⏱️ [00:19:38](https://www.youtube.com/watch?v=z_yg-thde28?t=1178) How anonymization actually works - A non-reviewer pre-reads submissions and redacts names, company names, and identifying details inside the description. - Amber guessed Steve Jones's submission from prior internal conversation; had no idea which was William Patton's. - Stated purpose: prevent selecting friends and colleagues, judge on talk merit alone. #### ⏱️ [00:22:14](https://www.youtube.com/watch?v=z_yg-thde28?t=1334) Speaker demographic figures - 24 speakers, 11 countries, 3 first-time speakers. - 50% female; male and non-binary percentages lost to transcription error. - 17% LGBTQ, 17% non-white identifying, 46% living with a disability, 1 speaker living in poverty. #### ⏱️ [00:23:49](https://www.youtube.com/watch?v=z_yg-thde28?t=1429) Merit-first ordering as an argument - Blind merit scoring ran first; demographic adjustment second and minimal this year. - Chris frames the sequencing as a data point against the claim that merit and diversity conflict. #### ⏱️ [00:24:28](https://www.youtube.com/watch?v=z_yg-thde28?t=1468) Fifth year, and the inverted gender problem - Fifth edition; started 2020 with one year skipped. - The team sometimes has to hunt for male speakers after de-anonymization — the reverse of most tech CFPs. - One prior year: all top scores were women, prompting a second pass. - More female volunteers than male in the organizer pool as well. #### ⏱️ [00:26:23](https://www.youtube.com/watch?v=z_yg-thde28?t=1583) Rater bias by design - Bracey, Amber, and Joe rate technical talks higher. - Marketing-team reviewers rate physical-space accessibility, case studies, and personal stories higher. - A wide committee is the deliberate correction for individual taste. #### ⏱️ [00:27:06](https://www.youtube.com/watch?v=z_yg-thde28?t=1626) First perfect-score talk in event history - Every reviewer gave 5/5 — a first across five editions. - "Accessibility is not enough: what we learned building the website for The Dot Experience." - Melody and Carl from Field Trip (director of strategy, creative technical director), agency located in Kentucky per Bracey's uncertain recall. - The Dot Experience is a museum for blind people and people with vision disabilities, built by the American Printing House for the Blind. - Danielle Burton, a blind screen reader user connected via APH, is a co-speaker and guided the build. - Notable method: the team drew design inspiration from the physical exhibit space. - Slot: 16:00 UTC, directly after the keynote. #### ⏱️ [00:30:06](https://www.youtube.com/watch?v=z_yg-thde28?t=1806) Compliance floor versus usability - Amber and Bracey both separate passing checkers and WCAG success criteria from delivering a usable experience. - Bracey corrects himself on air: APH is the organization building the museum, not just a referral contact. - Deliberate programming choice this year: more talks with lived-experience perspective. #### ⏱️ [00:31:24](https://www.youtube.com/watch?v=z_yg-thde28?t=1884) Keynote selection process - The keynote is the single invited slot; brainstorming starts at the beginning of the year, before the CFP opens. - Theme-driven: accessibility law two years ago (Lainey Feingold), German accessibility law last year, real-world lived experience this year. - No shared network with Ed Summers, so Bracey cold-messaged him on LinkedIn; Summers accepted almost immediately. #### ⏱️ [00:33:26](https://www.youtube.com/watch?v=z_yg-thde28?t=2006) Keynote content — the catch-up pattern - Ed Summers, head of accessibility at GitHub, blind, based in North Carolina, guide dog Loki. - Talk: "The Renaissance of Achievement for People with Disabilities." - Pattern: each new technology wave excludes visually impaired users first, then gets retrofitted. - Graphical interfaces weren't designed for non-sighted use; markup took years to become usable with assistive tech. #### ⏱️ [00:34:11](https://www.youtube.com/watch?v=z_yg-thde28?t=2051) Flash era, smartphones, and the AI exception - Roughly 10 years of Flash-era web described as mostly unusable. - Smartphones repeated the cycle on arrival. - AI is framed as the exception because it is text-based and therefore screen-reader-native from the start. - Summers describes using AI tools himself to bridge gaps left by graphical-interface-era technology. #### ⏱️ [00:35:35](https://www.youtube.com/watch?v=z_yg-thde28?t=2135) Loki, alt text, and how Summers was found - Loki does not appear in the talk; she appears in Summers's speaker profile photo, described in the alt text he wrote. - Organizers debated whether to list her as a speaker. - A viral post about an airline dispute over Loki, and a fellow passenger who swapped seats, is how June discovered Summers's work and recommended him. #### ⏱️ [00:37:37](https://www.youtube.com/watch?v=z_yg-thde28?t=2257) Chris's schedule picks - Kevin Andrews: "You didn't fix accessibility, you automated the problem: AI in the WordPress ecosystem." Background in higher education and WordPress. - Chris promoted the schedule on LinkedIn the day it dropped. #### ⏱️ [00:38:17](https://www.youtube.com/watch?v=z_yg-thde28?t=2297) Remediation pushback talk and the last slot - Speaker transcribed as "Jin Harry"; talk "Real world remediation pushback"; closes the schedule. - She DM'd Chris on LinkedIn asking for cases of clients rejecting accessibility recommendations; Chris had none, because Equalize Digital's inbound clients arrive already convinced by an external event. - US anchor times: starts 10:00 a.m. Central, final talk 9:00–10:00 a.m. Central the next day. - Final talks reliably draw several hundred live attendees. #### ⏱️ [00:40:56](https://www.youtube.com/watch?v=z_yg-thde28?t=2456) Amber's schedule picks - India Anderson, 22:00 UTC: "From plugin to product experience: accessibility lessons from improving WooCommerce extension onboarding," covering discovery in the WooCommerce marketplace through onboarding. - Amber's stated motive: Accessibility Checker has no onboarding wizard and the team keeps deferring building one. #### ⏱️ [00:41:36](https://www.youtube.com/watch?v=z_yg-thde28?t=2496) Simon Miner and the trust-versus-test question - Simon Miner, prior WordPress Accessibility Meetup speaker: "AI, accessibility, and WordPress: deciding what to trust, test, and do yourself." - Tagged intermediate; Amber argues it's beginner-friendly and the tag shouldn't deter anyone. #### ⏱️ [00:42:13](https://www.youtube.com/watch?v=z_yg-thde28?t=2533) Post-launch ownership - Speaker transcribed as "Santa," 06:00 UTC on October 8: who owns accessibility after a WordPress site launches. - Open questions the talk raises: client responsibility, maintenance contracts, unannounced marketing-department content edits. #### ⏱️ [00:42:59](https://www.youtube.com/watch?v=z_yg-thde28?t=2579) Equalize Digital's two speakers - William Patton, senior plugin developer, 20:00 UTC: "WCAG and CI/CD: catch accessibility bugs like any other bug," advanced. - Steve Jones, CTO, 13:00 UTC: "When does page zoom become an accessibility failure?" Amber expects to MC. - Amber, a non-developer, recorded William's session and reports it is genuinely technical with actionable content. #### ⏱️ [00:44:25](https://www.youtube.com/watch?v=z_yg-thde28?t=2665) Talk diversity as a separate axis from speaker diversity - Deliberate mix across roles: technical, business, design, personal story. - One talk on neurodivergence; others on lived experience of disability. - Deliberate mix across levels: beginner, intermediate, expert. #### ⏱️ [00:45:48](https://www.youtube.com/watch?v=z_yg-thde28?t=2748) Second Melody and where to find Bracey - Second Melody is a branding studio whose main offering is now strategy — reframing "I need a website" into what the client is actually trying to accomplish. - Services: naming, logo development, visual identity, verbal identity, messaging pillars, social media, physical collateral, and all web work under Bracey's team. - Specializes in pharmaceutical but works industry-agnostic; secondmelody.com. - Bracey on LinkedIn under his name; Instagram personal account and a bread-baking account. - Second Melody designed the WP Accessibility Day logo and website. #### ⏱️ [00:47:50](https://www.youtube.com/watch?v=z_yg-thde28?t=2870) Registration, cost, and what donations buy - Free registration, 501(c)(3) nonprofit, donations optional. - Donations cover the Zoom Events platform, human captioners, and paid sign language interpretation. - Registration URL spoken but corrupted in the captured transcript — verify before republishing. ## 📚 References - [WP Accessibility Day 2026 preview with Ryan Bracey — Accessibility Craft ep. 182](https://www.youtube.com/watch?v=z_yg-thde28) - [Accessibility Craft podcast (show notes and transcripts)](https://accessibilitycraft.com/) - [Equalize