Creating Tools to Build a More Accessible and Disability-Confident Workplace

Presenters: Tzviya Siegman & Christina Volpe, Wiley; Sylvia Izzo Hunter, Inera | An Atypon Company

Joni: All right, welcome to XUG day 2. We’re going to get started with “creating tools to build a more accessible and disability-confident workplace.” We’ve got Sylvia, and Tzviya Siegman, and Christina Volpe doing this presentation.

Let’s see. We learned all about Sylvia yesterday and Christina Volpe is the accessibility solutions lead for Wiley Research. In her current role she creates, drives, and socializes accessibility strategies that influence product and content development teams. Christina has more than a decade experience in digital publishing. Prior to joining Wiley in 2011, Christina held various editorial positions with Hospitality Technology Magazine and Pearson.

And we’ve also got Tzviya Siegman, she’s the information standards principal for Wiley. Tzviya serves as Wiley’s liaison to industry standards groups. She currently works in Wiley’s architecture strategy group, joining her interests in content structure, standards, accessibility and data. Tzviya co-chairs the W3C Publishing Steering Committee, and serves on the W3C advisory board. She has been working on eBook and web accessibility for more than 10 years.

And I’m not sure which one of you is starting first, but let’s take it away.

Tzviya: I think I’m is starting first, thank you. As Joni said, we’re going to talk about creating tools to build a more accessibility and disability confident workplace. We’re going to focus on accessibility maturity models. Christina is running the deck, so here’s our pictures. Feel free to reach out to us, there are email addresses. I think you’ll be getting a copy of the slide deck later on. So let’s talk about what accessibility means. We have this definition from the W3Cs web accessibility initiative, which is a really useful definition for accessibility in general.

“Web accessibility means that websites tools like technologies are designed and developed so that people with disabilities can use them, more specifically, people can perceive, understand, navigate, interact with, and contribute to the web. Web accessibility encompasses all disabilities that affect access to the web, and it also benefits people without disabilities.”

So let’s just take that apart for a minute. This isn’t just accommodating somebody who has a disability like using a wheelchair or somebody who is blind. It means that people have full access to everything that people without disabilities have access to. That means perceiving, understanding, interacting with, that means for a website, being able to build a website, fully navigate a website, and we’ll talk a little bit more about what types of disabilities exist, although that’s not the content of our talk today, and also contributing to the web. And something that’s also really important to keep in mind is that accessibility also benefits people without disabilities.

Some misperceptions that people have, are that making something accessible, especially digitally accessible, might take away from the beauty perhaps, or the complexity of a website or other digital materials, when in fact it often also benefits people without disabilities. So a quick example of that is that if you build a website with responsive design, usually that also benefits people with disabilities as well, because being able to strip away some of the complexity can benefit people with cognitive disabilities. So let’s move on to the next slide.

It’s important to recognize that there’s a reciprocal accessibility relationship. It’s very important for us to talk to our users, our customers about their needs, which often includes accessibility. If we’re not talking to our customers with disabilities, it means we’re leaving out a large percentage of our customers. I don’t remember the exact numbers, but people with reported disabilities in the US, it’s in the 20%, and there’s a significant number of people with disabilities who don’t report those. And disabilities can be situational or temporary, if you have a migraine, that might count as a temporary disability. I get migraines, when I have a migraine, I can’t really function and I might have to continue working, but I need to look at a screen that’s a little bit more simplified. Employees might have disabilities and they require accessible office places, accessible technology, and many other accommodations as well.

And employees need to talk to customers, customers need to talk to employees, employees need to be enabled to have the communication skills to talk to customers, you have a customer who says, calls your customer care team and says, “I’m having trouble with my screen reader,” your employees need to be able to understand, have the language and the tools to communicate with those customers in the same way that they would be able to answer a question about a broken website. So it’s a reciprocal relationship. We need to give our employees and our customers the tools to talk about accessibility just like they would about anything else.

So let’s talk about how to evaluate conformance progress. We talk about conformance or compliance with accessibility, and these are words that I don’t love, but they’re the best words we have right now. The reason I don’t love those roles is that it sounds like we’re just sticking strictly to legislation or strict guidelines. I like to think about compliance as a floor, we want to reach the ceiling or maybe go above the rooftops with accessibility. So accessibility maturity models are something that enable us to do that. There are a lot of maturity models that exist. The level access digital accessibility maturity model is one that’s been around for a long time. The W3C at the Worldwide Web Consortium is using that as well as other models as input to draft their own accessibility maturity model. These are tools that help us evaluate an organization’s compliance with accessibility for things like, what kind of the web content accessibility guidelines, as well as many other things such as the office space compliance. It helps you identify gaps, develop techniques and best practices, as well as inform roadmaps and project plan development. This is a long view of accessibility. They’re organized around dimensions, which are categories that address organizational needs by specific topics. And then you have proof points, which are evidence-based deliverables for outcomes that help you evaluate the maturity of each dimension.

And then there are maturity stages, these are ratings for each proof point that demonstrate the effort of each maturity stage. These are inactive, meaning you’ve done nothing to support the proof point, launch, which is just getting started, and integrated, which means you’ve begun to integrate this, you’ve probably done a lot of work to integrate this. The step between launch to integrated is usually a big jump. And then optimize, which means you’ve done everything you’ve set out to do to reach that accessibility stage.

Christina: Thanks Tzviya. Okay, so I want to spend a few minutes talking about, pardon me, the dimensions in a little more detail from a very high level and pull out examples of the proof points and just some questions that you should begin asking yourself as you are developing a compliance strategy.

So the very first dimension that we have is communications, and this relates to all direct internal and external communications. And this could be anything from emails that employees may be receiving from like corporate parts of your organization, such as HR, all the way to marketing materials, as well as activity on social media. So when you’re within this dimension, not only do those communications need to be accessible, but the preconditions for those communications need to exist. And essentially, these are the requirements that allow you to create those communications. So things to consider here are, do you have brand guidelines? Do you have a visual identity? Do you have requirements around the use of inclusive language? And maybe you even go as far as to having templated documents that you have to write to. So for example, if you’re giving a PowerPoint presentation, do you have corporate templates that your employees should be using as you are creating those presentations?

The next dimension is knowledge and skill. And this applies to the ability to obtain and retain accessibility knowledge, but also you need to be able to apply it to your job. And usually when people think about this, they may only think it applies to full-time employees or maybe contingent workers, but this actually applies to all external agencies that you may be using. So sometimes we hire subject-matter experts or vendors to come in and help fill a gap in skill or knowledge. And regardless of who we’re talking about, that person has to have the same level of accessibility knowledge as the rest of the people on your team. And how you do this is very similar to the communications dimension. So you need to have documentation in place that tells the standards for accessibility, so what standard are you following, how do you apply this to web development, how do you apply it to document development, and then finally, how do you apply it to multimedia development? And that type of documentation works best when it’s coupled with training. And there are lots of ways that you can offer training. Sometimes employees may, who are very motivated, may do the training themselves. Sometimes you may have specific pockets of the company that have a very strong focus in this area, so they’re going to give training for their groups. But the best way to do this is to have a robust, mature training strategy where all of your colleagues are getting access to the same training, and that training is incorporated into all onboarding practices.

The next dimension is support. And support has two parallel tracks, there’s the employee track, and then there’s the customer track. So the employee track is all about ensuring that your employees have not only the ability to do their job but that they’re supported. So this means providing them with accommodation requests. So occasionally you will have employees who may require access to a licensed screen reader or a very specific type of hardware to do their jobs. You need to ensure that they get access to those materials. And when we’re looking at things like HR, hiring managers and even employees’ direct managers, they need to have training, and disability awareness and inclusion to understand how to support their colleagues. On the customer side, this is all about support, both technical and non-technical. So for example, are your customer support agents trained in accessibility? Do they understand how to use assistive technologies? Are they aware of disability etiquette and awareness as they’re having those conversations? Do you have an escalation path in track, does it exist essentially? So if lower-level support agents can’t answer questions, it goes up to the higher levels with people who have more formalized training. And this also includes things like having self-service help support, and topics that have very specific accessibility slant. So what I have on screen right here, for example, is a screenshot of a support article for Knewton Alta, which is a Wiley product. And it’s all about how you support customers with disabilities who use this product.

And then finally, any kind of feedback that your support agents are receiving that should be incorporated into your product development lifecycle or your ICT development lifecycle, which is what they call it in this maturity model, which is the next dimension. And this is all about ensuring accessibility is incorporated into that entire product development process from conception all the way to maintenance and obsolescence. So how do you include accessibility requirements in the design, user experience, development, and even the quality review and release stages? So what I have here as an example, is an accessibility guide that we have at Wiley for web development, and it identifies the- This is the TOC by the way. It identifies the criteria for accessibility, the process for validating. So what is the manual requirement, what automated tools you’re going to use once you’ve done reviewing, what is the definition of done, is there a checklist that you need to follow, are there additional resources where you want to direct people should they have any questions?

The next dimension is personnel and this is all about HR. So the first proof point is about targeted recruiting, and this is ensuring that you are bringing diversity to your workplace so that persons with disabilities are directly involved in bringing their insights and lived experiences to decision-making processes that are occurring within your company. It also applies to how those jobs are being advertised. So do you have an accessible job application platform? Are the actual announcements themselves accessible? Do they include inclusive language? And do you include accessibility as a requirement in any of your job openings?

The next dimension, I promise we’re almost done going through them, there’s only this one and one more left, is procurement. And this applies to the sourcing, negotiation and selection of all goods and services. So whenever you hire an external agency that is either providing a solution or a service, you need to understand what their accessibility requirements are, and this begins at the RFP process. So when you are going through an RFP that should have standardized solicitation language, the templates in the form should be accessible for use. And then once you receive responses back, you need a way of evaluating those partners for risk, because ideally, they should be, if not at the same level as your company, very close to the same level as your company. And once you hire someone, you need to have a burden of proof for accessibility in all of your contract requirements to ensure that they are meeting any KPIs or requirements as a result of that partnership.

And the last dimension that I have here is culture. This is the largest one, and it is probably the hardest one to accomplish because it is about changing perception and advocacy for accessibility. And this covers a very wide range of activities, starting first with having executive sponsorship in place for accessibility, so you need someone who’s going to really advocate for you at a very high level, and you can also ensure that you have budgets tied to this so you can accomplish the things that you want to do. You need to have program managers in place who can implement accessibility policies and procedures to ensure that both your colleagues and external partners are following requirements so that your customers can use your products. And it also includes ensuring that accessibility is a part of your DE&I activities, so is it a part of your, do you have an executive statement about the commitment for accessibility? Do you have employee resource groups that are dedicated to accessibility and so on?

So I’m going to turn the deck back over now to Tzviya and she is going to go through examples of how you can get started with developing a maturity model as well as an example of how this can be worked into product development.

