Document Accessibility Unraveled
Chad Chelius and Dax Castro during an Accessibility Podcast with the Chax Chat Logo between them.

Guest Leonard Rosenthol: PDF/UA-2, Deep Fake Document Detection, and Acrobat Liquid Mode

Dax Castro
Welcome to another episode of Chax Chat. Join Chad Chelius and me Dax Castro, where each week we wax poetic about document accessibility topics, tips, and the struggle of remediation and compliance. So sit back, grab your favorite mug of whatever, and let’s get started.

Chad Chelius
Welcome, everyone. Today’s podcast is sponsored by AbleDocs, makers of axesWord, axesPDF, as well as document remediation services. So we want to thank them for being our sponsor on today’s podcast. My name is Chad Chelius. I’m an Adobe Certified Instructor, Accessible Documents Specialist, as well as consultant.

My name is Dax Castro. I’m an Adobe Certified PDF Accessibility Trainer, as well as certified as an Accessible Document Specialist by the International Association of Accessibility Professionals. Chad, how you doing man?

Chad Chelius
Good! How about yourself?

Dax Castro
I am trying to stay indoors as much as possible. The smoke here right now with the fires is unreal. The sun, you can stare directly at it. It just looks like an orange sitting in the sky. The clouds are still thick.

Chad Chelius
Well, the smoke is that thick over there. Well, no fire is here on the East Coast, thankfully. And I’m thankful for that. It’s really just kind of rainy and dreary today. So I’ll take what I got.

Dax Castro
I hear you.

Chad Chelius
So what do we got on the docket today, Dax?

Dax Castro
Well, we have a great guest. So Leonard Rosenthol is the Senior Principal Scientist. I didn’t know that Adobe had scientists. But he’s a Senior Principal Scientist with Adobe Systems. And we have him here to talk about a really interesting thing – I think – that a lot of people don’t think of initially when they think of accessibility. We’re going to talk about some of the validations of what makes a document accessible. And you know, we have these accessibility statements and verifications for, “Is a document accessible?” And I think a lot of people just think, “Hey, it passed the Acrobat Checker. So therefore, my document is accessible.” We’re gonna take a little deeper dive. And Leonard has get some great information about that. And he also sits on the PDF Association Committee. And without further ado, I want to introduce Leonard. Leonard, welcome to the program.

Leonard Rosenthol
Hey Dax, thanks for having me and Chad. Good to see you both the gentlemen. I feel sort of out of place here that I’m not certified by anybody to do my job unlike you two, but I guess, since Adobe continues to pay my salary, I’ll take that as the equivalent thereof. Yes, it’s great to be here.

Chad Chelius
Well listen, Leonard, anybody who knows you, knows that you’re beyond qualified to do what you do. Leonard, I’m going to take us back probably 20 years. Leonard and I kind of officially met at the… Was it called the Philadelphia AppleScript Users Group?

Leonard Rosenthol
Oh, there was indeed such a thing. Yes.

Chad Chelius
Do you remember that?

Leonard Rosenthol
I do.

Chad Chelius
So there was a time in my life where I was getting involved in Apple scripting. And Leonard is a very gifted scripter. And there were a bunch of other guys there too. Obviously, we had you on today, Leonard, because of your PDF knowledge. I’ve reached out to you many times over the years with really bizarre questions, I’m sure. But we have you on today because you wanted to talk about the content authenticity within the space of accessibility. And so can you tell us a little bit about that?

Leonard Rosenthol
Yeah, happy to do so. So about two years ago now, Adobe publicly came out. We actually did our first big announcement at MAX. So it’ll be MAX two years ago now, with the upcoming MAX. We made an announcement about a partnership between ourselves, the New York Times, and Twitter to form this organization called the Content Authenticity Initiative (CAI), whose goal is to address the problems of understanding whether the videos, the images, the documents, any of the assets that you consume online are real or not. I mean, we’re all probably familiar with the terms, “fake media, deep fakes”. Those are just one of… They’ve come to sort of encompass the whole thing, but that’s sort of the problem, which is how do you know whether something you were looking at or interacting with is authentic. And that was where the Content Authenticity Initiative started. I refer to it as CAI for short. And that has blossomed in the two years. We now have over 100 members of that organization. We’ve also partnered up with Microsoft, the BBC, ARM, Intel, and others to actually work on international standards in this space. So no surprise, this is something that you tell somebody [that] you’re working on this and it’s like, “Oh my god, where were you? We need to do before COVID. We need to do before the election.”