Digital](https://equalizedigital.com/) - [Equalize Digital Accessibility Checker](https://equalizedigital.com/accessibility-checker/) - [Equalize Digital merch shop](https://shop.equalizedigital.com/) - [Second Melody — Ryan Bracey's agency, and designer of the WP Accessibility Day identity](https://secondmelody.com/) - [American Printing House for the Blind — building The Dot Experience](https://www.aph.org/) - [The Dot Experience](https://www.dotexperience.org/) - [GitHub accessibility — Ed Summers's team](https://github.com/accessibility) - [WCAG 2.2 — W3C Recommendation](https://www.w3.org/TR/WCAG22/) - [WCAG 2.2 SC 1.4.4 Resize Text](https://www.w3.org/WAI/WCAG22/Understanding/resize-text.html) - [WCAG 2.2 SC 1.4.10 Reflow](https://www.w3.org/WAI/WCAG22/Understanding/reflow.html) - [WordPress Accessibility Meetup](https://equalizedigital.com/wordpress-accessibility-meetup/) - [WordPress Accessibility Team](https://make.wordpress.org/accessibility/) - Registration URL for WP Accessibility Day 2026: spoken at [00:47:50] but corrupted by automatic transcription — confirm from the event's own site before citing. ## 🗣️ Paraphrase - Ryan Bracey runs speaker selection for WordPress Accessibility Day and describes a pipeline most WordPress events don't run: a CFP opening in April, third-party redaction of identifying details inside submission text, blind 1–5 scoring by a committee chosen for divergent taste, and only then a de-anonymized balance check. This year that pipeline turned 88 submissions into 24 speakers from 11 countries, 46% of whom identify as living with a disability, with almost no post-hoc shuffling required. - The conference itself is free, nonprofit, and runs 24 continuous hours across October 7–8, 2026, with the schedule published in UTC. Donations pay for the Zoom Events platform, human captioners, and sign language interpreters rather than for speaker fees or venue. The organizers are aiming at 3,000 registrants. - The programming argument running through the episode is that compliance is the floor. The first talk in event history to score 5/5 from every reviewer — a case study of the website for a museum for blind visitors, built with a blind screen reader user guiding the work — makes exactly that case, and the keynote from GitHub's head of accessibility frames AI as the first platform shift that arrives text-native and therefore usable from day one instead of after a decade of retrofit. - The practical takeaways for a WordPress practitioner are concrete: the AI track was the most crowded and hardest to get into, WCAG-in-CI/CD and page-zoom failure thresholds are both on the schedule with named times, and post-launch accessibility ownership — who's responsible when marketing edits a page — gets its own session at 06:00 UTC on day two. ## 📰 Newsletter **WordPress Accessibility Day 2026: free, 24 hours, and the first perfect-score talk in five years** Block out October 7–8. WP Accessibility Day runs a full 24 hours online, costs nothing, and this year it's programmed from 88 submissions down to 24 speakers across 11 countries. The organizers are chasing 3,000 registrations. Three things worth your calendar: **16:00 UTC, right after the keynote** — "Accessibility is not enough: what we learned building the website for The Dot Experience." First talk in the conference's five-year history to score 5/5 from every single reviewer. Field Trip built the site for a museum for blind visitors, with Danielle Burton, a blind screen reader user from the American Printing House for the Blind, guiding the work. The premise: technically compliant and actually usable are different things. **The keynote** — Ed Summers, head of accessibility at GitHub, on "The Renaissance of Achievement for People with Disabilities." His argument: GUIs, the Flash decade, and smartphones each shipped unusable for blind users and got fixed years late. AI is the first one that arrives text-native. Bonus: Bracey cold-DM'd him on LinkedIn and he said yes immediately. Your network is not the only path to a great keynote. **20:00 UTC** — William Patton on WCAG in CI/CD, rated advanced and confirmed genuinely technical by someone who recorded it. **13:00 UTC** — Steve Jones on when page zoom crosses into an accessibility failure. **06:00 UTC on the 8th** — who owns accessibility after launch, which is the question every maintenance contract dodges. **For anyone running a CFP:** the process here is worth stealing. Someone outside the review team redacts names *and* company names *and* identifying details inside the description body before anyone scores. Reviewers are picked for divergent taste — the technical folks and the marketing folks rate the same talks differently on purpose. De-anonymization happens after scoring, for a balance check. This year that check barely changed anything. Registration is free and the event is run by a 501(c)(3). Donations go to the Zoom Events platform, human captioners, and sign language interpreters — not optional extras when your audience is this one. *Heard on Accessibility Craft episode 182, over a Vienna lager from Fast Friends Beer Co. Two thumbs up, all three hosts.* ## ✅ Action Items #### DevRel - Submit a WP Accessibility Day 2027 talk in April 2027, when the CFP opens — the cycle is predictable, so calendar the reminder now rather than reacting to an announcement. - Avoid the AI category if you want odds: AI was the most-submitted track in 2026 and rejection there was driven by pool size, not quality. Pitch WCAG-in-CI/CD, post-launch ownership, or lived-experience angles instead. - Attend and live-note William Patton's 20:00 UTC WCAG + CI/CD session on October 7, 2026; turn the notes into a DevRel workshop on wiring accessibility checks into a WordPress plugin's GitHub Actions pipeline. - Attend Simon Miner's "deciding what to trust, test, and do yourself" session and extract a decision tree for when AI output needs human accessibility review — that artifact is directly reusable in developer education. - Copy the blind-review mechanic for any DevRel CFP you run: third-party redaction of names and company names inside the description body, 1–5 scoring, de-anonymize only after scoring. - Adopt the divergent-rater rule: staff the review panel with people who rate technical and narrative talks differently on purpose, and track the split so one taste doesn't dominate the program. - Promote the free registration on October 1 and October 6, 2026, with the UTC schedule, not just US times — the event is built around 24-hour UTC coverage. #### OSPO - Policy and governance: publish the two-stage CFP process (blind score, then balance check) as a reusable governance template for open source events; WP Accessibility Day now has five years of practice behind it. - License compliance: no license issue is discussed in this source; skip. - Security management: not applicable to this source. - Software inventory: add Accessibility Checker to the org's accessibility tooling inventory with its actual capability boundary noted — real-time scan on save, unlimited free post/page scanning, bulk scanning behind Pro — so no one assumes it covers manual criteria. - Developer support: fund staff time for accessibility conference attendance on October 7–8, 2026, and treat the 24-hour schedule as an argument for asynchronous catch-up rather than mandatory live attendance. - Community engagement: sponsor or donate to WP Accessibility Day; donations pay for human captioners and sign language interpretation, which is a concrete, nameable line item for a sponsorship report. - Strategic alignment: report the 46% disabled-speaker figure as a benchmark when setting event KPIs; it's a published number from a comparable event. #### Community - Maintainers: watch the 06:00 UTC October 8 session on post-launch accessibility ownership and turn it into a maintenance-contract clause template for agency maintainers. - Contributors: volunteer for WP Accessibility Day. Bracey's own path — meetup attendee → volunteer → organizer → team lead in a few years — is the concrete onboarding story to point new contributors at. - Sponsors: build the sponsorship ask around the three named costs (Zoom Events platform, human captioners, sign language interpretation) rather than generic "support the event" language. - Ecosystem: the organizers publish speaker demographic stats. Ask other WordPress events to publish the same five fields — countries, first-time speakers, gender, LGBTQ, disability — so the ecosystem has comparable data instead of anecdotes. #### WordPress - **Core:** nothing in this source names a core ticket or component. Skip rather than invent a core action from it. - **Core Performance:** not applicable to this source. - **Core AI:** the keynote's claim that text-native interfaces are inherently more