Tzviya: Thanks, Christina. So I know this might seem a little bit overwhelming, but you can begin your journey by picking a model. As we mentioned, there are multiple models and customizing it for use within your company. Any model can be scaled up or down, you can make it bigger or smaller, tweak it to adjust to your needs.

You’ll want to start by focusing on a specific product, department, maybe the whole organization, maybe just one or two of these things. You want to review the model and make sure that you use your company’s lingo, your terminology, departments, processes, some people call production management. There are a lot of different terms that work for different companies.

And of course you need to get buy-in from many different people and think about how the audit can be useful. Business analytics visualizations can work really well, we’ve had a lot of success working with people who are on our product team, they’re the people who actually get the work into the hands of our developers. We’ve had some buy-in from legal, we’ve had a lot of buy-in recently from what we call our people organization, that’s an example of the lingo which used to be called HR. And the push at Wiley for DEI, diversity, equity, and inclusion has really given us a lot of support recently.

So you really, it often starts with a handful of people doing a lot of the work, and then there’s a lot of buy-in. We found that this maturity model really supports that, really helps gain support from the business side of the company, whereas the work on things like individual tickets to solve a problem will get you support from the developers.

And of course, testing it as you go to close any gaps before you were ready to use it, so in a recent discussion about this, we found that there was this kind of, the language that we discovered was that we were using the term digital accessibility in some places and accessibility in other places, and that was a little bit confusing. And while that might sound minor, we realized that there was a big gap in whether we were talking about physical accessibility versus digital accessibility, so that’s the kind of thing you might uncover as you’re working on it.

You’ll want to conduct an audit, interview stakeholders, the actual stakeholders, customers with disabilities, as well as employees with disabilities, as well as people without disabilities. And not everybody’s going to identify whether they have a disability and that’s okay, to understand the current state of your product, department and organization, and of course you have to find the gaps, figure out what isn’t working, what is working, what hasn’t been addressed, and where the biggest need is. So you might find that your biggest need is on your business.com website, you might find that your biggest need is on your site that supports your articles, you might find that your biggest need is actually in the HR organization. And of course, you’ll figure that out as you go.

And then you’ll agree on a focus with your key stakeholders. You’ll figure out what you want your biggest priorities to be in, in the coming years. And this is a long-haul effort, you’ll see in the next slide that we’re talking about really a five-year plan. You’ll start out perhaps with the smallest activities, perhaps with the largest activities, and you’ll want to align with business priorities, and slowly introduce new activities every year so that you can improve the maturity levels and scope them across the dimensions. So let’s look at an example with product strategy. This is where kind of the rubber hits the road.

So in the first year you’ll talk about requirements and documentation. I’m not going to read every line of this, but the idea is that you document in the first year exactly what it is that you’re trying to accomplish. You’ll have testing methodologies, user acceptance testing, you’ll identify a process, you’ll figure out how you’ll accomplish a VPAT, a voluntary product accessibility template. And you’ll agree on terms for contracts, possibly that includes writing an RFP, something along those lines, and of course, the KPIs.

In year two, you’ll start with a development workflow and training, you’ll select the partners for working on these things, you’ll identify a strategy, you’ll actually implement the VPAT, that probably means getting the VPAT accomplished, it’s a pretty lengthy process, and you’ll review what happens in year one. So every year you add on a review process as well.

Year three, you implement the workflow, prioritize the corrections based on the VPAT, the VPAT will give you a lot more work to do. You’ll roll out the training and you’ll review years one and two.

Year four, you start communicating about what you’ve accomplished, because by then you should actually have accomplished a great deal of work. You start the support FAQs so that your support team can help you train and support your colleagues as well as your customers. And you review years one through three, and you begin planning for what needs to go on for years five and forward. So that’s kind of a snapshot of what might happen in your product team. Is that our last slide, Christina?

Christina: It is.

Tzviya: Okay, we didn’t really have a good closer here, but this hopefully will give you a picture of what you can accomplish with an accessibility maturity model. And we’re hoping you have lots of great questions for us.

Sylvia: Okay, thank you. And we’re going to do questions at the end, I believe. So I am going to now attempt to share my screen, which is always an adventure. Okay, slideshow from beginning.

So I’m Sylvia, as you know, and I am speaking to you today from Toronto, which is on the unceded territory of the Haudenosaunee, Wendat and Mississaugas of the Credit, currently home to many Indigenous people from across Turtle Island. And I encourage you to visit native-land.ca to learn about the history of human occupation where you live.

I am the marketing manager at Inera but I also work in disability inclusion advocacy, and it’s in that role that I’m talking to you today. I’m going to give you a really fast tour of a resource currently in progress, which is called the Toolkit for Disability Equity in Scholarly Publishing. Hello. There we go. So what is this toolkit and why? So I hope– Whoops, did I go too–? I went too far. I went too far, I’m sorry. There we go.

So I hope you’re familiar with the C4DISC Toolkits for Equity Project, if you’re not, don’t worry, there’s a link at the end and you’ll get the slides. There’s lots of previous toolkits, and most of them exist primarily as static documents. So we wanted ours to be different in several ways, including a different format and approach. So first of all, it’s online and dynamic. The conversation around disability inclusion and equity is a rapidly evolving one, and for a resource like this to actually help people, it also needs to be capable of rapid evolution. It’s also more broadly targeted than previous toolkits because disability is a thing that can happen to you at any point in your life, as well as being a thing that you might be born with. And as we’ll see on the next slide, “people with disabilities” is a really, really big group, which means that over the course of our lives, pretty much all of us will either have a disability or live or work with someone who does, or both.

So why do we need this toolkit? As I just mentioned, “people with disabilities” is a very, very big category. 15% of the global population lives with some type of disability, which is just, like, a lot. In publishing, because we are making things for people to read or interact with, we naturally focus in on solving accessibility problems for people who are blind, visually impaired, or otherwise print disabled. So one way in which this toolkit project goes beyond a lot of the scholcomms resources that are already available is that it includes resources for all the other types of disability that people can have. And just a quick footnote here, you’ll hear me say both “people with disabilities” and “disabled people,” because within the disability community, people have different opinions about this. And of course, the most important rule is to call people what they specifically want to be called.

So another reason that we need this toolkit is that people with disabilities face significant obstacles to getting and then keeping jobs. So we want to make that easier, both for the employer and the employee because, if you have a disability, hacking your environment is a way of life. And speaking for myself and my toolkit team members, we are valuable colleagues, we want to work and we want to do the best work we can, which doesn’t mean that you’re not valid and valuable if you can’t work or don’t work, but we’re speaking for ourselves.

So this is going to be a real brief history of the toolkit so far, the original idea came from a group called Publishing Enabled, which focuses on advocacy within the publishing industry. We teamed up with C4DISC because the project we wanted to do naturally seemed to fit with their existing toolkit program. And then we took a look at the diversity tax and our existing workloads and said, look, we need a project manager, and C4DISC’s call for volunteers found as a whole team of them, as well as a bigger group of volunteers who form our other committees. C4DISC helps us coordinate and support those volunteers. PubPub is going to host the toolkit when it exists, and has offered to work with us on hosting any new types of content that we come up with, which is awesome. Because remember, this isn’t a journal, it’s a community resource hub.

So we know that previous toolkit teams had benefited from working at the Triangle Scholarly Communications Institute, so we applied to send a team there to work on this project. And if you’re not familiar with Triangle SCI, it is an annual retreat-style event that if your proposal is selected, they pay for your whole team to get together and work on a project for a week, which is basically the dream. What we didn’t fully get before we got there was how valuable it would be, not just to have this time to work together ourselves, but also to use the rest of the participants as a focus group for questions like, “okay, if you had a resource like this, how would you use it and what would you want it to include?”

There is a size limit on teams, we could only bring six people, so we couldn’t bring the entire steering committee, but those of us who did go represented a university library, three publishers, a standards organization, and a technology vendor. And don’t worry, we just took our masks off briefly for that photo, they had great COVID protocols at this event, which was very, very nice and inclusive.

So here’s our mission and goals, we don’t have time to unpack all this, but it is on the slide so that you can read it later. And also as a reminder to me and to everyone that there’s a whole big team of volunteers working on and thinking about this project, not just the little team that went to Triangle SCI. So we came to the Institute with some pretty well developed ideas, developed by the whole team about what the toolkit should look like. We wanted it to be a true toolkit, in other words, containing lots of different tools that different people could pick up and use. We wanted it to be dynamic with clear and easy mechanisms for keeping resources up to date, and with a variety of different entry points and pathways into and through the resources. We wanted it to be for our industry, so including resources that address specific issues in scholcomms. And we wanted to address people with disabilities who work in scholcomms or who would like to, people who hire and manage those people, and who run organizations in this space, and finally, anyone who wants to be a good ally to their disabled coworker or relative or friend.

Oops. Oh dear. So here’s the process that we went through over our week at the Institute. We first figured out what should go in the toolkit, what topics, what kinds of resources, not what resources specifically, because that’s a later stage and a different team. Then identify the minimum viable product, redefine the product workflow because we’re unlikely to get a second chance to do this much concentrated work, so we needed a strong process to move forward. And finally, figure out what we’re going to do next and when and in what order.

So this screenshot of a Miro board is a more organized and systematic version of where we started, which was with Post-it notes on flip chart paper. Our website will have a welcome page called You Are Here that helps visitors get oriented and offers multiple pathways into the content. And what’s the content? Ultimately it’s resources for every stage of the employee life cycle, from writing a job posting, to onboarding, to mentoring and promotion. We’ll address the broad job categories in our industry and different organization types like publishers, libraries, and vendors. And we’ll address both organizational and personal responsibilities. The types of resources will include some basics like, types of disability and a glossary of terms, as well as guidance on inclusive language, best practices and accessibility-related standards. We’ll link to other organizations that are related and helpful, and we very much want the site to include some kind of community functionality, but we’re still working out exactly what that might look like. Another future section we’re really excited about has the working title of “lived experience.” And here we’ll host, for example, scripts for navigating difficult conversations around disability and workplace adjustments, personal stories and case studies, and opinion pieces and that kind of stuff. We’ll have an FAQ with sections for the different stakeholder groups that I mentioned. And we want it to be welcoming and relevant and useful to a wide range of people, and have different starting points and ways in. And of course we’ve got an about section with the who, what, and why, thanks to our sponsors, the limitations, accessibility and privacy policies, how we’ll review new resources, and how to contact us if you have a question or a complaint or a suggestion.