Chad Chelius
Oh yeah, absolutely. I was curious if they had this for social media, because my goodness! If we need it anywhere, we needed there.

Dax Castro
Well, it’s funny, because at [Adobe] MAX, we had… Oh gosh! And so MAX is in San Diego. So this would have been 2016 or 2017, where they… In the Sneaks, they had someone come out, and they had the audio, a deep fake. And I forget who the guest celebrity was. But they made the guest celebrity say something they had not said at all, because they had downloaded or inspected a bunch of hours of audio of that person, and were able to then just type and it says it in that person’s voice. And it was amazing. And the first thing I thought was, “Oh boy, you can make anybody say anything. How do you do that?” And for those of our listeners, we just want to reiterate that on chaxchat.com we have a transcript of everything we talked about and links to the different items when they are some, and you can find more information about the Content Authenticity Initiative at contentauthenticity.org. Right Leonard?

Leonard Rosenthol
Yes, that’s the place.

Dax Castro
And it was interesting, because in that same Sneak, they talked about the ability to create birds out of anything, like literally birds that had never been invented before. The AI learning was able to just come up with new birds or new briefcases or new dresses or new anything. And then one of the other things that they talked about was the ability to analyze the noise inside a doctored photo to be able to look at that noise and tell the difference between what had been edited and what is not. And that kind of piqued my interest in the whole idea. And that was the last I heard of all that. So it’s great to see that the CAI has set some really great aspirations and got some really great players in the mix.

Leonard Rosenthol
Yeah absolutely, I agree. And let me actually connect some dots here. So I’m going to use your example Dax, to connect this into accessibility if you guys don’t mind.

Dax Castro
Sure, absolutely.

Leonard Rosenthol
Okay. So you talked about the fact that you can have fake audio. You said it. We saw it at MAX quite a couple of years ago. There actually was… I don’t know if folks realize that, but I don’t know maybe four or five months ago now, a new audio file came out of Anthony Bourdain talking about something… Basically, what the network had done was to combine pre-existing snippets of film he had taken with fake – or if you will – or produced audio with stuff exactly as you said. They had his old voice, they took some words, they put it together, and they released the program. Now, obviously everybody knew it was fake because he is dead. I mean, that was not like anyone is trying to fool you, but what’s interesting about that, and goes to the other thing you talked about is that, in addition to having the audio of this podcast, you have a transcript. You got the audio [and] you got to have the text version for people who require it.

Leonard Rosenthol
How do you know the to go hand in hand? How do you know that the transcript is a real value verified – if you will – transcript of the audio? Today, we don’t have a solution to that problem. You sort of take the word for it. Now obviously, people trust you. They go to your site. They trust your site. You say, “These two go together.” They trust that. [And] that’s the case. But what if somehow it gets disconnected? What if somebody downloads the audio stream or downloads the transcript puts it up on their own site? There is no connection be back to the other and there’s no way to say that somebody didn’t modify those things. And this is where content authenticity is going to help us not only in general, but in this specific, using this as a specific example. Because now we can put that stamp of authenticity. You can! You guys, when you put these up on your site can now say, it basically stamps its authenticness, if you will. Okay, this is an authentic transcript. You know, you can have all three of us vouch for the authenticity. You could do the same thing for the transcript that the three of us have vouch for that transcript and connect them. Let’s say these two go hand in hand with each other. In that way, not only will everyone know that, but if anyone tries to modify it, they’ll get warned: “Yo, this has been modified.”

Chad Chelius
Well, you know what’s interesting is, I used to work for the California High Speed Rail project. And we did a lot of meeting notes [for] board meetings and things like that. And people love to take things out of context and add the dot, dot, dots, and formulate a statement that might not, basically, be the truth of the statement that how it was presented. But I can totally see the idea of somebody taking a PDF of meeting notes, and maybe editing or modifying that in a malicious way. And having that stamp of authenticity would be a real comfort. The other thing that I think about though – Leonard, maybe, you can speak to this a little bit more – is, we get documents all the time, where they say, “They’ve got to be signed. They have a digital signature.” And then they come to us to be remediated. And we’re like, “We can’t touch it. It’s digitally signed. I can’t remediate this.” So we always tell people, “Look, we’re going to unlock it.” And we’ll either do a scan PDF or untagged PDF, and then retag it, and then you keep the signature live document on file. And then we just make a note saying, “Hey, live actual signed document is on file.” Is there…? Thinking about what you’re talking about with this content authenticity, are there plans at all in the works for how to get around that to where you can retain the digital signature while still being able to remediate the document?