screen-reader-friendly is worth testing against WordPress's own AI feature surfaces. Add "does this feature degrade to text?" as an evaluation question for any AI feature proposal. - **Core Program Team:** note the October 7–8, 2026 date on the cross-team calendar so release and triage rituals don't collide with 24 hours of accessibility programming. - **Mobile:** not applicable to this source. - **Playground:** build a Playground blueprint that reproduces a page-zoom reflow failure, timed to publish alongside Steve Jones's 13:00 UTC session on October 7. - **Community:** add the WP Accessibility Day CFP redaction method to the WordCamp speaker-selection guidance; most WordCamps anonymize the name field only, not identifying details inside the description. - **Training:** commission a Learn WordPress lesson on accessibility checks in CI/CD after William Patton's talk publishes; the advanced rating means Learn needs a beginner on-ramp in front of it. - **Training:** build a short lesson on what automated checkers do and do not catch, using the "accessibility is not enough" framing from the 16:00 UTC session. - **Marketing:** amplify the free registration across .org channels the week of September 29, 2026, with UTC times. - **Support:** collect the post-launch-ownership questions from the 06:00 UTC session into a HelpHub article on who maintains accessibility after handoff. - **Design:** review the speaker-photo alt text practice Ed Summers demonstrated — author-written alt text describing a guide dog — as an example for the design handbook. - **Documentation:** document the distinction between passing a WCAG success criterion and delivering a usable experience in the accessibility handbook, citing the 16:00 UTC session once it's published. - **Photos:** not applicable to this source. - **TV:** request that WP Accessibility Day session recordings land on WordPress.tv with the human-captioner captions intact — the event already pays for professional captioning, so the quality bar is already met. - **Polyglots:** 11 speaker countries means a multilingual audience; propose translating session descriptions for at least the top non-English locales. - **Accessibility:** run a WCAG 2.2 AA review pass on the team's own tooling recommendations after the CI/CD session, and add page zoom (SC 1.4.4, SC 1.4.10) to the testing checklist with an explicit failure threshold. - **Meta:** not applicable to this source. - **CLI:** explore a WP-CLI command surface for accessibility scanning that a CI pipeline can call, informed by the 20:00 UTC talk. - **Hosting:** not applicable to this source. - **Tide:** not applicable to this source. - **Openverse:** not applicable to this source. - **Themes:** not applicable to this source. - **Plugins:** India Anderson's 22:00 UTC session on WooCommerce extension onboarding is directly relevant to plugin-directory onboarding guidance; watch it and propose accessible-onboarding guidance for plugin authors. #### The WPOCC - Programs: propose a microgrant or stipend track for first-time accessibility conference speakers. WP Accessibility Day landed 3 first-time speakers out of 24 in 2026; funded speaker coaching could raise that. - Projects: offer The WP Open Community Collective's governance-clinic format to volunteer-run WordPress events that want to adopt the two-stage blind review process. - Funding experiments: pilot a matched-donation campaign for accessibility services at community events, using the Zoom Events / captioner / interpreter cost breakdown as the transparent ledger line items. ### FAIR - Technical independence: session recordings and speaker metadata for a nonprofit conference should be mirrorable with local metadata and provenance, not dependent on a single hosting platform's continued availability. - Policies and covenants: a 501(c)(3) running a free event on paid accessibility services is a good test case for a funding-transparency covenant — publish what donations bought, itemized. - Productized governance artifacts: turn the speaker-selection process described here into a publishable charter artifact (redaction protocol, scoring rubric, balance-check step) other events can adopt without rebuilding it. ## 🧪 Hypotheses 1. Blind review with in-body redaction produces speaker cohorts at least as demographically varied as review that starts from named submissions — testable by comparing the pre-de-anonymization ranked list against the final program across multiple event years. 2. Signaling a desired topic in a CFP raises submissions in that topic faster than it raises available slots, lowering per-submission acceptance odds in the signaled category — the AI track in 2026 is one observation. 3. Rater panel composition predicts program genre mix: panels weighted toward technical practitioners select more technical talks, independent of submission quality. 4. In accessibility-specific events, female-identifying speakers are overrepresented in top-scoring blind submissions relative to general WordPress events — Amber reports having to seek male speakers, which inverts the usual pattern. 5. Final-slot attendance in a 24-hour online conference is not materially lower than opening-slot attendance, because the audience is globally distributed rather than concentrated in one timezone. 6. Text-native interfaces (AI chat) reach screen-reader usability faster after launch than graphical interfaces did — measurable as time from public release to documented assistive-technology compatibility across platform generations. ## 📊 Methodology - **H1 (blind review and diversity):** collect the ranked anonymous list and the final program for WP Accessibility Day across its five editions. Compare demographic distribution pre- and post-de-anonymization. The organizers already hold both datasets; the analysis is a paired comparison, not new data collection. Confound to control: whether the applicant pool itself shifted year over year. - **H2 (CFP signaling):** for events that publish desired-topic lists, compare submission counts and acceptance rates by category against categories not signaled. WP Accessibility Day 2026 gives one data point (AI = most submissions, lower individual odds); five to ten events would give a trend. - **H3 (rater composition):** have each rater self-classify talk preference before reviewing, then correlate individual scores against talk genre. The episode reports this qualitatively (Bracey, Amber, Joe favor technical; marketing staff favor case studies) — quantifying it requires only the existing score sheets. - **H4 (inverted gender pattern):** compare blind-review score distributions by self-reported gender at accessibility-focused events against general WordPress events. Requires cooperation from at least two comparison events. - **H5 (final-slot attendance):** pull per-session live attendance from the Zoom Events platform for the 2026 event and regress attendance against slot position and UTC hour. Amber's "several hundred people" claim is a recollection; the platform holds the actual numbers. - **H6 (text-native accessibility lag):** define "documented assistive-technology compatibility" as the first published screen-reader-usable release, then measure the gap from public launch for GUI-era operating systems, the Flash-era web, smartphone platforms, and major AI chat interfaces. Ed Summers's keynote supplies the qualitative timeline; turning it into a measured comparison is the open work. - **Source limitation to state in any writeup:** all figures here come from a single conversational podcast transcript with known automatic-speech-recognition corruption. The 88-submission figure was corrected live and is reliable; the gender percentages are not recoverable from this capture. ## ❓ Questions - What are the male and non-binary speaker percentages that the transcript lost? The 50% female figure alone doesn't close the set. - Does WP Accessibility Day publish its scoring rubric, or is the 1–5 scale applied to holistic judgment? A published rubric would make the process copyable by other events. - How many of the 88 submissions were in the AI category, and what was the acceptance rate inside it versus outside it? - What happens to strong rejected talks? Is there a meetup pipeline, a waitlist, or a referral path to other WordPress events? - Who redacts the submissions, and how much time does that step cost? It's the operationally expensive part of the process and the detail most likely to stop another event from copying it. - Does the event measure whether the 46% disabled-speaker figure translates into disabled attendee registration and retention? - Are session recordings published with the paid human captions, or are they re-captioned automatically on the hosting platform? - What's the actual registration URL? The capture corrupts it, and every promotion action depends on getting it right. - How does the event handle speaker accommodation requests — is there a budget line for it alongside captioning and interpretation? ## 🔮 Future Research - **A cross-event study of CFP anonymization depth.