So if you were contemplating all of that and thinking, “Sylvia, that’s a lot,” welcome to the Minimum Viable Product. We asked ourselves what would be absolutely essential for the toolkit to be useful to people and pared it down to this set of materials that we’ll be working to have ready for launch. So from seven sections we’re down to four, and within those four we’ve identified the key elements that we’ll now be working towards finding, curating, and or creating over the next few months. So You Are Here, will include a welcome message that helps visitors find what they’re looking for and it’ll explain the purpose and feature a video message from the team. In the about section, we’ve got just the basics, background, mission, vision and goals, who we are, what the toolkit is and is not, and a thanks for our sponsors. The resources library in the MVP will focus on basics or disability 101, language guidance standards, and suggested reading. There are tons and tons of resources out there, that’s not the problem. The point is to find and provide those that are most relevant and useful to scholarly communications people. And finally, we’ll start our FAQ with the section that’s aimed at people with disabilities who are working or looking for a job in schol comms.

Oh dear, that’s not what I wanted. Here we go.

So here’s where we’ve landed at this point in the project, instead of the research committee feeding the framework committee, which was our original idea, we’ve put together the framework and we’re asking the research and writing committees to find resources that meet the needs we’ve identified for the MVP. That is research will find existing resources and then writing will fill in the gaps. And meanwhile, the website committee will be working with PubPub, and the outreach committee which was formerly known as marketing and has been renamed to acknowledge its community building role we’ll be planning how to spread the word.

This is our timeline for the next year and a half or so, and shout-out to my colleague Erin Osborne-Martin at Wiley for this very lovely graphic. The first milestone is a launch of the MVP site at or before the SSP annual meeting, so cross your fingers that we’ll get our session proposal accepted and that it will be ready to go. By September of next year we’re aiming to have worked out what the toolkit will be going forward, how it will be maintained and updated, what form or forms our community will take and so on. And then in November, we’re hoping to take our roadshow to the Charleston Conference and also launch our community. And this is so important to us because while big organizations like Wiley are likely to have resources for employees with disabilities, someone at a smaller organization may feel really alone and unsupported. And we want the toolkit to be a resource for people in that situation and to provide community support as well as just information. So that’s a big, big task and we’re still figuring out how best to do it. And the plan is that whatever that future state is, that gets figured out over the next year, it’s going to go into effect in 2024.

And I promise I’m almost done. So here are some resources for more information, and everything here is linked, the notes on previous slides are linked. And in about the next 10 minutes, these slides are going to be posted on the website and you can take a look at the ones that interest you. And now let’s do questions.

Joni: All right well that was fantastic. Thank you so much, Sylvia. In the chat. In case you missed it in the chat, Jennifer Kemp from Crossref dropped a link to the accessibility proposal for Crossref DOI links, which is an open discussion at the moment. They’re looking for feedback on that if anybody wants to go there and share that.

Tzviya: Does anybody want me to get the ball rolling with questions, I’d like to–

Sylvia: Yeah, go ahead.

Tzviya: People often ask what the cost of implementing accessibility is, I will say that there’s no fixed answer to that. If you have no budget, what I usually say is try to squeeze it in wherever you can. It’s very easy, quote unquote, to get started in small ways, for example, if you have a website, you can start to take a look at the way that your HTML is encoded, probably your website has a lot of JavaScript there. But you can start with simple things like, using an automated tool like Axe, Lighthouse, one of those. And that will get you only about 20 to 30% of the accessibility checks that are required but it’ll get you started. You can also run keyboard navigation, that’s not going to develop your maturity model in any way, but it gets you started, it gets people thinking in the right direction of accessibility. It might also open your eyes to just how much work there is to do. It’s a very good way to get started. The actual cost of accessibility, as you build up there are more and more costs, this is actually why something like a VPAT, it depends which vendor you go with, Christina can speak more about the exact costs, but it really varies with vendor and with how many pages you’re actually looking at. But that’s why it’s important to get the buy-in from leadership, and that’s why something like a maturity model can really be helpful. It shows how much it affects the entire organization, it’s not just that you’re looking at one website or platform.

Sylvia: And I think it would be interesting also to talk about the cost of not making things accessible because people can’t access it if it’s not accessible, and then- Bruce has a question.

Bruce: Sure, actually I have a few comments. The first is, one of the things that I think people have been learning in recent years, excuse me, is the value of creating something with accessibility in mind, it turns out to be universally usable. So I think a lot of us, particularly those of us who are older like myself, still think of closed captioning as something that’s an accessibility accommodation. I’ve read a number of articles recently that point out that particularly with people under 35, they turn on closed captioning whenever they’re watching something on streaming, why do they do that? Because it’s easier for them to multitask. So here’s something that was designed originally as an accessibility accommodation – for those of you who were at XUG three years ago, you may know the first ever closed caption program was The French Chef from Julia Child in the 1960s. But it’s now become mainstream and very usable on a mainstream basis. So it’s something to keep in mind that while you may look at it as a cost for accessibility, it can become a very, very important functional area for product.

The other thing, excuse me, that I’ve discovered recently is if you have a disability, it can be very tough if a company hasn’t done the right thing to reach someone who can help you out. So I discovered a couple of days ago when I had a power outage that I had to go onto one Zoom call using my iPad rather than my laptop. My laptop I can control with a mouse, my iPad, I’m at this point entirely controlling it with voice control. Well, it turns out there are certain buttons you have to hit in the process of joining a Zoom meeting, like, do I want to dial in or use WIFI for my audio. It’s unconditional, you can’t just, you have to tap on those, the moment you start the Zoom meeting voice control is turned off. So I had to go screaming for a caregiver to essentially tap two buttons to get me into the Zoom meeting. I’ve been on the Zoom website today looking around to see if there’s any way I can reach someone focused on accessibility to see if I can get help, and there’s no contact directly for accessibility needs. There are companies that do have that.

One last comment, I have a friend who lost her voice to ALS a while ago and she was all set up so that she could use a synthesized voice, she’d type things on her computer, they’d be read out. I can’t tell you the number of times she would call a company looking for technical support, they’d hear the computerized voice and instantly hang up because they thought she was a robot or some sort of a spam call. So building these resources to help support people with disabilities is a critically important part of the job.

Joni: Thank you, that is really-

Tzviya: I see a question about EPUB. Let me just, are there accessibility conformance preferences for delivery of long-form text in HTML browser or EPUB? We’ve heard about EPUB 3 being one of the best formats for delivering accessible content for years but is it universally preferable to delivering the same content in an accessibly designed website? So I’ve been involved in EPUB for more than 10 years, the answer is, it depends. HTML can be very accessible. It depends what kind of content you’re delivering. EPUB was never meant to replace HTML. For a scholarly article, I would say stick with your HTML, it’s fantastic. There are people who advocate for delivering birthday cards as EPUB, I’m not one of those people, you can do it in HTML. HTML is fantastic, make use of it, but make sure it’s accessible HTML. I have been involved in ARIA – accessible, rich internet applications, that is the tool that helps make HTML fully accessible. I could talk to you for days about how to make HTML accessible, but it doesn’t actually take days because it’s not actually that complicated. Make sure you’re making your HTML accessible. The same thing with EPUB, not every EPUB is accessible, you have to do it correctly.

If you’d like me to talk to you more about how to make EPUB and HTML accessible, I don’t want to monopolize the conversation, but please feel free to reach out to me. And there are many other people on this call who can talk to you about that as well. I want to address one of the points that Bruce made also, and then I’ll let other people talk. There are so many ways that tools that were originally made for accessibility have been shifted into the mainstream marketplace. Curb cuts made for wheelchairs, we’ve all used suitcases, strollers, whatever. Siri, Speech Recognition Institute, we all make use of that sort of thing on a daily basis also. The point is that, if you make your content platforms accessible, you will be improving the user experience for everybody. And now I’ll let Christina talk.

Christina: Sure, I’m going to talk about VPATs ’cause I saw Bill mentioned about VPATs. For anyone who doesn’t know what they are, VPAT stands for Voluntary Product Accessibility Template. It’s a form of accessibility conformance report that will tell your customers how accessible or not your product is. Essentially it’s an audit against your product, against the web content accessibility guides, it’s a lot of acronyms, there are 52 of them if you’re following WCAG 2.1 A/AA. You always want to really hire a third party because it’s very time consuming to audit because the audits are going to take into account a variety of methodologies. So you’re testing with an automated tool, yes, but you’re going to be testing with assistive technologies, you’re going to be testing with some manual processes like keyboard navigation, zoom. If you have people internally who are not as familiar or adapted to those types of testings, you are absolutely going to get inaccurate reporting. So you want to hire someone who is specialized in this and there are lots of vendors out there who do this by the way. And the cost will really vary, it’s going to depend on how many screens we’re talking about, and the complexity of your product. I’ve seen some VPATs quoted as low as 7,000 to some as high as like 20, 30,000. It’s really going to depend on what you are reviewing, but yes, absolutely use a third party and go through-

Tzviya: Oh, oh, Christina froze.

Christina: But also from the approach that company takes.

Tzviya: Christina, you dropped for a second. You said go through-

Christina: Oh sorry, go through the rigor. You have to put them through an RFP basically, find the companies that work best for you, not just from a budget standpoint, but also from like a compatibility standpoint, someone who has an approach that you’re happy with.

Sylvia: So I see that Joni has a question for attendees, which nobody has answered yet, which is, “If you publish audio or video clips are you providing transcripts? And if yes, how and where are you doing that? And if no, are you planning to add them in the future?” So I will just say that when we post the videos for this meeting for attendees or registrants to review or see again or view for the first time, we will have transcripts as well as closed captions. And they will be linked from the video description on Vimeo, so that’s how we do it.

Liz: Do you want to talk a little, just to give a quick snapshot of the process of how we do that?

Sylvia: Yeah, so we use a service called Rev. And the way Rev works is you upload, either you link, you post the URL for your video or you upload the video depending where the video currently lives. And either an AI or a human being, depending on how much you want to pay in money versus time goes through the video and creates basically a transcript. But then you say like, do I want, when you put in your order, you say, do I want captions, do I want transcripts, do I want both captions and transcripts, do I want– which caption format do I want? Because the front-end work is the same, but then you can get it exported in different ways. And then there’s an– this online platform allows you to watch the video back, watch the captions scroll by, make any corrections that are needed, and obviously the more a human is involved in the process, the less you need to correct. You can also provide a glossary, like if you have words like Inera and Atypon and eXtyles that will otherwise be incomprehensible to this trained but not jargon-trained person who’s doing the captioning. And then at the end you can download the captions as an SRT or VTT file or both and upload those to Vimeo. You can use a program like Adobe Premiere to burn them into your video if you want to do that. And you can also download a transcript, which you can then post on your website in HTML. And the cost is not zero but it’s really reasonable, for the amount of accessibility, additional accessibility that this provides.