Leonard Rosenthol
Absolutely. And it’s a great question. So we talked about it from two-fold. So one is people use digital signatures in that context, basically, as a form of locking. They don’t actually want a digital signature. What they want is to be able to put a lock on the file that says, “Either you can’t edit it or if you do, I want to know about it.” And in today’s world, that’s the technology we have in PDF that enables that functionality. But let’s fast forward to a world where we also have content authenticity, because content authenticity gives you all those same features, and then some. So now we’ll hopefully see is that instead of putting this huge heavyweight digital signature mechanism on things, they’ll instead do what they really want to do, which is they will put the content authenticity that says, “I’ve done my remediation.” So like, AbleDocs. So you guys do this for a living. You will be able to say, “When we’re done, we have remediated it.

Leonard Rosenthol
Here’s the information about the remediation. We remediated these 10 things. It might be something like the Matterhorn protocol reports. You put whatever information you want in there. And then it doesn’t lock the file, but it does make it tamper evident. Exactly! So I’ll know if anybody modified it. But it’s okay to modify it. Modification isn’t bad. Because it would be perfectly fine for you to send it back to the customer, the customer makes some modification, when you record that modification, so you can think about – we talked about – provenance. So the provenance here is you started with a remediated file. And obviously, probably earlier than that, somebody made the file, you remediated it; somebody edited it, and then you remediated it again, and all of the record, all of those pieces are in the provenance are in the authenticity. So I think it doesn’t mean not edited. It means I know everything that happened.

Dax Castro
Well, it’s really interesting Leonard that you brought up what was edited in the document. I think that’s actually kind of a really cool thing to be able to say, from a remediation standpoint, we edited the heading levels or we added tagged annotations, or what did we actually do to this file, because sometimes that’s important. I mean, if heading levels were different or we remediated, [or] we added tagged annotations, that maybe [an] additional information that wasn’t necessarily in the original file.

Chad Chelius
So since we talking about digital signatures, I mean, isn’t a digital signature in and of itself a form of content authenticity?

Leonard Rosenthol
Yes and no. So our content authenticity technology uses digital signatures. It’s absolutely a part of what we do. But the thing is, digital signatures, in general, serve two purposes. One is its tamper evidence, so I’ll know whether or not something’s been modified. The other thing it does is it identifies the signer. It’s who was it that did the actual sign. Who signed the file? But that’s all it does. And those are very important things, both in general as well as for content authenticity, but it doesn’t also give me any additional information about other things. So for example, we were just talking about what you did. So we call those actions. So what are the list of actions? So we have a defined way to declare the set of actions. It’s very common in PDF and other formats that you merge various pieces together.

Leonard Rosenthol
So in pages and PDF, in video or even in audio, you’re going to take multiple clips, [and] you’re going to bring them together. We refer to these things as ingredients. So you want to know all of the ingredients that make up this new combo. We call it composite. These are terms that everybody is familiar with, because we’re using them in this context. So same thing! Like that’s all information that’s part of the provenance, that’s part of the authenticity, that then gets wrapped up and signed, because – you’re right – at the end of the day that signature is still the core technology that the world relies on for that tamper evident.

Chad Chelius
I mean, I can’t tell you how often I’ve been teaching in Acrobat class, and we’re on the topic of digital signatures, and people are like, “Well, once a person signs it, I need the ability to go into that document and make edits.” And I’m like, “No, we can’t do that because that invalidates the signature.” We could see somebody sign it, and they signed it based on a price of $50,000, and then you change the price to $100,000. We can’t do that. So that’s really interesting. So getting back to how this relates to accessibility, because that’s one of the challenges that we face as document remediators, or in the case of Dax and I teaching people how to remediate documents. We can say that it’s WCAG 2.1 compliant, but there’s no way of making that known to anybody. And kind of going back to your analogy, Leonard, of, you can’t have to take people’s word for it. Now, you can, of course, run it through a checker that will confirm what somebody said, but fundamentally, there is no indicator anywhere in the document. Now, the one exception that I kind of wanted to bring up is, when you make a document PDF/UA compliant, there’s actually a little bit of XML code that gets added to the metadata that basically tells the document is PDF/UA compliant. And when that happens, there’s a little standards panel that opens up in the PDF that says that this is PDF/UA compliant, but what kills me is that there’s a little line under there that says, “Status not verified.” So it’s somewhat reassuring that we think it’s PDF/UA compliant, but the fact that it says “status on verified” is a little off putting. You’re like, “Well, is it PDF/UA compliant or is it not?” Can you speak to that?

