NISO Workshop on Metadata to Support Accessibility, November 15/16, 2021
NISO Workshop on Metadata to Support Accessibility, November 15/16, 2021
https://asa1cadmoremedia.blob.core.windows.net/asset-c90c45c1-f529-4ed1-bdab-f7fb52f1cc39/NISO Workshop on Metadata to Support Accessibility%2c November.mp4
GEORGE KERSCHER: OK. So my name is George Kerscher. I'm Chief Innovations Officer with the DAISY Consortium, and Senior Officer of Global Literacy with Benetech. I live in Missoula, Montana. We have the University of Montana in town here.
GEORGE KERSCHER: I'm blind. I have retinitis pigmentosa. It's a gradual degeneration of the retina. So over the many years from low vision to total blindness, I've experienced a lot of the different technologies and how people use a computer who have low vision or blindness. So I'm really well-versed in that arena.
GEORGE KERSCHER: So going into the second slide. So the one point here is that when we talk about DEI, I think we should be thinking of people with disabilities as part of the group that needs to be included. I used the term neurodiversity here really because last month I was attending a Microsoft virtual meeting with leaders from the blindness community around the world.
GEORGE KERSCHER: And we had presentations from several people on the Microsoft staff. And they're actively recruiting people that have neuro differences. So the autistic community is one where they're recruiting. And that's probably where the term neurodiversity started. And that neurodiversity term has been embraced by many, many others, including the dyslexic community.
GEORGE KERSCHER: And you can Google dyslexic advantage and read about the many people who have dyslexia. I usually point out that Charles Schwab, the financial genius, is dyslexic. Joe Montana, the great football player. Muhammad Ali is dyslexic. And the one that most people don't know about is Jack Horner, who attended the University of Montana here and flunked out of school.
GEORGE KERSCHER: He's dyslexic. He went on to Montana State University right down the road, 200 miles, and got his doctorate and became the famous paleontologist who developed the theory that dinosaurs were more like birds than anything else. And that many of the birds today are descendants from dinosaurs.
GEORGE KERSCHER: The University of Montana tried to give him an honorary PhD and he refused it. Happily a couple of years later, he was approached again and he softened and he did accept that well-deserved PhD from the University of Montana. But people with dyslexia just struggle with reading. And their perspective on things are different.
GEORGE KERSCHER: And so the people with neurodiversity can bring a lot to the table. So also want to mention that in the DEI movement, people with sensory disabilities, blind, low vision, hearing impaired need to be included. And of course, physical disabilities. So next slide.
GEORGE KERSCHER: What we're looking to do is build a society where publications are born accessible. Right out of the box when you download it, when you read it online, it's accessible. So we're going to focus on people with print disabilities who cannot read standard print. And we want to make sure there's a great reading experience with eyes, ears, and fingers. People without disabilities need to have a great reading experience.
GEORGE KERSCHER: And people who cannot see the information will be reading it or hearing it and reading with their ears, or refreshable Braille with tactile. And I believe this will benefit everybody in the mainstream. What we have to look at is personalizing the reading experience. And this is much different than in the past where we had print, which couldn't be personalized in any way.
GEORGE KERSCHER: We want to be able to make adjustments in the size of the text, the colors, the fonts that are used, the spacing, all of these different visual characteristics that can be modified. And we want to make sure the text to speech can be used. And the improvements in text to speech are absolutely phenomenal. The work that Google and Microsoft and Amazon and others are doing using AI to improve text to speech is just really, really good.
GEORGE KERSCHER: And people love read aloud. So next slide. So I've got it takes two baby. It;s a combination. You've got content that needs to be accessible. And you've got a reading system that presents that information. And it's the reading system that allows the user to make the adjustments in the color, the fonts.
GEORGE KERSCHER: And that would activate text to speech. So take the text that is there and send it to the TTS engine for voicing. So these publications need to be born accessible. And the reading system delivers the reading experience that people will just love. We're seeing many companies developing their reading system and doing a really, really great job.