** Most events anonymize a name field. This one redacts identifying content inside the description. Nobody has measured whether the deeper redaction changes outcomes, and WordPress has enough events per year to answer it. - **Attendance economics of the 24-hour format.** A single-track, 24-hour, UTC-scheduled online event is a specific design choice. Per-session attendance by UTC hour, across five editions, would tell the community whether the format serves global audiences or just spreads a US audience thin. - **The text-native accessibility claim.** The keynote's argument deserves an empirical treatment. AI chat interfaces are text-native at the model layer but wrapped in graphical applications with their own accessibility debt. Measuring the gap between the two layers is unexplored. - **Automated checker coverage.** The "accessibility is not enough" framing implies a measurable gap between checker-passing and user-usable. Pairing automated scan results against user testing outcomes on the same set of WordPress sites would quantify the gap instead of asserting it. - **Post-launch accessibility decay.** The 06:00 UTC ownership talk raises it; no one has measured it. Scanning a cohort of WordPress sites at launch and again at 6, 12, and 24 months would show how fast accessibility regresses under normal content editing. - **Volunteer-to-lead pipelines.** Bracey went meetup attendee → volunteer → organizer → team lead in roughly five years. Whether that path reproduces, and what makes it stall, is an open community-health question. ## 💡 Implications - **Practice.** Agencies should stop treating an automated checker pass as a delivery criterion. Two separate sessions on this program — the perfect-scoring case study and the page-zoom talk — are built on the gap between passing and working. Add manual zoom testing at 200% and a screen reader pass to any accessibility definition of done. - **Practice.** WCAG checks belong in CI, not in a pre-launch audit. A talk is now dedicated to this at an advanced level, which means the technique is mature enough to be taught rather than invented. - **Policy.** Events that publish speaker demographic figures create a comparable benchmark. Once one event publishes 46% disabled speakers, "we couldn't find anyone" stops being a defensible answer for other events in the same ecosystem. - **Policy.** The merit-first, balance-second sequencing is a defensible answer to CFP fairness objections and is cheaper to defend than either quota-first or merit-only approaches, because the order itself is the evidence. - **Security.** Not applicable to this source — no security topic is discussed. - **Education.** The program is stratified by both audience level and job role, which is a structural argument that accessibility education fails when it's pitched at one persona. Training curricula should mirror that: separate tracks for developers, content editors, designers, and business stakeholders. - **Funding models.** A free event funded by optional donations, with costs named as platform, captioners, and interpreters, is a transparent-ledger model other WordPress community projects can copy. Naming the line items makes the ask concrete and makes the spending auditable. - **Hiring and representation.** The keynote comes from GitHub's head of accessibility — a signal that accessibility leadership is a defined senior role at major platforms, which is useful evidence for anyone building a business case for such a role. ## 🧠 Learning Styles Assessment - **Reading level estimate:** Flesch-Kincaid Grade Level approximately 8–9 on the transcript text, with a Flesch Reading Ease in the 60–70 band. Spoken conversation keeps sentences short and vocabulary common; the jargon load is light (WCAG, CI/CD, screen reader, UTC, alt text) and every term is either explained in passing or inferable from context. The automatic transcription's run-on sentence structure inflates raw sentence length without raising actual difficulty. #### 👀 Visual learners - Present: the video format shows three speakers, the beer can's label design gets described verbally, and talk titles are read aloud. Speaker profile photos and the published schedule are referenced but not shown on screen in any described way. - Improve: put the UTC schedule grid on screen while it's discussed; show the speaker demographic figures as a chart rather than reading percentages aloud; display each talk title as a lower-third when it's named. Eleven talk titles and times are spoken with no visual anchor. #### 🔊 Aural learners - Present: this is a podcast — the entire content is audio-first, with conversational pacing, three distinct voices, and audible turn-taking. Theme music by Steve Jones bookends the episode. - Improve: chapter markers at the segment boundaries (beverage, selection process, keynote, schedule picks, closing) would let aural learners skip the 13-minute beverage opening. A 3-minute audio summary of just the conference logistics would serve listeners who want the dates and URL without the conversation. #### 🗣️ Verbal learners - Present: strong. Terms are defined in flow ("we anonymize... it's not just that we don't see your name"), the keynote's argument is retold as a narrative sequence (GUIs → Flash → smartphones → AI), and analogies do real work (the beer-to-memory link, the "nothing about us without us" callback). - Improve: a spoken glossary moment for WCAG, CI/CD, and UTC would cost thirty seconds and serve newcomers. The episode assumes listeners know what a success criterion is. #### 🧩 Kinesthetic learners - Present: little. The content is descriptive, not procedural. The closest thing to a hands-on step is "go register" and "use coupon code accessibilitycraft." - Improve: a companion checklist — register, add the UTC times to a calendar, pick three talks by role, donate — would convert the episode into something a listener can do while listening. A downloadable ICS file of selected sessions would do more for this group than any additional talk. #### 🧠 Logical learners - Present: good. The selection pipeline is described as an ordered process with a clear decision sequence, and the numbers are given (88 submissions, 24 slots, 4:1 ratio, 1–5 scoring, 11 countries, 46% disabled). The merit-then-balance ordering is presented explicitly as an argument structure. - Improve: a flow diagram of the CFP pipeline would make the process copyable. The demographic percentages are read as a list with no denominators restated per item — a table with counts alongside percentages would prevent the ambiguity that the transcription damage exposed. #### 🧑‍🤝‍🧑 Social learners - Present: strong. Three-way conversation with genuine disagreement and self-correction, an explicit invitation to comment and review, and the whole subject is a community event. Bracey's path from meetup attendee to team lead is a social-belonging narrative. - Improve: name a discussion channel for the episode. The hosts ask for comments but don't point to a Slack channel, forum thread, or meetup where the conversation continues after the episode ends. #### 🧘‍♀️ Solitary learners - Present: the full transcript and show notes are published, which supports self-paced reading. Individual talk titles are specific enough to research independently. - Improve: a written session-picker with descriptions, levels, and UTC times would let a solitary learner plan without re-listening. A reflection prompt — which of the 24 sessions matches the accessibility gap on your own site? — would convert listening into planning. #### 🌿 Naturalistic learners - Present: thin but not absent. The seasonal framing is sustained (fall, Oktoberfest, bonfires, leaves, the east coast air off the ocean), and the beverage segment ties taste to place and memory. - Improve: the event's environmental case is never made. A fully online 24-hour conference eliminates travel emissions for 3,000 attendees — that's a concrete sustainability argument the organizers leave on the table, and it would serve naturalistic learners while also being a sponsorship talking point. ## 📊 Accessibility Assessment *This section evaluates the accessibility of the source video and its published artifacts — not the accessibility of the conference being previewed.