Leonard Rosenthol
I can. And actually, it’s great, because those two pieces go hand in hand. So that great question. So yeah, I’ll answer the second one. And given that it’s also my fault that it does those things that way. You definitely have to…

Dax Castro
We found him. We found the person.

Leonard Rosenthol
It is my fault. I did that. I was part of that. Well, or at least Acrobat’s behavior is responsible for that. So the answer actually goes to the PDF/UA specification. So during the development of PDF/UA, one of the – I would say – most contentious item, the most contentious item that came up over and over again, was around validation of UA. And the question specifically was, “Could you do validation of a PDF/UA file without a human in the loop? Could you do it purely by machine?” And the decision was – and I think it was the right decision – no. It is impossible to do PDF/UA validation without a human. And I’ll give you the primary example. You have to have alternate text on images. You have to have a caption. You have to have that alt-text for an image. How does a machine verify that there’s something there? Like if there’s just a space, is that okay? If I put the word “FUBAR”, is that…? Like, a human would never let any of those things through, but the machine, in general… I mean, obviously, remember that we’re a couple of years in… We’re much further along in artificial intelligence and things than we were when we did the standard, but I still think it’s true. Your machines are not going to necessarily give you the same results for that particular validation aspect that a human will.

Dax Castro
Right. You look at Microsoft, and Microsoft’s AI engine for understanding images when it tries to write the auto supplied text, I mean, I don’t know how many images I’ve seen that say, “Image of sunglasses” as the alt-text description, because it just didn’t know.

Leonard Rosenthol
That’s right. Exactly.

Chad Chelius
I feel like the backup on that… I can’t tell you how many times it has told me that it was an image of a close up of a sign. And when Microsoft first started doing it, it was funny, because in addition to the alternate text, then it would put a line that would say, “Description generated with extremely high confidence”. I don’t know if you guys have seen that or not, but it almost kind of broke the rule of why we use alt-text in the first place – you put it little copyright symbol on there at the bottom. So yeah, that’s a really good point. I guess another good example would be tagging. I mean, there is no way that a machine – and we know that from the Acrobat checker – where it always tells you that you need to check the logical reading order, because a machine can’t do it. So that’s the reason behind it.

Leonard Rosenthol
Yeah. And those go hand in hand, if you use the Acrobat accessibility checker, all the ones that forced you to manually fix them are those same ones out of you… We actually go read the text in UA. It says base on something along the lines of a human has to verify that. That’s why we designed the checker. So that answers your question. But now let’s talk about how we get away from that, and we actually don’t have this problem in the future. So the way we’re going to do this is two-fold. So half the work has already been done. The other piece, we were just sort of talking about it, now bring it back in. So the W3C, the WCAG committee actually defined something called the Accessibility Statement. So this is how in the context of the web or an ePUB, you state, “This will use ePUB.” This ePUB is WCAG AA compliant. That’s an accessibility statement that you make, and you have to make it in a certain way, and there’s extra data like who you are and whatever, you can put in that. So you mentioned the PDF Association. So for those of you who don’t know the PDF Association is an organization made up of about 150+ vendors, educational institutions, [and] all sorts of other things.

Leonard Rosenthol
Basically, it’s the PDF industry. It’s a place for the industry to come together and do things: educate users, work on standards, etc. And AbleDocs, and Adobe, and you name them, they’re members of the association. So accessibility is a huge area of interest for the association. And they have a very active set of working groups in that area. So one thing that group did, maybe about two and a half [or] three years ago now at this point is – I know, we all get dates and times a little weird with COVID – but they created the equivalent for PDF called an accessibility declaration. So there’s a document if you go up in a PDF Association website, you can actually download the specification around accessibility declarations. It’s not really used yet. So we built the spec. We said, “This is a great idea.” It’s a little ahead of its time because we are connecting it with PDF/UA-2. Unfortunately, PDF/UA-2 is taking much longer than we expected it to take. Not entirely its fault, because we all saw PDF 2.0 took longer to do than we anticipated, but UA-2 is coming along. UA-2 will mandate the use of an accessibility declaration. So now you’ve got to put one of these things in your UA file. So now we’ve got the new version of UA. we have a standardized way for you to declare this information about it. And third, now we’re going to connect it to content authenticity, because now you’re going to put that stamp of authenticity on it, so that not only are you putting in your declaration but we know who you are and when you did it and then nobody touched it ever since, or if they did, they documented that they did it.