Tzviya: I’ll also add nobody’s mentioned legislation in this discussion. I try to talk more about the carrots than the sticks, but in the United States we have the Americans with Disabilities Act, and section 508, the Rehabilitation Act. We won’t get into exactly what is required and what isn’t required, but I’ll just say that places of public accommodation require accessibility, whatever place of public accommodation means. In Europe, in the EU, we have the European Accessibility Act, which beginning in 2025 will allow for fines. Exactly what those fines will be is not yet set in any country, except Italy’s beginning to come to terms with that. But that’s part of why a lot of us are starting to take accessibility a little bit more seriously now. So if you sell anything in Europe, the European Accessibility Act is absolutely explicit that it includes anything digital, it expressly says eBooks, eBook readers, websites and so on, learning management systems. And then there’s also legislation all around the world, that I’m not going to go into detail about, but Ontario in particular, has some pretty strict legislation.

Joni: We have a few more minutes if there’s any more questions? And full disclosure, so a couple of us here are on the JATS4R accessibility committee, which is part of why I am trying to get everybody to talk to me about transcripts, which is the thing that I want to talk about all the time.

Christina: So I want to ask you a question then, so we’re talking about transcripts, are you doing anything with audio descriptions for video? So if there’s something happening on the screen, but it’s not acknowledged by the narrator, are you hiring people to write the audio descriptions for people who are blind? Who can’t see what’s happening on the screen and need that description of what’s actually happening but it’s not captured in the narration.

Sylvia: Speaking for us, currently no. And that’s kind of the next step I think.

Bruce: I think some of the major studios are doing that. There was an article I read in the Wall Street Journal six months or so ago about Netflix, which is doing that for a lot of their programming now.

Joni: I’ve got a question from Bill.

Bill K.: I guess I’ve got a comment on that, that if you don’t provide audio descriptions and your content is used in education and university, etc, very likely there’s a person in the disability services offices at that university that has to do that. It’s an enormous amount of work, it’s enormously costly, and the university is obligated to provide that accessibility for that resource to that student. So it has to be done. If you don’t do it, somebody else is going to do it, and oftentimes there are many universities remediating the same resource at the same time, it’s a crazy situation.

Sylvia: Yeah, yeah.

Bill K.: So just do it.

Tzviya: Well “just do it” is very easy to say.

Bill K.: Yeah, I know, I know. And audio description is the hardest part, I think.

Tzviya: Well, there are many-

Bill K.: Especially to do it well, to do it well it’s a real art. But just be aware that downstream of you getting your content into the library or into the university classroom, there very likely is still somebody else that’s having to deal with something that’s not sufficiently accessible. They do all kinds of work, they do image descriptions, they fix tables, they do all kinds of crazy stuff, and it’s usually very labor-intensive work. So a lot of people just aren’t aware that that’s happening.

Joni: I’ve got a question from Kevin O’Donovan.

Tzviya: Okay, “how about reducing the length of documents in simplifying sentence and paragraph structure? Is that considered a part of accessibility, readability as accessibility?” I can begin to address that. So, the documentation for cognitive accessibility has just begun to be a part of WCAG, the web content accessibility guidelines, which actually are going to be called the W3C accessibility guidelines because they want to expand it beyond just the web. It’s begun to be documented and it’s not necessarily that you need to do reduce document length or reduce complexity for something like a lot of us are in scholarly publishing, if you’re writing an article about quantitative chemistry, it’s not necessarily possible to reduce the complexity. What is recommended though is to provide a simplified language alternative. And what remains to be seen is how this will play out in areas that are inherently complex. If you give me a minute, if we’re still on, I can find a few samples in the drafts of, I think it WCAG 2.3 where they show how this is done, ’cause standards language is also generally very complicated and they show simplified version of the actual documentation. But like I said, when we’re working in the area of scholarly publications, it’s not necessarily a topic that can be simplified.

Jo: I have a comment, that is both comment and question, and I’ve been trying to figure out how to articulate it. So I’ve found that a lot of people, like in the lower millennial and Gen Z generations, that there’s a lot of undiagnosed, what I’m going to call a mental disability. It’s not actively visible like ADHD or anxiety, but very much does – migraines! – does affect day to day work. And I know that for me, a lot of the, what started out as things for just general accessibility, like closed captions and transcripts has been instrumental in my development and ability to work well. And I’m just curious how these toolkits apply to those types of disabilities and accessibility as well. Because some disabilities are very obvious, like hearing impairment or eyeglasses, but some of those less visible accessibility needs are often not disclosed in a workplace. And so how do you address those if they’re not actively disclosed? I guess is the question. Does that make any sense at all?

Sylvia: I think we have to assume that they exist even when individual people don’t disclose them, that’s kind of step one. But I’m going to let Christina and Tzviya speak to that too.

Christina: I mean, so the model does take into- It’s supposed to take into account the full range of disabilities that exist, not just the visible ones, but also the invisible ones. So that’s why it’s good to look at a model against something else like the W3C requirements because it does really try to marry the two together. And it’s going to come down to again, how you are training people to work with their colleagues, how you are training people to create materials that are going outside of your company. So it’s not so much as, like a separation, it’s looking at it as a way of, including the full range.

Tzviya: I think also that something, adding on to what Christina said, once you’re making materials accessible or accessible for everyone, you don’t have to know that you have a disability to benefit from accessible materials, this goes back to what Bruce said. I don’t have to turn on an accessibility button to benefit from something. This is why you might have seen some websites that have like a little button that gives you access to accessible materials, this is why many people in the community of people who develop accessibility tools are so vehemently opposed to that. I don’t need to say, I need to turn on the tools for people with vision disorders to access that material because I just should benefit from it naturally, because it’s built into the website.

And I could talk more about that also, but somebody keeps mentioning long COVID, I developed an autoimmune disorder after I had COVID in November 2020. This is my TED talk, right? I know all about long COVID. I didn’t know what it was for a long time, I knew that I had all sorts of problems with COVID – that stemmed from COVID – I had fatigue, I had this weird chest pain, I had all sorts of problems. Yes, it’s a huge problem but this is just like all sorts of other disabilities. People don’t necessarily know what it is at the outset, it creates cognitive dysfunction, it creates all sorts of problems, but people will still benefit from this model. We’re building the same resources into our technologies and into our physical workplaces. And it’s not that we have to do something special for somebody with long COVID, versus dyslexia versus blindness, a lot of the tooling overlaps. And so this model embraces everybody, people with vis- visible disabilities, invisible disabilities, temporary disabilities, situational disabilities. Situational might mean I broke a wrist, that’s temporary. I’m in a noisy room and I can’t hear what’s going on, so I turn on captioning. I’m nursing a baby, I can’t use my hands, that sort of thing. So disability doesn’t necessarily mean I have an impairment, it might just mean I can’t use a certain part of my body right now. So I hope that helps answer some of your questions. I do encourage you to look at some of the links and read up on these models. And again, feel free to reach out to me, Christina, Sylvia, hopefully we can help you move forward with these.

Joni: We are out of time, but this was a great conversation. Thank you all so much.

eXtyles in the Cloud

Presenter: Bill Fox, Inera | An Atypon Company

Liz: I think we’re moving on to our next talk before the break, which is going to introduce Bill Fox who heads up our development team. And he’s going to talk a little bit about what we’ve been working on and what’s coming up for cloud technology. So just to orient us a little bit, we have currently one commercially available cloud product which is Edifix, the sister product to eXtyles, which is the reference processing web service. But we’re working on other things and Bill’s gonna give you an overview.

Bill: Sounds like we’re missing the sound on this.

Jo: Thank you, Bill. I forgot to optimize for video clip. Well, I will restart the video.

Bill: Hi, welcome to eXtyles User Group. My name is Bill Fox, Director of Software Development. Today I’d like to share with you development and strategy in our eXtyles cloud progress. In past XUG meetings, I’ve talked about Edifix. Well, we’ve been in the cloud with eXtyles technology since 2014. Edifix is our reference parsing, styling, linking technology using the same background tools that we employ in eXtyles desktop. We’ve also been in the cloud with a prototype service for our Arc technology which is automatic paragraph recognition for journal articles.

So what’s new? Today, we’re showing off our prototype of Edifix with DOCX. In the past, Edifix required that you cut and paste your references into a text box and then ask it to process. Now you can drag your journal article document directly into Edifix and it will find those references and process them accordingly. Let me demonstrate. This is an Edifix prototype service where I’m going to select an editorial style and choose a Word file that I wish to be processed.

This will take approximately two minutes to process, so I’m going to go ahead and jump forward to our finished jobs list and show a previous job that I ran. Here you can see the references that have been captured from that document, corrected, styled, linked, et cetera, with other metadata information for you. So Edifix with DOCX, how’s that married to our eXtyles cloud strategy? Well, what’s in there? We’re using a new DOC/DOCX translation capability that is crucial to being able to process your journal documents in the cloud. We’re using the same citation parsing, styling, linking technology that have always been in eXtyles desktop and in Edifix. But we’ve also prototyped our new architecture for delivering most of eXtyles’ capabilities in a cloud service. We’ll still offer that Edifix API 2.0 service with this new DOCX capability.

Importantly, we’re not using Microsoft Word and we’re not using Microsoft Windows to scale this process in the cloud and to build out these components that cover most of the eXtyles features. So we’ve done a prototype and we’re ready to move forward. So one of our top development priorities in 2023 is to bring the batch processing capability into this cloud architecture. Robin Dunford will share in a few minutes information and a description of our eXtyles SI offering.

We’re also going to offer or port over our Arc journal paragraph recognition into this capability. Finally, we need to bring over a full export capability for eXtyles. So input, process, and output gives us a full flow to bring your documents into the cloud and return a fully styled journal article. Many components will follow this initial development, such as document cleanup, auto-redact, and more. So in conclusion, we’re going to offer Edifix in production with this DOCX capability very soon. We’re going to be working on those additional features that I just mentioned to build out the infrastructure to allow us to bring most of the eXtyles capabilities to the cloud. Ultimately, we will have a full cloud service. Thanks for listening. I hope you have questions. We’ll take them now.

Liz: We have a question. Thank you, Bill. From Charles O’Connor. “Will Edifix return DOCX as well?” Robin said, “I knew someone would ask that.”

Bill: Apparently, I was muted. At this time, no. It will not deliver DOCX. That’s something we’ve discussed. We can’t promise that at this time.

Liz: So the question then from Kevin at ICF. “A full cloud service for just journal articles or for books as well?”

Bill: Part of the batch processing is that we’ll take in DOCX and we’ll deliver XML out of the process. So there’s not a user interaction. Key to this is the ability to recognize components of the article or the document piece. Therefore, right now, Arc technology works on journal articles. Should you need just metadata or references, it is conceivable that our batch processing could handle a book component, but we cannot, at this time, recognize all the components of book content.