* #### Screen reader semantics - Present: the episode exists as a YouTube video plus published show notes and a full transcript on the show's own site, which gives screen reader users a structured document alternative to the media player. Episode number 182 is stated aloud, giving a findable identifier. - Improve: the captured transcript shows no speaker labels beyond ">>" markers, which is ambiguous for a three-person conversation. Publish the transcript with named speaker attribution and semantic headings per segment. Maps to WCAG 2.2 SC 1.3.1 Info and Relationships (Level A) for the transcript page, and SC 2.4.6 Headings and Labels (Level AA). #### Keyboard navigation - Present: not assessable from the transcript alone. The YouTube player supports keyboard operation by default; the show's own site was not inspected. - Improve: verify the show site's episode pages for visible focus indicators and a skip link to the transcript, which is the longest block on the page. Maps to SC 2.4.7 Focus Visible (AA), SC 2.4.1 Bypass Blocks (A), and SC 2.4.11 Focus Not Obscured (Minimum) (AA, new in 2.2). #### Contrast and readability - Present: not assessable from the transcript. No on-screen text is described in the capture. - Improve: any burned-in lower-thirds, talk titles, or schedule graphics added to future episodes need 4.5:1 minimum against their background, and 3:1 for large text. Maps to SC 1.4.3 Contrast (Minimum) (AA) and SC 1.4.11 Non-text Contrast (AA). #### Text resizing and reflow - Present: not assessable from the transcript. The video itself scales with the player. - Improve: test the published transcript page at 200% zoom and at 320 CSS pixels of width for clipping — long transcripts are a common reflow failure. Maps to SC 1.4.4 Resize Text (AA) and SC 1.4.10 Reflow (AA). Fitting, given that one conference session is specifically about page zoom failures. #### Media and captions - Present: this is the strongest area. A full transcript is published for every episode, the audio is clean with three clearly distinguishable voices, and the conversation has minimal crosstalk. Music cues are marked as `[music]` in the transcript and non-speech sounds like `[clears throat]` are annotated. - Improve: the captured transcript contains multiple material errors — the registration URL, "Vidily Freriedman" for a speaker's name, "Jin Harry," "Santa," "Cliff" for what is likely a different word, and a dropped clause in the gender percentages. If the published captions carry these same errors, a deaf or hard-of-hearing viewer cannot retrieve the event's URL or the speakers' names from this episode. Human-correct the caption file, at minimum for proper nouns, URLs, and numbers. Maps to SC 1.2.2 Captions (Prerecorded) (A) — accuracy is part of the requirement, not just presence. - Improve: no audio description track is indicated. The episode is talking-heads plus a beer can, and the hosts describe the can's design verbally, which handles most of the visual content — an extended audio description is likely unnecessary here. Maps to SC 1.2.3 and SC 1.2.5 Audio Description (Prerecorded) (AA). - Improve: the transcript is referenced as available on the show site but no downloadable format is mentioned. Offer a plain-text or Markdown download. #### Forms and interactions - Present: the only interaction the episode drives to is registration and donation on the conference site, and a merch store the hosts built partly to test WooCommerce's out-of-the-box accessibility. - Improve: the registration form and donation flow should be verified against SC 3.3.2 Labels or Instructions (A), SC 4.1.2 Name, Role, Value (A), SC 2.5.8 Target Size (Minimum) (AA, new in 2.2), and SC 3.3.7 Redundant Entry (A, new in 2.2). The hosts' own WooCommerce store is a reasonable place to apply the same checks, since they framed it as an experiment. #### Standards reference - The material gaps in this source map to a short list: SC 1.2.2 Captions (Prerecorded) for caption accuracy, which is the one real failure risk here; SC 1.3.1 Info and Relationships and SC 2.4.6 Headings and Labels for transcript structure; SC 1.4.10 Reflow and SC 1.4.4 Resize Text for the transcript page at zoom; SC 2.4.1 Bypass Blocks for skipping a long transcript. Nothing in the source indicates a Level A structural failure in the media itself — the gap is transcription accuracy, which is fixable with a human editing pass over proper nouns, numbers, and URLs. ## 🌍 DEIB Assessment #### Cultural breadth - Present: 11 countries represented among 24 speakers, and the schedule is published in UTC with 24-hour coverage specifically so the event isn't a US-timezone event with international stragglers. Prior keynote programming covered German accessibility law, and the organizers describe actively seeking voices "from all around the world." - Improve: the episode itself anchors times to US Central and only converts to UTC when reading specific slots — a pattern that quietly recenters a US audience. Lead with UTC. No multilingual option for the event or the podcast is mentioned; session descriptions in the top non-English locales would extend the reach the speaker roster already implies. #### Inclusive language - Present: careful throughout. Speakers use "people with disabilities" and "living with a disability," "blind individual," "non-binary," and "LGBTQ community" without stumbling. Danielle Burton is described as "a blind person who uses a screen reader," which is accurate and non-euphemistic. Jargon is defined in flow. - Improve: some idioms won't translate for a global or non-native-English audience — "deep down the rabbit hole," "pull out all the stops," "like a moth to a flame," "bookending," "slide out of your chair." Accessibility conferences draw international audiences; a plain-language habit costs nothing. #### Neurodiversity support - Present: the conference programs a talk on neurodivergence, which the hosts flag. The episode has a consistent, predictable structure listeners can rely on: intro, guest background, beverage, sponsor, topic, closing. - Improve: the beverage segment runs roughly 13 of 49 minutes before the conference content starts, with no chapter markers or in-episode signposting of when the topic changes. For listeners who need to budget attention or who came for the conference details, that's a long unmarked runway. Add chapters and a spoken "here's what's coming" at the top. #### Representation - Present: strong, and unusual. 46% of speakers identify as living with a disability, at a conference about disability — the episode explicitly connects this to "nothing about us without us." A blind screen reader user is a co-speaker on the highest-scoring talk. The keynote is delivered by a blind technologist who leads accessibility at GitHub. One speaker identifying as living in poverty is counted and named, which is a category most events don't track at all. - Improve: the quoting balance in the episode itself tilts toward the organizers — Bracey and Amber speak most, Chris moderates. No disabled speaker from the program appears on the episode to describe their own talk. Inviting one of the 24 would put a first-person voice into the preview rather than describing those voices secondhand. #### Psychological safety - Present: notably good modeling. Bracey names his imposter syndrome out loud and the hosts affirm his contribution. Chris discusses his own rejected submission without defensiveness, and the hosts explain the structural reason (pool size) rather than letting it read as a quality judgment. Bracey corrects two of his own errors on air — the submission count and the APH relationship — without any face-saving. Chris flags "not that I want to get political here" before making the merit-and-diversity point, signaling awareness that the topic carries risk for some listeners. - Improve: no content warning or community guidelines are mentioned, and the invitation to comment doesn't come with a moderation statement. On a topic that attracts bad-faith argument — merit versus diversity in speaker selection — stating a moderation policy alongside the call for comments would protect the people most likely to be targeted in the replies. ## ✅ Logical Validity #### Strong reasoning - **Sequencing as evidence, [00:23:49].** Chris's strongest argument is structural rather than assertive: because merit scoring ran first, under redaction, and the demographic check afterward required almost no adjustment, the resulting diversity cannot be attributed to the balance step. This is a well-formed causal argument — the ordering of operations rules out the alternative explanation. It would be stronger with the actual delta stated: how many swaps, out of 24. "Hardly any shuffling" is the weakest link in an otherwise tight chain. - **Redaction depth as a validity argument, [00:19:38].** Amber's supporting evidence is the right kind: she could identify Steve's submission only because of independent prior knowledge, and could not identify William's at all. That's a natural experiment on the redaction's effectiveness, with a positive and a negative case, from inside the process. Strong because it's falsifiable — if she'd identified both, the claim would have collapsed. - **Panel composition as bias correction, [00:26:23].