GEORGE KERSCHER: But unfortunately there are companies out there that really are not doing much in terms of the accessibility or the personalization in those reading system. And as information technology professionals, we need to be aware of these differences. And what are the good and what are the bad. Next slide.
GEORGE KERSCHER: I'm going to talk quite a bit about trust and being welcome in a particular environment. We want to welcome people who have in the past been really unwelcome. We've not been able to provide a reading experience to people with disabilities, but we can today through digital technology. And so we need to build trust.
GEORGE KERSCHER: In that community, they need to trust us. They need to trust the libraries, the bookstores, the retailers, the reading systems that are out there, and the publishers creating the content. So this trust begins with the EPUB 3 specification. It was developed with people with disabilities in the room, at the same room with the technologists and the publishers developing the specs.
GEORGE KERSCHER: And so we made sure it was designed from the get go with accessibility in mind. The publishers are embracing EPUB and they're embracing accessibility. I talked to publishers all the time in many different arenas from trade to textbooks. And we're seeing EPUB being adopted broadly. So VitalSource, a major player in this field and very active in higher Ed, reports that 90% of the top titles that people are using are in EPUB 3.
GEORGE KERSCHER: Now not all titles are in EPUB 3. There's still a lot of titles in PDF. But when it comes to personalization, changing fonts, the size of things, I mean, PDF was not designed to be flexible. We also have the EPUB Accessibility Conformance and Discovery Specification, which is a companion to the EPUB 3 standard.
GEORGE KERSCHER: It integrates with EPUB 3 and previous versions of EPUB. And it's the key to being born access. It builds on WCAG. So we don't reinvent anything. We point to wait WCAG for accessibility requirements. And we go on and add additional items that are specific to publishing that are not part of the web.
GEORGE KERSCHER: And in this specification are the accessibility metadata requirements. So those are in the EPUB Accessibility Conformance and Discovery Specification. And of course, the discovery side of things is the metadata that we want to focus on today. So here's the slide where I roll my eyes and say, oh, my God, are you kidding?
GEORGE KERSCHER: I have to explain to a NISO community the importance of metadata? But there is a little difference here because it's not about the publication. It's not the author title ISBN copyright. It's about the accessibility of that publication. So yes, it's metadata. But it's the kind of metadata that we're not really used to seeing until now.
GEORGE KERSCHER: So next slide. So everybody wants the publication to look good. That's not a question. But how you get there is a big question. So if you've got a block of text that is an image and for some reason the publisher decides to put a picture of a heading, they really want a fancy heading, so they doctored up and put an image of it in there, that's not going to be accessible for modification, for personalization.
GEORGE KERSCHER: And it won't work with TTS. If you've got a table that you say, oh, gee, I really think this table what would work better is a picture, that will defeat the accessibility. You just won't be able to get at that information. We want to make sure that the publishers are using semantically rich information. So EPUB is really just a zip file with HTML and CSS inside.
GEORGE KERSCHER: And the semantics that are built into HTML are essential to use. So use the heading structure that's there. And we can see that it's a heading. It's not just larger in bold, but it's actually a heading. And using my screen reader, I can easily navigate from heading to heading. And by using Cascading Style Sheets, we can apply a different style sheet if we want to change the font, or the spacing, or any of the visual characteristics of the publication.
GEORGE KERSCHER: And it all comes from using HTML and the open web platform. And I think I'll mention here that we're talking mainly about EPUB, but HTML is also a great format for publishing documents. So we might talk about this a little bit in the journal space. But HTML is equally as accessible when done right as EPUB.
GEORGE KERSCHER: So on to the next slide. So the features that you see in the metadata-- and there's a whole flock of these metadata items. I'm not sure how many. I think it's around 15. Screen reader friendly is probably one of the most important. And that's what the end user wants to know is that this is something I can use as a blind person with my screen reader, or with low vision that it is screen reader friendly.
GEORGE KERSCHER: That will also impact any person with wanting to use the read aloud feature because it's the same functionality for read aloud that the screen reader is using. They're both taking the text and sending it to a speech engine. We want to make sure that there's no accessibility hazards. And flashing is one of the big ones that can cause problems with people who have epileptic seizures.