Liz: Yeah, it’s also conceivable that if you have a fully structured manuscript in Word already that this could, correct me if I’m wrong, that this could then work on non-journal content. But for the full automation of Arc that’s taking an unstructured manuscript and doing all of the intelligence that applies paragraph style recognition, that’s strictly for journal content at this point. I have a question. The Edifix being able to take in Word documents and recognize the reference list, I think that it seems to me that that’s gonna be most beneficial for as you say sort of batch processing and for use with the API, but is that also going to be part of the UI so that Edifix subscribers will be able to upload Word documents as well as organizations who are using it integrated into other systems?

Bill: Well, as the prototype demo shows, we’ve already got it in the UI, and in fact we’re going to be adding it to our API access. I would actually put it back to you as like how do we want to offer this capability?

Liz: Yeah, I think we’ll explore that, but I think it would be very cool to have it in both.

Bill: Absolutely. Obviously, we don’t want it to be out there on part of our trial subscriptions on Edifix, but if there are particular customers, we have heavy UI users within Edifix and they may find incredible value out of being able to put Word documents in and get references out.

Bruce: Bill, if I can come back to Charles’ question for a second. Under the new 2.0 API, can Edifix return formats other than JATS such as the HTML format that we have in the UI?

Bill: Absolutely, Edifix has the ability to export in some 15 or so formats. Maybe I’m over exaggerating, 12 or so formats. One of those is DOCX but it’s all in references in DOCX. I’m not sure I’m understanding the question. Maybe if Charles could qualify his question or I was thinking more along the entire document. Am I getting this right, Bruce?

Bruce: I think what Charles is looking for is, if you use the API with new feature to put in a DOCX file, can you get something back that you can open in Microsoft Word? ‘Cause you can’t open JATS in Microsoft Word but HTML can be opened in Word.

Bill: Right, so the short answer to that is yes. Once we deploy this into Edifix, you can put in a DOCX and you can get out DOCX and a variety of other document formats. Part of my thinking was that with eXtyles in the future, we can’t just deliver back an entire DOCX for a full journal article.

Bruce: Right. Now Charles, jump in and turn on your microphone if I’m missing something. I believe the context of your question was strictly with the reference list. Can you get back something you can load into Microsoft Word? Is that correct?

Charles: Yeah, I mean, of course the idea would be to send it to edit a full journal article to Edifix and get back a full journal article with the references styled. But if it said you sent the full journal article and got something back that you could then get into Word, that would also work.

Bruce: Okay, interesting point. We can certainly think about that. Thank you, Charles.

Bill: So I’ll conclude on that. If you put in DOCX to Edifix, which is all about references and not other parts of the document and you want DOCX out, that will be there very soon for references.

Liz: Great. Any other questions? Thank you. We’re doing great on time this morning, which always pleases me.

eXtyles SI—All the Power, None of the Clicks

Presenter: Robin Dunford, Inera | An Atypon Company

Christine: Hi everybody, I’m Christine Benson and I’m the editorial quality assurance specialist at an Inera. You’ve probably gotten some emails from me for routine updates. It is my pleasure today to be introducing Robin Dunford, who is one of our senior solutions architects. Robin’s presentation today is titled, “eXtyles SI: All the Power and None of the Clicks”. And I’m sure you’re all really excited to hear more about eXtyles SI. So I’m going to go ahead and hand it over to Robin, but during the presentation I’m going to be moderating the chat for questions and we will have some time at the end for a Q&A. So, go ahead, Robin.

Robin: Oh, thanks, Christine. I feel under real pressure now cause everybody’s stuck so well to time so far. But yeah, let’s see how it goes. And I’m, all this chat about not being very well, a few people on the Atypon stand or the Wiley stand in Frankfurt have reported back with Covid since they came back. I am feeling a little bit croaky today, but I have tested negative. But yeah, I’m hoping I don’t start to-

Liz: We’ll keep our fingers crossed for you.

Robin: Yeah, thanks. I’ll struggle through. So SI, what is eXtyles SI? SI stands, as I understand it, predates my time with Inera, stands for server implementation, though I should say I run eXtyles SI on a MacBook. So, do most of the Inera team. So server is a kind of a euphemism. Don’t think of it as being some massive box on a rack somewhere. You can run eXtyles SI anywhere you can run regular eXtyles. But eXtyles SI allows eXtyles processes to be run automatically without the need for a user to be present. So that’s essentially what we’re talking about. We’re talking about automated processing. And steps to be run are specified in an XML manifest that complies to a DTD of our own creation. And essentially if it can be done by pressing a button on the eXtyles ribbon, it can be done by eXtyles SI, except for paragraph styling, and I’ll talk a bit more about that later on.

So if I’m not pressing the buttons, what is? As I say, eXtyles SI uses an XML manifest to control the entire process. Looks something, here’s a little snippet of a few angle brackets. So we’ve got a few steps here. We’re opening a log file, we’re then opening a Word document and we’ll notice, I’ll talk about that later. We’re not opening one particular Word document here, and then we’re logging a message to say we’re doing something, we’re starting processing a file. So instead, as you’ll see there, instead of the dialogue boxes that you’ll be used to from desktop eXtyles that come up and tell you eXtyles has done things and this has happened and we found this, or this or that, eXtyles SI writes messages to a log file so you can see what steps have been run and see what the output, or the outcome of the different steps has been.

So here’s a typical eXtyles SI log file. Just pick out a few highlights here for it. So we’ve got a build number there in this particular log. So this tells you which build of eXtyles was used to run this manifest, can be useful. We’ve got a timestamp there so we can see when a particular step happened and that can be useful if you want to see how long did things take, for example. And we do a lot of that with our Arc, work to develop eXtyles Arc, looking at how long different steps take and things like that. We’ve got some error classes here. So these are, there are different classes for different types of message. As a rough rule of thumb. The lower the class number is the bigger the problem is. And so these high numbers are usually just information, they’re just telling you things have happened. But you could have a process run that looks out for logs that have particular low class numbers and that tells you that something went wrong with that file. And then you’ve got an individual error number. And again, those are very specific messages that are called out at different places in the eXtyles code.

Then we might have the result of a process. So in this case we calculated a timeout value and I’ll talk a bit more about timeouts later on, but in this case we decided that the timeout value for this file was 2,037 seconds. And we calculated that from these various things, which the message tells us here, how many words were in the file, how many table cells, how many MathType objects, how many Office Equation Builder objects, how many references do we think there are in the file, so on. How big are the images that are left, how long has already passed, three seconds. So we got some information there and we got another process result here. So we’ve been removing style aliases in this particular manifest and it’s telling us which aliases were removed from the file. So there’s a lot of information there in the log and that’s designed to help you debug things if something’s gone wrong, but also to monitor how eXtyles SI performs.

So how does eXtyles SI know which Word file it should be using? In your manifest, you can either specify a location and a file name of a specific file. So you can say run these steps on this particular file. So we have a bunch of test suites that we run every night and some of those run particular steps on particular files every night, so we can have a manifest that spells out exactly which file to run or as I showed you in that little snippet before, you can use wildcards, like the little asterisk, and the question mark, and so on, to pick up for example, all words that match a particular pattern, all Word files that match a particular pattern or to pick up any Word file. So I had a *.doc? So that’ll run any .doc file that appears in that location or any .DOCX file. And you could set up a watched folder, so essentially you’ve got a script that’s watching a folder and as soon as a Word file gets saved into there, it’ll kick off a manifest and the manifest will pick up that file and run with it.

So what are the end products you can get out from eXtyles SI? I mean similar to regular desktop eXtyles, so SI will automatically save the Word file when it finishes. You don’t have to worry about telling SI that it needs to save the file once the manifest’s done. Baselines are created automatically anywhere they would be in desktop eXtyles. So typically before Auto-Redact for example, SI will save a baseline. You can use the create a baseline function, you can add that into your manifest to say I need to save a baseline at this point. You can put that in your manifest. One thing that the desktop SI doesn’t offer, because it’s pretty simple to hit F12, or save as, or whatever, from the Word menu, is that you can save a copy of the Word file. So if you want an intermediate copy at a particular stage so that you can go back and look at what the Word file looked like at a certain point rather than using a baseline. So you can save it into a different directory. For example, specify where you want it to go to, and you can stick a prefix on the front or a suffix on the end to, on the end of the working Word file. So you could have, you know, file name dash before Auto-Redact, after Auto-Redact, whatever. So, you can just put a prefix on or suffix on there.

And of course as with desktop eXtyles, you can run one or multiple XML export steps anywhere you like in the manifest. Doesn’t have to be at the end, though typically that’s where you would expect to run them. So another thing that typically you are used to doing by hand, if you’re using desktop eXtyles, is adding document information to the file from the document information dialogue.

So how do you do that in SI if there’s nobody sitting at the computer? So with desktop eXtyles you can either add metadata into the document information dialogue by hand, typing it in, copying and pasting, so on. Or if you have metadata browse set up, you can browse to a metadata XML file that contains information that you want to store as document metadata. So SI has two analogous options. You can hardwire information into the manifest. So for example, if a particular manifest is only used with a particular journal, you could hardwire in the journal name into the manifest and you don’t know it’s just there and it gets picked up that way. Or similarly to how you can do with metadata browse from the desktop version of eXtyles, you can pull in metadata from a metadata XML file, or you can use a combination of the two. So again, you’ve got a manifest that only runs on one particular journal, you can set the journal in the manifest just by, you know, hardwiring it into the manifest. But then you can pull in other things like author names, and ORCID IDs, and publication dates, and so on, from the metadata XML file.

So I just want to talk about a few possible SI workflows that you could set up. And I guess this is what you might call the sort of traditional SI workflow. So I’ve got three manifests here and then I’ve got two steps where a human being is still involved. So essentially paragraph styling, manifest that runs through activation and cleanup and pulling in the document information, which could come from an XML file. Then I’ve got a manual step here, which is paragraph styling. So the manifest has run, finished, that part of the SI process is finished, and then you’ve got, you might save the output to a particular location that someone else then picks up and then does the paragraph styling by hand. Then they could pass that onto another manifest, which then runs Auto-Redact, runs all your advanced processes, your reference processing, your PubMed reference checking, and so on, your citation matching, your author processing. And then you could copyedit, so you, you pass that file on and then somebody is coming in and copyediting, they’re reviewing eXtyles’s comments, they’re checking, you know, all the warnings that eXtyles put in about references and so on. And then they could throw it over the fence again to a third manifest that does the export or might do multiple exports, for example, you might be doing Crossref deposit, so it could run that export and it could run an export to give you JATS XML as a second step.