Dax Castro
Well, even in California, they passed AB-434, which mandated that every website has to have an accessibility statement. And I will tell you, there are so many organizations even now [who] put the statement on there, but their websites are still far less than accessible. So more than just the statement, there has to be that keep behind it and the validation.

Leonard Rosenthol
And knowing that it hasn’t been modified [and] websites is the worst, because the whole point of a website is for it to be completely ephemeral. Now you’re pulling libraries from here and CSS from there, and if any of those things… You know, somebody thought they were changing something simple on a CSS string, and it totally broke accessibility of this site way over here. And that’s… Again, but even [with] PDF, we were talking earlier, [has] the same problem. Somebody goes in and uses the editing tools in Acrobat because they’re just so easy, [and] it breaks accessibility without knowing that.

Dax Castro
Well, like in Australia, they… Not Australia, in Canada, they actually have a mandate saying every two years, you have to go in and re-evaluate your website. Every other year, you have to provide a valid accessibility statement stating that you’ve gone through the website and your content still accessible as they’re getting more and more people to conform to the Web Content Accessibility Guidelines. So I think more and more companies and entities overall are really taking more than just a one and done approach to: “Yes, my website is accessible now.” And that’s enough.

Chad Chelius
So Leonard, if I’m envisioning this correctly, so currently, for PDF/UA, we need to add a PDF/UA identifier, which is a little piece of metadata that gets embedded in the document that says, “Hey, I am PDF/UA compliant?” I claim to be… What’s that?

Leonard Rosenthol
I claim to be compliant.

Chad Chelius
Exactly. You’re right. And Dax and I have talked about this before [that] just putting that in there does not mean it’s compliant. You know, that’s one of the problems with that as you can just kind of throw that in there and say that it is, but it doesn’t mean it is. But then from the content authenticity standpoint, is that another chunk of metadata that gets added to the file?

Leonard Rosenthol
It’s a little more complicated than a little bit of metadata, but conceptually, sure. You can think about it more like a digital signature [which] we were talking earlier. You know, at that same level of infrastructure within PDF. Again, I don’t want to get into gory details here, but it’s much more of that a core aspect of the PDF data structure rather than a little piece of metadata you stick on the side would be a better way to say it.

Chad Chelius
And maybe I went out of order. Maybe the accessibility statement is that more a piece of metadata that would get added to the file.

Leonard Rosenthol
Yeah.

Chad Chelius
Okay. So identifier in the case of PDF/UA and then the accessibility statement, but then the content authenticity, that’s going to be a more robust.

Leonard Rosenthol
Absolutely at all.

Chad Chelius
Yeah, that’s going to be more robust. I’m not sure how that’s going to happen. I mean, you know how that actually gets injected into the file. What the mechanism is?

Leonard Rosenthol
So the answer will be, when that happens… You know, there’s the technical aspect, which I don’t think is what you’re asking me. But from a practical standpoint, what I envision is that it happens in the same way that you run your current tools today that say, “Make this file PDF/UA”, and it goes through the same thing. And it’s just going to finish with the content authenticity. Our goal is that it’s just part of it. Like, you can’t not do today. As you said, “You stick into a little UA flag.” This just becomes part of it. You stick it in the flag, and you stick in the accessibility declaration, and you put in the content authenticity. And now you need three bits instead of just one. Well, it’s just the requirements.

Dax Castro
You know, we talked about making the PDF/UA identifier. And I will tell you, if anybody from NVDA or JAWS is listening to our podcast, it would be amazing if you could listen for or analyze the PDF and say, “Hey look, it’s PDF/UA. It’s got a marker there and start taking advantage of some of the rules in PDF/UA that expand the capabilities of a PDF.” I’d love to be able to go beyond <H6> in a PDF. And PDF/UA says you can go beyond <h6> infinitem if the software can detect it. The use of the <H> tag is a great example as well, where no software out there right now recognizes the use of the <H> tag, because they’re looking for the web, H1, H2, H3, but you could totally make the title of your document be the <H> tag and not have to take up one of those precious six heading levels, all by just going, “Hey look, I see your PDF/UA identifier, and I’m going to open up some more doors to allow your document to be richer.”