GEORGE KERSCHER: That's a real big one. And one of the schema.org metadata items is accessibility summary. And this is just text that can be written to explain what is the accessibility features that the publisher have put in to that publication. And that must be in every born accessible EPUB outdoor.
GEORGE KERSCHER: So next slide. So we're going to talk a little bit about conformance and the metadata associated with that. ConformsTo is one schema.org item. And these are built into the EPUB. They're actually inside. And what is referenced there is normally a WCAG value of A or AA.
GEORGE KERSCHER: And most countries are settling on AA as the requirement. Certainly in the United States it is. It's in Europe as well. And also the EPUB accessibility. So that pair both WCAG and EPUB accessibility. What's in the EPUB Conformance and Discovery Spec must be there to be born accessible.
GEORGE KERSCHER: Another item is CertifiedBy. And that identifies who reviewed the publication and can attest to its properly accessible. And there's also Certifier Credential, which indicates what is the badge, or the credential that established is the authority of that party.
GEORGE KERSCHER: So moving on to the next slide. So again, we're on the idea of trust. How can you trust this accessibility metadata? The community has come together to build these accessibility standards. And so it really does come from that community. And some of the most important-- Ace by DAISY, the Accessibility Checker for EPUB.
GEORGE KERSCHER: It's freely available, open source. And we encourage all publishers to use this throughout their production process. And we encourage others that have access to the unprotected EPUB to use the Accessibility Checker Use it if you're creating your own publications. The Accessibility Checker will guide you in correcting your errors.
GEORGE KERSCHER: So it'll give you a list of the things that found through the automatic checking and will point you to the knowledge base that is maintained. And it explains in publisher friendly ways how to correct the errors and what is the appropriate way to do things. We have SMART-- the Simple Manual Accessibility Reporting Tool, which will take the JSON output from the Ace Checker, you load it into the SMART tool.
GEORGE KERSCHER: And it'll provide you with a checklist of all the things that need to be checked manually. It filters out things that are not in the publication. So if you don't have video, for example, it won't put that in the list of things you need to be checking manually for correctness. And the automated checking is really a great first step, but it requires manual checking to take it all the way.
GEORGE KERSCHER: So a little bit more about ConformsTo and CertifiedBy. It's the publisher that's making the statement in the ConformsTo. They're attesting that this publication is meeting this level of accessibility, WCAG AA and EPUB 3 requirements. We'd like to see it in every single EPUB out there.
GEORGE KERSCHER: And it has to be there to be called for an accessible. CertifiedBy should be in there. The publisher may self-certify. And that's perfectly fine. But there's a great deal of trust that the community places in a third party neutral review of that publication.
GEORGE KERSCHER: Certainly in higher education, it is really, really advised to have third party sort of certification. So next slide. Global Certified Accessible-- GCA is a program. It's a service for publishers that Benetech is running.
GEORGE KERSCHER: It certifies the workflow from the publisher. So not each title is certified, but the publisher goes through a lengthy process where they submit titles that come through a particular workflow, and it gets reviewed by Benetech's experts. They provide feedback and guidance on how to improve that workflow, so that it's a born accessible publication.
GEORGE KERSCHER: The publisher needs to put three different types of titles through the service and get it approved. It must pass in order to get the GCA certification. At that point, they're allowed to use the mark that Benetech has developed, and use the CertifiedBy property in their publications.
GEORGE KERSCHER: There's a dozen publishers out there that already have GCA certification. And I've seen many of these publications and they are just really terrific. 60 organizations are in process, eight countries and three continents. And every area from higher Ed, k-12, the trades. And there's hundreds of titles out there that are available already, hundreds upon hundreds.
GEORGE KERSCHER: So going to the next slide. So where is this accessibility metadata exactly? So the schema.org metadata is inside the EPUB. And publishers will also create ONIX metadata for the supply chain and getting that information out there. Incidentally, the SMART tool has the feature where it can generate the ONIX metadata from the schema.org metadata inside the publication.