** The reasoning here is explicitly mechanistic: individual raters have known genre preferences, so widening the panel across those preferences produces a program that no single preference could have produced. This is a causal design argument, not an appeal to fairness, and it's the most transferable idea in the episode. - **Historical induction in the keynote summary, [00:33:26] and [00:34:11].** Three prior cases (graphical interfaces, the Flash-era web, smartphones) establish a pattern, and the fourth case (AI) is argued to break it for a stated reason — text-native interfaces don't require visual translation. This is proper inductive reasoning with a named disanalogy rather than a blind extrapolation. Clarifying it further would mean separating the model layer from the application layer: the argument holds for text generation and is much weaker for the graphical applications wrapped around it. - **Live self-correction, [00:16:48] and [00:30:46].** Bracey corrects 102 to 88 by checking a document mid-answer, and corrects his description of APH's role. Both raise the reliability of everything else he says, because they demonstrate he's reporting rather than performing. #### Fallacies - **Hasty generalization — [00:23:49].** Chris generalizes from one year at one event ("you did the merit first and then you barely had to adjust") to a counter-argument against merit-versus-diversity skepticism in general. One event-year is a data point, not a refutation, and this event has an atypical applicant pool — an accessibility conference self-selects for disabled and underrepresented applicants in a way a general tech conference does not. Chris partly hedges it as "maybe a data point," which softens but doesn't remove the leap. **Fix:** state the scope explicitly — "at an accessibility conference, with this applicant pool, merit-first ordering produced a diverse program five years running" — and publish the year-over-year adjustment counts so the claim rests on five observations instead of one. - **Anecdotal evidence — [00:19:38].** The redaction's effectiveness rests on two cases from one reviewer: Amber identified one submission and missed another. Two anecdotes from a single participant can't establish that redaction works at scale, and the reviewer with the most internal context is the least representative test case. **Fix:** after each cycle, survey all reviewers on how many submissions they believe they identified and check those guesses against the real authors. That turns an anecdote into a measured leakage rate at almost no cost. - **Appeal to authority (mild) — [00:32:45].** "He is the head of accessibility at GitHub. So, it's safe to say he knows what he's doing." The job title is offered as sufficient warrant for the talk's quality. It's a reasonable heuristic and Bracey backs it elsewhere by saying he recorded the talk and found it good, but the inference as stated runs from title to competence. **Fix:** lead with the substance he heard in the recording — the historical pattern argument — and let the title be context rather than proof. - **Survivorship bias in a stated non-finding — [00:39:40].** Chris tells the remediation-pushback speaker he has no examples of clients rejecting accessibility recommendations, because Equalize Digital's inbound clients arrive already convinced. He identifies the selection effect himself, which is good, but the exchange could leave a listener with the impression that pushback is rare. It isn't rare; it's invisible to a specialist firm with self-selected inbound clients. **Fix:** say the second half out loud — the absence of pushback in this dataset is a property of the client acquisition channel, not of the industry. - **Equivocation risk in the keynote summary — [00:34:11].** "This one is unique in the field because it's a textbased technology. It is very friendly for people with visual disabilities." "AI" here slides between the underlying text-generation capability (genuinely text-native) and AI products (shipped as graphical web and mobile applications with their own accessibility defects). The claim is true of one referent and shaky for the other. **Fix:** distinguish the layers — "the interaction model is text, which is screen-reader-native; the applications wrapped around it still need the same accessibility work as any other interface." - **Note on what isn't here.** There's no straw man, no false dilemma, and no ad hominem anywhere in the episode. Disagreement is minimal and handled by adding information rather than attacking a position. #### Tone of voice - Collegial and self-deprecating, with an unusually low defensiveness floor. Bracey volunteers his imposter syndrome at [00:02:59]; Chris treats his own rejection as a data point about pool size rather than a grievance at [00:37:37] ("I am glad it's this one and not mine. zero zero issue with this"); Amber apologizes for possibly having scored his talk low before learning she'd given it 5/5. - Enthusiastic and promotional but not inflated. Claims come with numbers — 88, 24, 11, 46%, 3,000 — rather than adjectives. The sponsor read is clearly demarcated. - Careful on identity language and explicitly conflict-aware. Chris marks the merit-and-diversity segment with "not that I want to get political here," which signals he knows the ground is contested and is choosing his framing deliberately. - Warm and digressive in the first third. The beverage segment is nostalgic in register — bonfires, beaches, a college basement — and functions as rapport-building before the substantive interview. The shift at [00:13:47] is clean. - Precise when it counts. Both Bracey and Amber flag their own uncertainty rather than papering over it: "Kentucky, I believe," "I hope I did not get that wrong," "his name sounds very familiar," "I'm not even going to try and say their last name." ## 🛡️ DARVO Analysis No DARVO pattern is present in this source, and the structure of the conversation makes it unlikely to appear. There's no accusation, no accused party, and no dispute between the participants. - **Deny:** no denials or minimizations of wrongdoing appear. The closest structural analog is the opposite behavior — Bracey corrects his own errors twice unprompted ([00:16:48], 102 to 88; [00:30:46], the APH relationship), and Amber concedes Accessibility Checker's missing onboarding wizard at [00:40:56] rather than defending it. Chris names the selection effect in his own non-answer to the remediation-pushback speaker at [00:39:40] instead of letting the gap stand unexplained. - **Attack:** no attacks on any person or their credibility. The one place where a target exists — people skeptical of intentional speaker diversity, at [00:23:49] — is handled by offering evidence and explicitly declining to escalate ("not that I want to get political here"). The airline passenger who objected to Ed Summers's guide dog at [00:35:35] is described neutrally, with the story's emphasis placed on the passenger who swapped seats rather than on condemning the objector. - **Reverse victim and offender:** no recasting occurs. The disability-community framing runs the correct direction throughout — disabled people are described as excluded by technology design decisions, not as imposing on anyone, and the "nothing about us without us" callback at [00:23:04] centers disabled authority rather than inverting it. Chris's rejected talk is the one place a speaker could have claimed unfair treatment, and he explicitly declines to. - **Emotional cues and rhetoric:** the emotional register is nostalgia (the beverage segment), enthusiasm (the keynote and the perfect-score talk), and mild self-deprecation. No grievance, outrage, urgency, or persecution framing appears anywhere in 49 minutes. - **Related fallacies:** the fallacies identified in the Logical Validity section — hasty generalization, anecdotal evidence, mild appeal to authority — are ordinary informal reasoning errors in a conversational register, not manipulation tactics. None serve a DARVO function. ## 🎯 Conclusion - **Bottom line.** WordPress Accessibility Day 2026 runs free and online across October 7–8, 2026, with 24 speakers from 11 countries selected from 88 blind-reviewed submissions, 46% of whom identify as living with a disability. The programming argument is consistent from the keynote down: passing automated checkers and WCAG success criteria is the floor, and the sessions worth your time are the ones about what happens above it — CI/CD enforcement at 20:00 UTC, page-zoom failure thresholds at 13:00 UTC, post-launch ownership at 06:00 UTC on day two, and the first perfect-scoring talk in the event's history at 16:00 UTC right after the keynote. - **Next steps for a WordPress practitioner.