Leonard Rosenthol
And I’ll go one further, if I might, which is, if any of those folks are listening, please come join the PDF Association and please do participate in the industry, because we’re working to improve accessibility. Those are your customers. These are the people you’re making money on. Come help us do the right thing for all of these people, so that it works better together. That’s what we want at the end of the day. And your participation is going to be a big strong point.

Dax Castro
And for those of you who are listening to the podcast, who are remediators, holy cow, how many time, how many hours a month do you spend, and rounding certain things, adding actual text, creating different structures, because whatever screen reader doesn’t really read the compliant structure the right way, you end up doing all this extra work. And if we could get JAWS and NVDA and Apple voiceover – oh my gosh – that would be great to be able to come into the fold and actually participate in these standards, it would be great. I don’t know… Literally, every time… We run a PDF accessibility Facebook group. We have little over 1500 members. Anybody new coming in, the first thing they ask is, “Is there any screen reader for a Mac besides Apple voiceover? How do I test this? Or I’m only testing with Apple voiceover.” And part of that is Apple voiceover is only so compliant. It’s really great in a web environment or in a native app environment. It is not really designed for a PDF environment, which actually brings me to one thing I wanted to ask you about, Leonard, is the idea of liquid mode that’s coming out for Acrobat. Can you tell us a little bit more about that and where we are with that?

Leonard Rosenthol
Yeah, so it’s actually out. It’s been out to 100% of our customers worldwide. I think, it’s got to be at least 8 to 12 months now, I think it’s been 100%. It’s been quite a while. We had it out earlier in smaller segments, but I think at this point it is 100% worldwide. It’s only available on our mobile clients at this point. You need Acrobat [in] mobile and either iOS or Android. And basically, what it is, is it serves two purposes. Its main purpose is to enable, and it sort of encompasses both, but it’s really about improved consumption of your documents. So we all know that on mobile devices, especially consuming PDF is a real pain: pinch, zoom, scroll; nobody likes it. So what liquid mode does is it gives you an alternative presentation of the same content. So you can think of it – if you will – as a responsive mode. We don’t like to think of it as reflow because reflow is sort of old school – very basic concept. This really goes above and beyond, because what it does is – and this goes well to the accessibility – we’re leveraging that exact same set of technologies. So all of those tags you’re familiar with, and the attributes, and the off tax, and all those things you folks are putting in there to make it accessible are the things that are driving liquid mode.

Dax Castro
It’s interesting, Leonard. While you were talking, I’m like, “You’d have. It’s a thing.” I didn’t realize that it was in the app. So I opened my phone while you were talking. And there it is! There is a liquid dot like a water drop. And I turned it on a PDF in my phone. And holy cow, this is great! It literally restructured the document. I’ve got a bunch of lists in here that now actually fit on the screen, and I don’t have to scroll left and right. The background looks like it maybe got injects; kind of little wonky, but not every documents going to be perfect. But it’s got whole sections. Like I can collapse a next step section that’s inside this PDF. I’m impressed. Definitely, if you’re listening to this podcast right now, and you’ve got Acrobat mobile on your phone, give it a try on an existing document. That’s pretty neat. Now, from a user, from a creator standpoint, do we have to do anything different to our documents to take advantage of this?

Leonard Rosenthol
Such a great question. So hold that thought because I want to get into accessibility first. And then I’ll answer that because I think they sort of go hand in hand.

Dax Castro
Absolutely.

Leonard Rosenthol
So the other thing that you didn’t try, but if you turn on voiceover right now, for example, while you’re in liquid mode, you will get a better accessibility experience exactly for the reason you mentioned. Again, I assumed you’re on iOS, I apologize for…

Dax Castro
I’m not. I haven’t drunk that Kool-Aid. I am an Android user. But that…

Leonard Rosenthol
It was total connect to the Android accessibility API, but it is still better for the reasons you mentioned, which is, the iOS and the Android and the Mac accessibility model is tied to the web. 100%. And so is liquid mode. So because it meant by doing this, we could get that improved accessibility. So people who require accessibility, access to PDFs will find liquid mode, huge! I mean, maybe even more so than you do as a user doesn’t require those accessibility tools.