GEORGE KERSCHER: So a retailer like VitalSource, or RedShelf, or ProQuest will be taking the-- EBSCO is what I was trying to think of, will ingest the publication and open up the EPUB and grab the accessibility metadata out of the publication and use that to display. What we want them to do is display that accessibility metadata to end users.
GEORGE KERSCHER: In the Accessibility Checker, it'll also pull out that accessibility metadata and show it to you as something you need to check in your publication. Both RedShelf and VitalSource are displaying accessibility metadata. And we're going to go to the next slide. So the schema.org and ONIX metadata is not intended to be consumed by end users.
GEORGE KERSCHER: And the first company, VitalSource, when they started doing it, did display that schema.org metadata. And it was like, oh, boy, we're going to have to help explain what this metadata is. So we developed the user experience guide for accessibility metadata. And we've got a link to where that is there.
GEORGE KERSCHER: And I've got normal human beings cannot understand this kind of metadata. And with all love in my heart, I don't mean to imply that librarians and people that work on metadata day in and day out aren't normal human beings. But you get what I mean, I think. So it needs to help inform the end user about the publication.
GEORGE KERSCHER: Whether it's a customer looking to purchase, a patron looking to borrow a book, or an educator looking to identify a title they're going to be using in their coursework, they need to know about the accessibility of publication before they buy, before they use it. So next slide. The guide details how to do this.
GEORGE KERSCHER: So there is a principles section, which we don't think is going to change too often. It establishes what are the principles that we want to deliver to the editors. So for example, excess mode sufficient equals textual, means that the entire publication can be consumed via text. The images that are in there have alt text.
GEORGE KERSCHER: And so they can be described to the end user through that alt text. My screen reader reads that alt text. And read aloud functions can present that alt text as well. So what we call this in the user experience guide is screen reader friendly. And that's one thing we want to see in the publication. And the metadata is presented using terms that are understandable to the general community.
GEORGE KERSCHER: So next slide. The techniques-- so I talked about the principles in the user experience guide. There's a technique section for schema.org. So if you're developing your presentation of the accessibility metadata, and you're going to be presenting it from an ingested EPUB, you can use the techniques for schema.org. But if you're using ONIX metadata as the basis for your presentation to end user, we've got another set of techniques for ONIX metadata.
GEORGE KERSCHER: What we don't have is anything for MARC. And so that's something we need to work on. And we're looking for help in this arena, so we can get a set of techniques built for MARC metadata. Next slide.
GEORGE KERSCHER: So VitalSource, RedShelf, they're exposing that metadata. Libraries, we want to see them do it. I don't know of-- well, Niles in Canada is a library for people with disabilities. They're exposing accessibility metadata now, but that's the only library I know of that's doing it. EBSCO is participating with us. I don't think they're showing that metadata yet. And we're just beginning to talk to ProQuest about this issue as well.
GEORGE KERSCHER: And both all of these companies are partners with the accessibility community and with Benetech. So trust throughout the supply chain. I as an end user want to go to a bookstore or a library, read about the accessibility of a publication.
GEORGE KERSCHER: Using the accessibility metadata, I will be able to understand what I'm buying, what I'm getting. I want to be able to trust that accessibility metadata that comes ultimately from the publisher. I trust that the publisher has trained their staff on accessibility how to create publications that are accessible, and how to pick vendors that know about accessibility.
GEORGE KERSCHER: And I trust that the vendors that are out there are using Ace and SMART and the knowledge base to create content that is fully accessible. And they're putting the accessibility metadata into the publication as they build that document. When I see certified by GCA, I have confidence that a third party vendor has certified it.
GEORGE KERSCHER: And there's other companies that are doing certification. I know that Lea in Europe, in Italy is doing third party certification as well. But when I see that it's certified by a third party and I look at their credentials, I have a great deal of confidence that it's going to be so.
GEORGE KERSCHER: So next slide. So in this final slide, and I think we're going to leave that up there for a while, I've got links to a variety of resources on inclusivepublishing.org, where there's links to many of the different specifications and guidelines, and link there to Benetech's GCA program. And with that, I would say I'm ready for questions.