So this is a second example and this is a more automated process. So here we’ve got a manifest similar to the first one where we’re doing the activation step and the document information gathering and we’re running cleanup. Then we’ve got somebody still by hand doing the paragraph styling. But then essentially we’re telling SI to do everything else. Run Auto-Redact, run all the advanced processes, run citation matching, run author processing, output an XML file, and then you review that XML. So that might be in some kind of WYSIWYG environment or some, yeah, some kind of XML editor. So in this case we’ve only got two manifests and only one manual step with the Word file. And then my third example here, is essentially a totally automated workflow. So in this case we’ve got a manifest that does the lot, soup to nuts. So it runs activation, it runs cleanup, it runs eXtyles Arc to do automated paragraph styling, and then it passes that file on to Auto-Redact, onto reference processing, onto citation matching, author processing, and gives you an XML file which you then review. So you’re still going to have those same comments that have come in from PubMed, Crossref, you might still have your citation matching warnings, you might still have even, you know, Arc might not know what exactly what to style some things as, you might have to review some of that. So this is a fully automated build showing that kind of cloud solution where you’re essentially uploading a Word file and what you get back is JATS XML and your processed Word file. And obviously in this environment it would depend what you wanted to do with that XML. If you, if there were problems with the XML you might say, actually we need to go back to the Word file, fix something in there and then just hit export again. But this workflow gives you a completely automated process that gives you JATS XML at the end, which might be good to go but might, you know, then again, might not be.

So I thought it would be useful to look at what’s kind of quite new in eXtyles SI. The development of eXtyles Arc has really kick-started some new features in SI that really, we needed to add to make that whole automated workflow work. So if you are familiar with SI already, you’ll know that this sort of the elements in the XML are essentially sort of descriptions of what eXtyles processes are doing. And one you can do in SI, which you can’t do in desktop eXtyles is the SetConfigFlags element, which was added just around two years ago, which allows you to, so Config.ice is a configuration file for eXtyles that has a bunch of settings for your whole eXtyles configuration. So these are always active. For example, there’s a Config flag, if you don’t want to see the Auto-Redact dialogue when you hit the Auto-Redact button, if you don’t want that dialogue to show up, there’s a flag to switch it off. There, if you want to specify where the temp directory is for eXtyles, that’s a Config.ice flag, if you want to say whether to flatten footnotes and endnotes when you activate a file, that’s a Config.ice flag.

So typically with desktop eXtyles, those are really a config.ice file and almost, you know, all desktop users out there never go in and touch it or change it. But with SI you can change individual flags on the fly in a manifest, even on an individual file. So you might say, you know, for this file I want to use a different XML format instead of UTF-8 or something like that. So there are cases where you want to, or in this case I want to delete the images on activation, I don’t want to keep any images, I just want to delete them. But in my regular build I keep the images so you can set individual Config.ice flags just for the duration of this manifest. So you’re not actually editing the Config.ice file but you’re just resetting them. Another one you can use which, so this is the analogous feature to metadata browse from the document information dialogue where you hit a button and you browse to the XML file. So we added this slightly longer ago for SI, and as I say, it allows SI users to essentially mimic the behavior of metadata browse.

But one thing that it also allows you to do, if those who are familiar with metadata browse will know that essentially what’s happening is eXtyles is reading an XML file with the metadata in it, is picking out bits of that XML that you have specified that you need and storing them in the Word file as specific document variables. And so that mapping means that we have to know what the XML is going to look like. It’s ScholarOne or it’s Editorial Manager XML or it’s come from a certain place. So at the moment, again, most desktop customers would set that up for one particular mapping. So they’re expecting one particular kind of XML to come in. But if you had a journal that was using ScholarOne and a journal that was using Editorial Manager or using eJournalPress, you might want to be able to use different mappings because their XMLs don’t all look the same. So with SI you can do that, you can say this is the mapping file I want to use, this is the XML, or you might only for some, in some cases you might want different metadata to be grabbed out of the XML, same source XML but you want to pick different pieces of it.

Calculate timeout, I already showed that timeout thing in the log file that I showed, again that’s about two years old. And so this allows, so at the moment eXtyles does calculate a timeout and it calculates that based on essentially just the size of the Word file. It’s a pretty rough estimate. It gets stored and certain eXtyles processes will keep an eye on that timeout and if we get past that timeout, eXtyles will time out and quit. So that’s a very rough measure, just a rough measure of how big the file is. So we needed for Arc, we needed, to run it on a server we needed some more nuanced abilities there. So you can recalculate the timeout now based on a whole raft of properties of the document that I showed you before, what’s the word count, what size are the images in the file, how many table cells does it contain, how many references are there? Because we know from experience and you’ll know yourselves that certain eXtyles processes take certain amounts of time. So if you are running PubMed reference checking and you’ve got 300 references, that’s going to take quite a while. So one of the factors that goes into this calculation is how many references are there in the file? Which is relevant if you’re running a step like that, but if you’re not then it isn’t relevant. So you can recalculate this timeout. If you delete the images at the beginning of the manifest on activation as I said, then the file now is suddenly a lot smaller and you don’t need that original estimate of the file size that eXtyles calculated and always has calculated. You can now recalculate it based on your file being much smaller and actually cut your time, trim your time out down, optimize your process.

LoadStylesFromTemplate is something that you absolutely cannot do in desktop eXtyles. It’s just not possible. Feature doesn’t exist. We added this just over a year ago and this allows you to just pull in the paragraph and character styles from a template. So at the moment, if you remember, if you run update document information, one of the things that does, as well as allowing you to change things on the document information dialogue, update document information actually also reloads all the paragraph and character styles. So LoadStylesFromTemplate and SI allows you to just pull the styles back in without worrying about fussing around, updating document information. And this can be useful in eXtyles Arc because eXtyles Arc is configured to use a certain set of paragraph styles, and you can then, if your styles are different, which they most likely are, you can then map them afterwards.

We have a style mapper feature that can map style X in, that comes out of eXtyles Arc, to style Y on your palette, you might call references bibliography or something like that. So you map that, but Arc also has a whole set of Auto-Redact rules that it uses to maximize its ability to identify content automatically. So if you want to use those Auto-Redact rules, you kind of have to use the Arc paragraph styles, but then you might also want to pull in your paragraph styles afterwards and then map the Arc styles. So that’s what LoadStylesFromTemplate does.

SaveWordDocumentCopy, I kind of already mentioned this as well, that functionality has existed in eXtyles for quite a while, in eXtyles SI. But it used to be that you just had to specify, hardwire what the file name of the copy you wanted to make was in the manifest. Which was tricky if you were running multiple files from the same manifest cause you’d just overwrite your copy every time. So you can now put a suffix or a prefix on the end of the name of the file that’s running. So if you’re running 10 different files, you can save a copy of each of them with a suffix, you know, dash before export, you know dash after citation matching, so you can have a copy of each of those files.

So as I say that’s useful in situations where you’ve got a manifest with multiple files and we added support for kill files, what we call kill files, again about two and a half years ago now, which allow you to actually stop a manifest running. So if you’ve kicked off a manifest and it’s got 50 files to deal with and you suddenly realize you made a mistake and you didn’t actually want to run that manifest, you can drop a kill file in, it’s just a text file with a particular file name, you drop it into the working file folder and next time eXtyles SI reads the next step of the manifest and it sees that kill file, it’ll say, oh hang on, I need to stop now, and it will cleanly exit. And you can exit one particular doc, you can kill one particular document, and move on to the next one. If you think, oh actually I didn’t mean to run this file, I know it’s going to take ages, you can just kill that one particular file, you can kill the entire manifest or you can kill all manifests. So if you have a script that’s supposed to run several manifests, one after another, you can have a kill file that’ll actually abort that completely. So those can be handy if you realize, but as I say, caveat, they only work between manifest steps. So if the file, you’ve sent a file off that you know is going to take an hour to activate, I’m afraid these can’t help.

So if you’re interested in eXtyles SI, sorry, if you’re interested in eXtyles Arc, you will need SI in order to run it. Lost SI off the end of that sentence, apologies. Another thing to bear in mind, eXtyles SI doesn’t run out of the box. It’s not like desktop eXtyles where you just install it and you are basically ready to go. You will need some in-house expertise to analyze what you’re trying to achieve, to create and to maintain your manifests and potentially to script other aspects of the workflow, copying files around, looking for that timeout, and deciding to kill processing if timeouts, if things have gone on for too long. Things like that, we can help you with that. We’d certainly, you know, any new SI customers, we’ll give you sample manifests, we’ll give you lots of documentation, we’ll explain what everything does, the documentation is very comprehensive. And come talk to us about what your requirements are, let us know, you know, we’re always glad to talk to people about SI.

So thank you very much for your attention and I think I have just left time for questions. How about that?

Patrick: I have a question.

Robin: Sure. Fire away.

Patrick: Yeah. So the manifest, I find that, you know, we’re not using SI and-

Robin: Mm-hmm.

Patrick: Almost none of the processes ever run cleanly. We have to go in and fix something. So it would, does the, is there a point where the manifest would say, wow you’ve got a lot of, like for example in the citation matching, we’ll get the error, oh this is going to flatten stuff, are you sure you want to run the next step? Would it stop there or what happens?

Robin: So I showed you the log file. So it will, anything that would’ve normally come up in a dialogue will go into the log file and it will go into that log file with an error class and with an error number. As I say, the lower the error class, the more significant the problem. So what you would typically do, if you see, so there are, I don’t think there are many eXtyles errors where you would say, “Actually no I don’t, everything has to stop now, we mustn’t go on, otherwise something’s going to be destroyed.” Now if there is, you would obviously want to have your manifest end at that point. You would want a manifest that stopped there. And then what will have happened is, if there is an error, like citation matching, if it reports, there are figures and tables cited that are missing from the file, that will have a certain error class and you can then have a script that says, just have a look at every log file and flag up to me files that show me errors of class 200 or below for example. So 100 to 200, those are the really significant ones that mean we just couldn’t finish. So some things have gone seriously wrong. So you could have a, you can have a, as I say, a script that reads the log files after you’ve finished and flags up to you the real problem files and the log will then tell you what actually happened. You know, what was the thing that caused us to throw that error, we found, you know, there were five unknown references or there were, you know, two figures cited that are, that are not in the document or something like that.

Patrick: Are there still comments put into the Word file?

Robin: Yeah, yeah. So any comments that would, that the desktop version would’ve put into the Word file, SI will also put into the Word file.

Patrick: Now, I’m curious about the automated styling. That’s very interesting to me. I don’t know that I would, I don’t know that I believe it.

Robin: Many people have said that. It’s not optimized for standards, so I should alert you to that. So if you wanted us to do that, that would be a very different conversation, at the moment it is optimized for journal articles. And it’s pretty good, but that’s another conversation.

Christine: It is quite good. We do have a couple of questions. It looks like Stacy and Monica’s questions overlap a little bit. Stacy says, “Can you remove paragraph styles not part of your template, not just add them, in our workflow paragraph styling is done before we run our eXtyles SI and bloated styles crush our pipelines.” Monica says that that’s a problem for them too. They run post-processing cleanup to remove unused styles and reload the styles before sending to SI, but sometimes still get the error, wondering if SI will be able to handle this better at any point?