Chad Chelius
Well, I will tell you Leonard, you and I had a little bit of a spirited debate – I don’t know, it’s been probably six or eight months ago – about liquid mode. And initially, I was like, “Holy cow, we’ve got all this extra stuff. And now we’ve got liquid mode and all.” I will tell you, my opinion has changed just from actually seeing it in action. It definitely creates a better user experience just from the get-go. And you know, I’ve done… You know, as I’m scrolling through this, as I scroll past the heading level, the top of the page actually changes and updates the heading level action. That’s pretty dang slick, man. That is really cool.

Leonard Rosenthol
There’s a lot of very interesting subtle things you will discover as you play with more and more and different document types.

Chad Chelius
Yeah, absolutely.

Leonard Rosenthol
And it will continue to evolve. So we’ve been working again with the association. And this is another place where this accessibility declaration comes into play, which is, if you’ve got a file with an accessibility declaration in it, you can now state in addition to I prepared it according to WCAG AA or a PDF/UA-2 that I’ve done it in accordance with the rules of Adobe Liquid Mode. So we’re going to publish, and in fact, we’re working very closely with the association that had been for quite a while now to publicly publish those rules, because we want them to be open and available, so everybody can do it. So we absolutely want people to be able to prepare documents that so we don’t have to do anything. We want to take it and just boom. Here’s the great new experience. And we don’t have to do anything behind the scenes. But we need to know that you’ve done your side of the work. So it’s not our fault, if you will.

Dax Castro
And it goes back to preparing the document correctly. And this is the thing that Chad I talked about every podcast, that literally 90% of accessibility is, “Did I create a file the way the program was intended? Did I use this the right way? Am I using styles? Did I use the right containers? Am I using tables to layout my three column layout rather than actual columns, [and] those types of things?” That’s why it’s so important for everyone listening to really understand how the program work. And whether it’s Word or InDesign or PowerPoint or whatever it is that you’re using to get the most out of it, because then it’s more flexible. Now you don’t have to, Leonard doesn’t have to create an SDK for us to like have this whole other language that we need to know. We just need to have created the document correctly in InDesign, and it’s going to be pretty accessible coming out.

Leonard Rosenthol
And again, it doesn’t even have to be InDesign. And again, it doesn’t have to be Adobe tooling. It’s part of what we want. It’s not anything coming from Adobe because PDF… We didn’t talk about this before, but just many people don’t always remember, PDF has been an open standard since 2008. So yeah, 13 years now. Keep that in mind. So it’s not about Adobe. We don’t want people to always think that way.

Dax Castro
Chad and I would love to… I mean, literally, I would feel like my life is fulfilled if we had any influence on Adobe to make a button that says, “Match the reading order to the tags order.” And it’s just a one click or a preflight fix up that says, “Walk through the tag structure and make the read order the same thing.” Because we spend hours and hours and hours. There are so many associations out there, mostly, in Canada and some other places. HHS has a standard that says, “The reading order panel must match the tag tree.” And so right now you try to edit that and it starts playing havoc with the tag tree. We’d love to be able to just have a one click solution where it just matches both of them or goes away entirely.

Chad Chelius
Leonard, do you have any insight as to why we have two orders at all?

Leonard Rosenthol
Yeah, it’s a very easy answer. And the answer is simple. PDF started its life as pure reading order, or technically as content order. Because it’s the order in which everything happens on the page: draw this, draw this, draw this, draw this. So for the first 10 years of its life, give or take, that’s all PDF did. Then [in] PDF 1.3, 1.4, 1.5, we started bringing in tags you PDF and structured PDF, at which point, those things got added on. I hate to use the word bolted on. I’ll say added on. And so now you have this set of information and this set of information. And they are separate. I mean, they are in a lock and they only go together, obviously. But they are still… And this is… And we have debates about this in the standards body as well. Because you’re right, if you run through the content or the reading order, and you run through the tags order, you can end up with two very different sets of things. And the fact is that, admittedly, sometimes one is better than the other. You can’t say thou shalt always use tags or thou shalt always use content.

Dax Castro
Well, I’d love to be able to say, if you mess with the reading order, don’t mess with my tags tree. You know, one of the things we’ve been teaching a lot recently is that we found out that if you use the content panel to set your read order, not the read order panel, but the content panel, the tags don’t change. But if you start going into the read order panel, and you have those nice little numbers that say 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, but you’re like, “Oh, well 7 needs to be 3”, you drag that up, then you go back and look at your tags, all of a sudden, two tags have been combined now. And there’s an empty container tag where that used to be. And so we’ve been teaching people, “Look, stay out of the read order panel. Just edit the content panel.” You’ll get what you want, which is the read order to be restructured and the tags to stay the way that you’ve made them.