** Register now and pick three sessions by role rather than by topic. Add the UTC times to your calendar directly — the 24-hour schedule means your local slots are scattered. Verify the registration URL from the event's own site, because the captured transcript corrupts it. - **Next steps for anyone running a CFP.** Steal the process: redact identifying details inside the description body, not just the name field; use a rater panel deliberately composed of people with opposing taste; score blind first and check balance second. This event has five years of practice behind the method and reports it needed almost no correction at the balance step. - **Next steps for a team lead or sponsor.** Donations buy three named things — the Zoom Events platform, human captioners, and paid sign language interpretation. That's a concrete sponsorship line item you can put in a budget request and report against afterward. - **What to carry forward.** The single most reusable idea here isn't the conference. It's the claim that merit-first ordering and speaker diversity aren't in tension when the review is genuinely blind — and that the way to make that argument is to run the merit pass first, publish the adjustment count, and let the sequence be the evidence. ## 🔄 Refactor Appendix — WP Accessibility Day 2026 Preview with Guest Ryan Bracey > Density: 2–5 Literature atoms (H3), 0–3 Permanent candidates (H4). > Defaults: 3 Literature, 1 Permanent. Resonance filter: if you can't > articulate in one line why a claim matters, drop it. > > Konik claim-header rules: present tense, declarative, ≤10 words when > possible, searchable, unique, mirrors annotation focus if present. > > Atomicity: one declarative sentence per atom. "and"/"also" → split. > > Tag pool: #bmf · #addendum · #xref · #articleseed · #storystem · #share · #aphorisms ### 📚 Literature Split Pad (H3) #### [H1] Blind CFP review redacts company names, not just names > [!quote] [00:19:38] > So it is very anonymized. It's not just that we don't see your name. It's like we don't know what company they're from. **Context & use**: - WP Accessibility Day 2026 speaker selection, [00:19:38] in the episode; a non-reviewer pre-reads submissions and replaces identifying details with "redacted." - This is the operational detail that separates real blind review from name-field anonymization. Most WordPress events, including WordCamps, do the shallow version. Directly reusable in any CFP process design. **Tags**: #xref #share **Highlight**: [view highlight](https://www.youtube.com/watch?v=z_yg-thde28?t=1178) <!-- atomicity: 1 idea ✓ --> #### [H2] Merit-first ordering makes diversity outcomes self-evidencing > [!quote] [00:23:49] > just based on the merits of the talks, which I think counteracts some of the arguments you hear sometimes around being intentional about having diverse speakers and it should just be about the merit. **Context & use**: - [00:23:49], Chris Hinds on the two-stage process: blind merit scoring, then a de-anonymized balance check that needed almost no adjustment in 2026. - The argument's force comes from sequencing, not assertion. If the balance step changes nothing, the balance step can't explain the outcome. Worth keeping because it's the rare diversity argument that's structurally falsifiable — publish the adjustment count each year and the claim stands or falls on data. **Tags**: #xref #articleseed **Highlight**: [view highlight](https://www.youtube.com/watch?v=z_yg-thde28?t=1429) <!-- atomicity: 1 idea ✓ --> #### [H3] Compliance passes; usable is a different test > [!quote] [00:30:06] > And that was their whole crux of we know what is technically compliant, but what's actually compliant for people **Context & use**: - [00:30:46], Bracey summarizing the first talk in the event's five-year history to score 5/5 from every reviewer — a case study of the website for The Dot Experience, built with a blind screen reader user guiding the work. - This is the framing to use when a client or stakeholder treats a green automated scan as done. It's now backed by a named, recorded conference session rather than by assertion, which makes it citable in a proposal. **Tags**: #xref #share **Highlight**: [view highlight](https://www.youtube.com/watch?v=z_yg-thde28?t=1806) <!-- atomicity: 1 idea ✓ --> #### [H4] Text-native interfaces skip the accessibility catch-up cycle > [!quote] [00:34:11] > He then dives a little bit into AI, what the emergence of AI means, and how this one is unique in the field because it's a textbased technology. **Context & use**: - [00:34:11], Bracey summarizing Ed Summers's keynote. The prior cases: graphical interfaces, roughly 10 years of an unusable Flash-era web, then smartphones — each shipped inaccessible and got retrofitted years later. - Useful as an evaluation question for any new interface: does this degrade to text? The caveat to carry with it is that AI products ship as graphical applications, so the claim holds at the interaction-model layer and not automatically at the product layer. **Tags**: #xref #articleseed **Highlight**: [view highlight](https://www.youtube.com/watch?v=z_yg-thde28?t=2051) <!-- atomicity: 1 idea ✓ --> #### [H5] Signaling a CFP topic lowers that topic's acceptance odds > [!quote] [00:18:56] > It's almost like if you picked that, you had a lower chance of getting selected. **Context & use**: - [00:18:56], on the AI track at WP Accessibility Day 2026. Organizers signaled interest in AI topics; AI became the largest submission category; individual odds inside it dropped. Chris Hinds's AI talk was rejected despite a 5/5 from one reviewer. - Actionable in two directions: as a submitter, avoid the signaled category if you want odds; as an organizer, publishing a wanted-topics list creates a selection artifact you should either cap by track or disclose. **Tags**: #xref #share **Highlight**: [view highlight](https://www.youtube.com/watch?v=z_yg-thde28?t=1136) <!-- atomicity: 1 idea ✓ --> ### 🧠 Atomic Split Pad (H4) #### [P1] Process ordering is the cheapest evidence a selection process can produce **Claim** (one sentence, Courtney's voice): When you run the contested step second and it changes almost nothing, the order of operations itself becomes the evidence that the first step did the work. **Evidence pointer**: [[#H2]] — the WP Accessibility Day team scored blind, then de-anonymized for balance, and reported almost no shuffling in 2026. **Scope & conditions**: Holds when the second step's adjustments are counted and published; collapses into an unfalsifiable claim when "hardly any shuffling" stays qualitative. Also assumes the applicant pool isn't itself the explanation — an accessibility conference self-selects for disabled and underrepresented applicants in a way a general tech CFP does not, so the argument transfers weakly to events with different pools. **Links** (≥1 with annotation): - [[Blind review design]] — because extends - [[Quota-first selection]] — because contrasts on where the intervention sits in the sequence - [[WordCamp speaker selection]] — because applies-in volunteer-run WordPress events <!-- atomicity: 1 idea ✓ --> #### [P2] Anonymization is only as deep as the field you forgot to redact **Claim** (one sentence, Courtney's voice): A review process is blind at the depth of its weakest redaction, so identifying details inside free-text fields determine whether anonymization works at all. **Evidence pointer**: [[#H1]] **Scope & conditions**: Holds wherever submissions include prose the submitter controls — CFPs, grant applications, code review, hiring. Breaks down when reviewers have independent prior knowledge of the submission, as Amber Hinds did with one colleague's talk and did not with another's. Deep redaction costs a dedicated non-reviewer's reading time, which is the practical barrier to adoption. **Links** (≥1 with annotation): - [[Process ordering is the cheapest evidence a selection process can produce]] — because extends - [[Name-field anonymization]] — because contrasts on redaction depth - [[Grant review processes]] — because applies-in funding decisions <!-- atomicity: 1 idea ✓ --> #### [P3] An automated accessibility pass measures the tool's coverage, not the user's experience **Claim** (one sentence, Courtney's voice): A green automated accessibility scan reports what the scanner can detect, which is a statement about the tool rather than about whether anyone can use the site. **Evidence pointer**: [[#H3]] **Scope & conditions**: Holds for every automated checker including Accessibility Checker, axe, and Lighthouse; the gap is largest for criteria requiring judgment — alt text quality, focus order sense, error message clarity, cognitive load. Does not mean automated scanning is low-value: it catches the mechanical failures cheaply and at scale, which is exactly what CI is for. **Links** (≥1 with annotation): - [[WCAG 2.2 AA]] — because extends - [[Accessibility Checker]] — because applies-in WordPress plugin testing - [[Manual accessibility testing]] — because contrasts on what each method can detect <!