Robin: So SI, you can do exactly that. So a lot of our manifests, we will run the post processing clean up step to remove unused styles as the very first thing we do before activation. So we will strip out any styles that are not in use and then activation will then pull in all of your styles. That does not guarantee that you will not hit the Word, if anyone who’s, anyone familiar with that, says something about your document contains too many styles for the clipboard, you want to use Normal instead, and you always have to say no. SI cannot prevent that happening. If there is still, if your style palette is very large and the document also contains a reasonable number of styles, we’re going to hit that. We use PTFB Pro, I will not, before the watershed. If the people know what watershed is, explain what PTFB means, but it means “press the something button”, and it means that you can use PTFB Pro to close that dialogue automatically. So that’s what we do with SI because yeah, if Word has that warning that it puts up and you can only script within Word to close a Word dialogue by choosing the default option, which is to say “yes, use Normal instead,” which is the default option for that dialogue. So we, as I say, we use PTFB Pro.

Christine: I know we have only, oh we’re right at 10:30, but Cindy asked, “What is the best way for an existing eXtyles SI customer to get reskilled in eXtyles SI configurations?”

Robin: Talk to us about some admin or developer training, I would say, and we’ll put something bespoke together because it is a specialized set of skills.

Christine: I don’t know if we’re out of time, Jo, I just wanted to check in with you.

Jo: If there’s any questions we can go ahead and take one more, but I think that was the last one. So we’ll go ahead-

Robin: I can certainly follow up with Stacy’s last thing in the comment, we can follow up on that after today.

Industry Updates

Presenter: Bruce Rosenblum, Inera | An Atypon Company

Jo: We’ll get started with Bruce Rosenblum and the fan favorite, the XUG favorite, I should say, industry updates. Let me share my screen, Bruce.

Bruce: I’m going to drive my own slides, Jo.

Jo: Oh, okay, my bad. Let me stop the share then and you can go ahead and share. All right, this needs no intro, so I’m just going to go ahead and let you let you start.

Bruce: Okay, thank you very much. Thank you everyone both colleagues at Inera and our customers who are attending, for being here for yesterday afternoon, this morning, and also for sticking with us as loyal customers over the years. This will wrap up the meeting today. It’s my annual industry update presentation. It’s a mix of updates on a variety of topics that have certainly engaged me and if they haven’t crossed your desk, they’re things that it would be good I think for you to be aware of.

So things that I’m going to cover today, my general industry updates. Last year I gave a full presentation on the statistics for our Cabells integration with Edifix. We haven’t talked about that yesterday and today, I’ll be talking about that a little bit. I want to touch on three topics that I think are particularly hot in the industry right now. You should definitely be keeping your eye on them, and as I’ve done the last few years, I’m going to close with a few personal thoughts.

So those of you who’ve been attending the last four or five years know that I am somewhat of an AI skeptic. I think AI can be very useful to identify patterns that then need human intervention, but I am skeptical that we will ever have completely reliable, for example, self-driving cars. And it seems that people keep trying to bring AI into scholarly publishing. Here’s an example actually, courtesy of our colleagues at J&J, that came to my attention a few months ago. Someone who is in the AI development space logged into their AI open account and gave an instruction to the engine they were using to write an academic thesis in 500 words about GPT3, which is their AI engine, and to add scientific references and citations in the text. The paper was spat out by that engine, he went to submit it to a journal and then discovered that the first problem he had with the journal is that the submission system wanted a first name and surname for the author, and since he considered the AI program, rather than his programming skills, to be the author of the paper, he put none for the surname. I don’t know how many of you find this troubling, I definitely do find it a bit troubling that a paper written by an AI engine without significant human intervention could actually have something significant to say. So something to ponder, you can read the full article about this at the link I provided in Scientific American.

Another thing I like to update each year, Unicode keeps growing, we’re now up to Unicode 15.0 which was released just over a month ago, about 4,500 new characters including 20 new emoji, a total of almost 150,000 characters. You may wonder where all those characters come from, but they keep finding new characters or in some cases just new scripts that they haven’t covered before.

I think I’ve talked about this on and off over the years. Metadata quality at Crossref in my view continues to be a problem and I’m going to step back and immediately say, this is not Crossref’s problem. It’s our problem because we’re the providers of the metadata at Crossref, Crossref doesn’t just create that metadata. I have said for years, this is my hypothesis, I’ve never done a scientific study, that given a hundred records at random, probably three in 10 have minor errors and one in 10 have major errors. This is my exhortation to all, to all of you to do a better job with your Crossref metadata, and Crossref really agrees with this. They put out a blog this past year about data quality issues.

And I abstract a couple of simple examples. It is not uncommon and we certainly see this all the time in the eXtyles and Edifix reference processing and linking, that we’ll see the entire author name in the surname element. Instead of splitting it out into given name and surname. Or if the pages aren’t known, we’ll see values like no or N/A or zero. If the data is not known, don’t include it! If, for example, you use article IDs and not page numbers, don’t include page numbers. There’s a correct way to include article IDs. So as publishing has changed over time, people have not always necessarily updated their Crossref deposits properly. For example, in their transition from traditional pagination to article IDs, it’s worth going back and auditing. And just one more thing I’m seeing in here, I didn’t even notice this before, the second surname in the upper part of the slide “Asta L. Anderson (“. Probably there was something like an author degree that followed that name, that open paren has no business being in there, and that can definitely trip up linking. So it’s worth once in a while to go back and audit your Crossref data to make sure it’s good.

Edifix and Cabells, this is something that we partnered with Cabells with their Predatory Reports database a little over a year ago. We went live in June 2021. When we started the project, we had no idea how frequently we would actually see journal references citing journals that are in the Cabells Predatory Reports database, we had no idea how often we’d see them and why. We now have, last year we gave an update with just four months of data, we now have an update with a full year of data, September 1st of last year to August 31st of this year, and here are the results.

But before I get to the results, just a few caveats. First of all, we excluded all jobs from the sample set that include our sample references because those– our sample references include both a Cabells reference and a retracted reference. Second, I will talk a little bit about retractions, and I’ll point out that when Edifix flags a retracted reference, we have no context as to whether the author is citing it as retracted or whether they’re citing it without knowing it’s retracted. So in some cases, retractions could be cited knowingly in other cases unknowingly. Finally, our Edifix retractions are based on Crossmark data and PubMed data. We know that there are other articles that are not listed in either or, as you’ll see in a few slides, are listed in Crossref but without Crossmark metadata.

So in that 12-month period that we looked at ending August 31st, Edifix ran approximately 2.2 million references, of those almost 5,000 references matched entry in the Cabells Predatory Reports database, which means we’re seeing approximately one in 458 references submitted in the wild to Edifix as matching a journal listed in the Cabells Predatory Reports database. For context, a year ago that number was one in 560. So we are seeing some reasonable degree of consistency in the frequency of citation of journals that appear in the Cabells Predatory Reports database. I need to point out however, that while the average across all customers was one in 458 references, the actual frequency varies widely across different customers. It is not identical across customers. So we ran a secondary analysis of customers who ran at least 2,000 references in those 12 months.

And here’s what we found. First, 11% of those customers had no matches in the Cabells Predatory Reports database. For the remaining 89% of those customers, The ratio varied anywhere from one in 52 references to one in over 2,000 references. So you can see that the ratio varies widely based on the customer. So there is no consistency there and a small number of customers had absolutely no matches. We also thought it would be useful to look at the five Edifix customers who were the heaviest users of Edifix, and those ratios ran in the range of one in 236 to one in 658. So that does give you an idea of what you may see with a larger volume of content in terms of Cabells matches. We continue to monitor this. And we’d also are able to collect some other interesting statistics. Over the course of those 12 months, Cabells Predatory Reports had approximately 16,500 titles. I say approximately because we get monthly updates of the data from Cabells, and I would say on a typical month, and Robin, you can jump in and correct me if I’m wrong, on a typical month we’ll see between 150 and 250 new journals added to the database. So the database is constantly growing. Every once in a while, but very rarely, a journal will be deleted. One of our customers got a report from Edifix about a year ago, questioned the Cabells methodology on that one specific journal, Cabells revisited that journal and they did remove it from the database, but that’s the only title that we’re aware of that that customer has questioned. All the other titles that have been reported to them, they’ve accepted the Cabells report as is.

Out of those 16,500 or so titles, almost a thousand unique titles were matched. So it’s a very broad distribution of journals that are being matched out of the Cabells database, it’s not just 10 or 20 or 50. And approximately 50% of those 959 unique titles had only a single match. So they didn’t occur very frequently. At the other extreme, there were 34 titles that were matched 20 or more times. And of those 34 journals that were matched 20 or more times, they represent 18 different publishers. So this issue is not concentrated all with just one, two, or five publishers. And just to give an example, one of the publishers that appeared with two journals in that most cited list of 34 titles was OMICS International. I call them out because they have been cited by the American FTC for predatory practices. One of those OMICS titles was cited more than a hundred times across those 2 million plus references. And of the 34 most cited journals, there were two titles that were cited more than 400 times. So there are some titles out there that people cite very frequently that are in fact on the Cabells Predatory Reports list.

Also, I will point out that one of these two titles matched more than 400 times, it was matched from jobs across more than 50 Edifix customers. So what this is showing is that the matches in the Cabells Predatory Reports are not all concentrated with a single publisher or with a single area of publication. They are spread quite widely. Finally, a couple of words on retractions. There were 326 jobs–and I’m looking now at jobs, not individual references–where at least one reference was flagged as retracted. So that came through as a total of one in 391 jobs that had a retracted reference. I’d love to work with a larger data set and I expect a year from now we will have much more data on this. But like the Cabells references, the frequency of references flagged as retracted varies widely among customers, we aren’t seeing that it’s the same ratio across all customers. So that number is something we’re also keeping an eye on. And also a reminder with just the reference list, we don’t have the context in which these references were cited. So some of them probably were cited with the knowledge where they were retracted, but many more may well have been cited without that knowledge.

Some of you may be familiar with the NISO CREC working group. Formerly it was known as CORREC. There was a big announcement yesterday that this group in conjunction with the University of Illinois Champaign-Urbana, I’m sorry, Urbana-Champaign, was just awarded a Sloan Foundation grant of nearly $250,000 to work on figuring out how to better spread, or, I’m sorry, prevent the spread of retracted research. Sylvia Hunter from the Inera team is on this, and it’s important to keep an eye on this work, what they’re looking to do ultimately is to create consistent community practices to enable publishers, repositories of all sorts to identify and signal that a publication has been retracted or has an expression of concern. There is not that consistency today.