Leonard Rosenthol
Yeah, I know you don’t want to hear this. But believe it or not, that’s a feature. In both cases, the reading order panel is absolutely designed to go hand in hand with structure and tagging. So the idea was to… Actually what we thought we were doing, we’d like to think in many cases, we are hope users more easily create that constraint reading plus content. So it tries to do the right thing. You know, it’s software, it’s never perfect, but it combines both things. The content panel, as its name implies, is just about content, which admittedly is also the same thing what the tags panel. The tags panel is just about tags, because you can’t screw with the contents in the tags panel, you can’t screw with the tags from the content panel.

Chad Chelius
But it makes sense when you describe that Leonard, regarding the order panel is the order that it is drawn. And I think that naturally equates to your read order. Your stacking order, which… I think of the stacking order more in the content pane, because that’s where you can drag items and hide them inadvertently, because you sent it behind another object, but based on what you said, it makes sense why when we modify something in the content pane, the reading order is affected. So that really makes a lot of sense. And hopefully that helps our viewers or listeners to kind of wrap their head around it, because I can tell you [that] this is the hardest concept for us to teach. And then the first question we get is why are there two orders. And I’m like, “I don’t know.” But now I actually do.

Leonard Rosenthol
Now you do.

Chad Chelius
Yeah. So very good.

Dax Castro
Awesome. Leonard, this has been an amazing. We can spend… Literally, this is almost an hour now. I’ll have to edit this down. You’re gonna challenge my editing skills to narrow this down. But it’s really great to talk with you. We (Chad and I) started this podcast because we have these conversations between ourselves. And one day I was just like, “You know, more people need to hear this.” And because people start asking questions, we’ll talk about stuff we hear on the Facebook group. The stuff that we’ve talked about today is definitely some of the things that have kind of been looming out there as these kind of unanswered questions. I get out one thing, will you share with us after we get done with this podcast the link to the accessibility declaration statement?

Leonard Rosenthol
Yeah.

Dax Castro
I’d love to post that for our readers so that they could go and download that again on chaxchat.com. Awesome. Well, we are at the end. We do not have time for a “Who’s on Twitter” this week. We just don’t. I just cannot fit it in. But I promise you, next week we will have a “Who’s on Twitter”. Leonard, I don’t know if you know, [but] we do a spot every week where we highlight an accessibility focused Twitter account. We call it “Who’s on Twitter” and we’re just pick somebody and hopefully they’ve got something good to say. We do have one and we’ll mention who they are next week. So stay tuned. Again, Leonard, thank you so much.

Leonard Rosenthol
Thank you guys. It was great.

Dax Castro
We really appreciate you. If there was any one last message you’d like to give to our remediation audience out there, impart your wisdom.

Leonard Rosenthol
You know what, I’ll sort of turn it around rather than imparting wisdom, I will ask for wisdom, which is that, we in the industry and the standards groups, obviously, we’re out there and we’re trying to do the right things for you, but we want to hear from you. I mean, we do have their members of your stuff, AbleDocs is a very active participant, but the more we hear from you, the more you folks get active, whether it’s actually participating in our virtual meetings, whether it’s responding to messages, participating in social media, whatever it is, get a hold of us. You reach to us. Reach out to the association to me directly, whatever works, but tell us what you need, tell us how we’re screwing up or how we’re doing well, because it’s only through that feedback that we know we’re on the right or wrong path.

Dax Castro
And what’s the best way for them to give you that feedback?

Leonard Rosenthol
The best way is go to the PDF Association website, which is pdfa.org, and you’ll see a bunch of links there specific to accessibility. There is even a remediators forum. There is all sorts of great places there and easy ways to contact us.

Chad Chelius
Awesome. Thanks a lot, Leonard.

Leonard Rosenthol
My pleasure.

Chad Chelius
Well, that wraps up our episode [for] today. Once again, we want to thank AbleDocs for being our sponsor. AbleDocs, is the makers of axesWord, axesPDF, as well as document remediation services. So we want to thank them for their support. My name is Chad Chelius,

Dax Castro
and my name is Dax Castro, where each week Chad and I bring you accessibility and unravel accessibility for you.

Chad Chelius
Thanks, guys.

No Comments

You can be the first one to leave a comment.

Leave a Reply

Your email address will not be published. Required fields are marked *