-- atomicity: 1 idea ✓ --> ### ✏️ Fleeting Queue - [F1] Verify the real WP Accessibility Day 2026 registration URL — the captured transcript corrupts it at [00:47:50] as "wpaccessibility.2026" — before using it in any promotion #fleeting #inbox - [F2] Recover the male and non-binary speaker percentages lost to transcription damage at [00:22:14]; only the 50% female figure survived #fleeting #inbox - [F3] Confirm the correct spellings of three speaker names garbled by ASR: "Vidily Freriedman" (likely Vitaly Friedman, last year's keynote), "Jin Harry" (remediation pushback, final slot), "Santa" (06:00 UTC Oct 8, post-launch ownership) #fleeting #inbox - [F4] Build a UTC-to-local calendar file for the five sessions named in this episode: 13:00 (Steve Jones, page zoom), 16:00 (The Dot Experience), 20:00 (William Patton, WCAG + CI/CD), 22:00 (India Anderson, WooCommerce onboarding), 06:00 Oct 8 (post-launch ownership) #fleeting #inbox - [F5] Open question: does WP Accessibility Day publish its 1–5 scoring rubric, or is it holistic judgment? A published rubric is what makes the process copyable #fleeting #inbox - [F6] Article seed — "Steal this CFP process": deep redaction, divergent rater panel, merit-then-balance ordering, with the 88/24/11/46% numbers as the spine #articleseed #fleeting #inbox - [F7] Test the keynote's text-native claim against actual AI product interfaces: the model layer is text, the application layer is a graphical web app with its own WCAG debt #fleeting #inbox - [F8] Ask the WordPress Training team whether a Learn WordPress lesson on accessibility checks in CI/CD exists; William Patton's advanced talk needs a beginner on-ramp in front of it #fleeting #inbox - [F9] Note for the accessibility handbook: Ed Summers wrote his own alt text describing his guide dog in his speaker photo — a clean example of author-written, contextually-relevant alt text #xref #fleeting #inbox - [F10] Follow up on the Equalize Digital merch store as a WooCommerce accessibility test case at [00:07:32] — did they publish findings? #fleeting #inbox ### → Things 3 (max 2 actions) - [ ] Register for WordPress Accessibility Day 2026 and add the five named UTC sessions to the calendar — verify the URL from the event site first, the transcript garbles it #things - [ ] Draft the "steal this CFP process" writeup from the 88 submissions / 24 slots / deep-redaction / merit-then-balance material and pitch it to the WordPress Community team before the 2027 CFP opens in April #things <!-- deep-read:sections end --> 📄 **Full text:** [[Full Document Contents/Articles/WP Accessibility Day 2026 Preview With Guest Ryan Bracey.md|WP Accessibility Day 2026 Preview With Guest Ryan Bracey by Equalize Digital]] ## Highlights - [00:18:56] category that had the most submissions. It did. It's almost like if you picked that, you had a lower chance of getting selected. >> Totally. Totally. >> Yeah, cuz if I remember correctly, you did score pretty well in our anonymous review. You just were in a large competition pool there. surprised. >> What's funny though, but this is why I think it's good that we anonymize applications because so two people from our team at Equalized Digital, Steve who co-host this podcast and our developer William who's been on a few times for people who listen, they did end up getting selected. I I was able to tell with the anonymous [00:19:38] because like we have someone who's not on the speaker team who's not going to be reviewing it go through and they even like in the description if someone says their name or their company name or some important identifying information in the description they'll like redact it. They'll put redacted. >> So it is very anonymized. It's not just that we don't see your name. It's like we don't know what company they're from. We don't know like all that kind of stuff. and I was able to guess Steve's. I was like, I think this might be Steve because I knew he'd applied and we'd literally had a conversation about Cliff internally as a company because we were ([View Highlight](https://read.readwise.io/read/01m48t8brh2bzgmw0h7gscq06y)) - Note: [H1] Blind CFP review redacts company names, not just names - Tags: atom/literature/h1 - [00:23:04] giving them a diverse perspective of speakers. Yeah, that's really really cool. I I love all of that, but in particular, I love the percent identifying as having a disability stat. >> I think that that for an accessibility conference, that is of particular importance and super great to hear. Super great to hear. >> Yeah. Very much calls back to the without us quote that is the famous rallying cry of accessibility. >> And I like the methodology there, right? because there's the there's the anonymized based on subject matter and description and then there's uh configuration around demographic to make sure you retain diversity. And I think [00:23:49] it's particularly cool or telling that even before you had that second adjustment period, you were already pretty much there >> just based on the merits of the talks, which I think counteracts some of the arguments you hear sometimes around being intentional about having diverse speakers and it should just be about the merit. Not that I want to get political here, but I think the fact that you did the merit first and then you barely had to adjust >> is maybe a data point that people could take away if they find themselves typically skeptical of these kinds of processes. >> Right? You know what I'll say too that ([View Highlight](https://read.readwise.io/read/01m48t8e9tzhrswk8q5vycxyta)) - Note: [H2] Merit-first ordering makes diversity outcomes self-evidencing - Tags: atom/literature/h2 - [00:30:06] for the fact that building a really good accessible experience is not just about compliance or like passing, you know, the checkers or a checklist of any sort like web content accessibility guidelines. and I have not seen that talk. So, I am very much looking forward to tuning in live and I will definitely be watching it live. But I think from what they've said in the description, they're going to talk a lot about like >> what is this really? Because it's not just did I pass this success criterion. >> Yeah. And that was their whole crux of we know what is technically compliant, but what's actually compliant for people [00:30:46] are usable. And I did have a slight error in what I said. Not only did they reach out to the American Printing House for the Blind, that's actually the organization that is building the museum. So that's how they were connected with this project. >> Gotcha. >> But no, Chris, that was a good note you made, too. And we have tried this year especially to get and choose more talks that did have a personal perspective of what living with a disability is like or how to work with those people who live with disabilities and trying to get a more personal perspective on all of that. So it was very wonder if you want to talk about the keynote because we ([View Highlight](https://read.readwise.io/read/01m48t8gp9jxppfpn5msxxe9em)) - Note: [H3] Compliance passes; usable is a different test - Tags: atom/literature/h3 - [00:34:11] we got into a few years there of flash design and no accessibility going on the web. So he said there was about 10 years there where the internet was just mostly unusable for people. Then he said smartphones came out. It was the same thing for a little bit of time. There was they just weren't usable for people who could not see. He then dives a little bit into AI, what the emergence of AI means, and how this one is unique in the field because it's a textbased technology. It is very friendly for people with visual disabilities. And then he goes into why this is a renaissance of achievement and how he himself is leveraging these AI tools to [00:34:53] help kind of bridge the gap left behind by other technological advances for people who can't always see a graphical interface. So really good talk, great guy. I was lucky enough to meet with him a few times before and then record his talk with him and I think everyone's going to really enjoy it. He's a great speaker, great topic. His dog does not show up in the talk. I'm sorry for anyone who got excited when I brought her up. You can see her or hear her description on his speaker profile photo because he did upload a picture of himself with Loki and I think he mentioned that in the alt text when he wrote the alt text for the image. So ([View Highlight](https://read.readwise.io/read/01m48t8mfz5dhyr0wetzmbde2v)) - Note: [H4] Text-native interfaces skip the accessibility catch-up cycle - Tags: atom/literature/h3