And let me give you an example. I went looking at some articles in Crossref metadata in the last few days, and I found inconsistent publisher practices. The most reliable way that we have in Edifix and eXtyles To pick off that an article is retracted is to use the Crossref, I’m sorry, the Crossmark assertion article metadata, which gives a CRM item element that specifically, which gives a claim of retracted. Some articles I’ve found instead have Crossmark metadata that have an update type of retraction. These are more often in retraction notices rather than the original articles. But again, we have two different ways of expressing the same thing. Finally, what we discovered somewhat recently and very much to our surprise is that many publishers don’t participate in Crossmark, which we knew, but as a result, we’re not depositing Crossmark metadata. And what they do is they update the title of the reference in Crossref without any consistency.

So here are three examples I very quickly found, one has “retracted:”, the second has “retracted article”, and the third is actually a retraction statement pointing back to an article statement of retraction. But the only way you can reliably pick these off is if you start pursing the article type. This kind of inconsistency needs to be cleaned up and I’m hoping that part of it will come out of the NISO CREC project is an update that will provide much better guidance and that Crossref will join in on that. A few quick comments on some hot industry topics. Research integrity, particularly with the spread of predatory journals, I think has become a much hotter topic. I’ve already somewhat addressed it here just talking about Cabells and retractions. That’s not the only area, there are other areas, for example, simply making sure that authors are sharing their data in an open fashion, making sure conflict of interest statements are complete, et cetera. It’s an area that I expect will see more activity in the next 12 months, so keep your eye on it.

Accessibility is a hot industry topic. We had some great talk about that earlier this morning, I won’t talk about it further here, but in part because of the new European Union regulations, I expect there will be more and more focus on this. And again, if you have questions about how you can move forward on that, particularly in an eXtyles workflow, please reach out to us and talk to us we’re more than happy to help you out.

Finally, open access is becoming a bigger and bigger issue, particularly with the latest OSTP memo in the United States. More and more research is becoming openly available, but not without challenges. I want to talk about that in just a moment. But before that, I just want to point out something that’s not a hot industry topic, just a reality check because those of us who work at in Inera love XML. It is what we eat, drink, and breathe every day because that’s what eXtyles produces, but it’s not a hot industry topic. On the other hand, if we go back and look at everything on that previous slide, what you’ll notice is that all of these require XML. Doing a better job of detecting research integrity requires XML, accessibility absolutely requires XML. PDFs aren’t good enough. Open access requires XML. So while XML itself is old news, it’s our job to make sure that we still do a great job of it and it’s vitally important to make sure that we do it well. So let me talk about open access for a second then I will tell you that as someone who is reading a lot more of lit literature because of my own medical issues, I am a huge fan as a patient of open access. It gives me access to a lot of material.

On the other hand, open access has fueled the rise of predatory journals and really has changed the nature of what publishing is. In particular, what I see now are a lot of publishers really chasing volume of articles for profitability because it’s the author that’s paying, rather than chasing quality in a day when 20 years ago when publishers had to convince librarians to purchase subscriptions based on the quality of research. That’s a huge shift and it will continue to impact the business.

What I’m going to present right here, however, is something that hits a little closer to home for me. I was exchanging emails with a researcher I worked with closely a few months ago, and she wrote to me, “it’s crazy that authors have to pay such a huge amount of money to make their articles open access, but I’m glad that my PIs bite the bullet for this sake of accessibility.” I very quickly reminded that researcher that actually those quote unquote huge fees pay for me and my team and a lot of other people to do all of the R&D and management required to build electronic systems for publishing. Publishing electronically is a lot more complex than publishing in print, and building and maintaining those systems, particularly as requirements keep expanding when we add new persistent IDs, when we add new required sections, when we require data to be openly available, that doesn’t come for free, that all has a cost.

So, when in reply to my comment, the researcher said, “actually, I appreciate the explanation, but the consensus within the scientific community is that the fees are crazy.” And she proceeded to send me a humorous video, which I will not play for you, you can grab the slides and play it on your own time, of one researcher’s possibly humorous view on open access fees, but it does highlight a certain amount of tension going on between the research community and the publishing community.

In the process of all this exchange, I found an article in The Scientist that I thought gave a fairly balanced view of why is it that scientists are paying to have their papers published open access. And the sentence I thought that said it best is that what researchers are paying for is a certification service. Something that says this paper has some credibility. And believe me, I’ve seen a lot of papers that are not at all creditable, they never should have made it through peer review, typically in lower quality journals. So yes, some sort of certification service for science is needed. Today it’s peer review, maybe that will change in the future. But again, managing peer review systems, managing the publication process, it all takes money. So the problem that we all face, we meaning all of us who are in the publishing industry, is that there are negative perceptions of open access fees outside of our offices.

Part of the solution is we need to all make it part of our job to do a better job of explaining ourselves and what those fees are. So the next time you happen to bump into a researcher if you get into a conversation about open access fees, take advantage of the opportunity and give them a different perspective, one that includes what it is we do every day to help make sure their research is published and published in a high quality and sustainable manner.

Finally, I want to close on again my third or fourth year doing so a bit of a personal note. Almost three years ago I woke up one morning to an email from a good friend linking me to this article in the Wall Street Journal about a positive first report of an ALS drug trial. I knew about the drug trial, I had not yet heard the results prior to this newspaper report. This for me was incredibly exciting news because for those of you who are watching who are not aware of this, I was diagnosed with ALS five years ago. Three years later, just a month ago, barely a month ago, the FDA approved this new medication. Now for someone with ALS with an expected lifespan of two to five years from diagnosis, three years is an awfully long time. So I want to take a moment to walk you through what happened in those three years. The first thing that happened for me nine long months after that first article, is that an article, is published in the New England Journal of Medicine with the clinical trial results. And by the way, this article is published with eXtyles, I’m incredibly proud of that fact, and it’s hosted on Literatum by our Wiley Partner Solutions company.

This had the results, the data that I was looking for to tell me how effective this drug would be. Then just six weeks later, the same research team, and by the way, this is the research team that’s also taking care of me, publish in a Wiley journal called Muscle and Nerve, a survivability analysis that showed, gee, I can start adding potentially months to my life. Well, then what the company did, after much discussion because it’s a new small biotech, is they filed a new drug application with the FDA. And finally, last March they had an advisory committee hearing with the FDA, and the FDA actually voted six to four against approving the drug. Part of why they did that is because they had some statistical questions with the Muscle and Nerve article in terms of this survivability analysis. And part of that is because they were being asked to approve a drug with only a single phase 2/3 drug study instead of multiple drug studies. Well, again, the scholarly publishing world helped out here. A second article was written and published in Muscle and Nerve to respond to the FDA’s statistical concerns and that was finally published last May. And then subsequent to that article being published. And then most importantly for me, this past June Health Canada – and if there’s anyone from CADTH still on the call, thank you, thank you, thank you, even though all of you personally had nothing to do with this – Health Canada gave a provisional approval to this new drug in Canada.

There was also a lot of community advocacy by ALS patients and their caregivers. And at the beginning of September, there was an unprecedented second advisory committee at the FDA. This virtually never happens. And this time a complete turnaround from last March, they voted 7 to 2 for approval. On September 29th, not even a month ago, the FDA approved this new drug. For those of you who are wondering if I’m on it, I’m waiting for insurance approval. That does not happen overnight in the US. But it presents some promise. For me it presents, hey, maybe it will keep me going just a little bit longer.

But the reason I share all of this with you is because those published clinical trial results in scientific and medical journals were a critical part of the submission to the FDA. So I share this with you, this story as a simple reminder that what we do every day has an impact on the lives of real people. People like me, people like all of us on this call. It doesn’t matter whether we’re talking about a new ALS drug, a new cancer treatment, research about global warming. All of you may publish in different areas, but what we do when we publish that research, that does make a difference.

So with that as a closing thought, I want to share a few thank yous. First of all, again, thank you to all of you for working with us as customers, and especially over the last years of COVID, which have been challenging for continuing to work with us and work closely. Second, a huge thank you to the Inera team. As was mentioned in Liz’s talk yesterday, you’re now an entirely virtual team, we now have a minority of people in the Boston area, which is something new to us, but particularly through, again, two and a half challenging years of COVID, and again through living with me and my medical situation, you have been absolutely incredible and I thank you for that.

The final thank you is to someone who’s not on the call today but who I think all of you know, and that’s Irina Golfman Rosenblum, Inera’s founder who retired three years ago. She retired because she needed to take on the role of caregiver for me. And I can tell you that is an incredibly difficult and challenging role. Far more difficult than starting and building the company Inera and the product eXtyles that you all use all the time. And an extraordinary thank you to Irina for taking care of me in part so that I can just continue to work with all of you. So thank you very much and we look forward to seeing all of you next year. And Jo, I don’t know if there are any questions, but I will unshare my screen.

Jo: I did not see any come through in the chat. A couple great comments throughout, but specifically by Debbie that just is a great tweet. Just electronic publishing is more complex than print, yes, but if anyone wants to raise their hand and ask a question, now would be the time. I know we’re right at 12 o’clock a little over, but if you don’t mind staying on for questions, we’re happy to take them. I see a hand raised by Bill, all right, go ahead Bill.

Bill K.: Well, I think in this group, obviously we all are very grateful for to Bruce for what he’s done with Inera and enabling so much progress for the work we do. But I just want to make sure everybody’s remembering how much of a contribution he’s made to the industry in general. I mean the whole XML model that we live and breathe with and depend on, originated with work done by Bruce and also Debbie and Tommie way back in the day. So, you know, huge, huge kudos to Bruce for getting us where we are in the industry as a whole. Whether or not we use eXtyles, we all use the work of Bruce Rosenblum.

Bruce: Thank you very much, Bill, that’s very kind of you.

Jo: All right, any other question?

Debbie: And Bill, remember that we also had a whole lot of fun doing it.

Jo: Some great comments in the chat.

Liz: Yes, and thank for being so lively during the chat. It’s fantastic and so much fun and impossible to keep up with at some points, which is great.

Jo: Okay, if we don’t have any questions, I’ll finish us off. Again, thank you for coming. I’m glad everyone seems to, oh, words. I’m glad everyone seems to have enjoyed another virtual XUG. I– there will be an email going out afterwards with a survey, ’cause we really do appreciate your feedback and it helps us build next year’s program and future conferences. I also know that some people have had trouble getting the emails this year, so I’ve added a link to the day 2 landing page. And we really appreciate everyone coming out and hope to see y’all for future years to come. Liz, anything to add?

Liz: No, that’s it. Thank you, like I said, we’re scanning through the chat. I think we do have a record of the chat.

Jo: Yes we do.

Liz: So we can follow up. I think there’s been a couple questions in there, but I think all of the materials that people are asking for, you’ll be able to access once we have them posted.

Jo: And I’m also posting the survey in the chat right now.

Liz: For anyone who wants to get started right away.

Jo: We do use that feedback. All right, if that’s it, unless has anything else to add, this closes the 18th annual XUG, thank you very much.