Eight days ago, I wrote an article here called “The Death of Screen Reader Innovation” and posted it on Saturday July 20, 2013. For a variety of reasons I didn’t expect this article to be much of a hit. I posted the item on a Saturday, typically a slow day for traffic on this site, I hardly said anything that I hadn’t written a bunch of times in my old “Blind Confidential” blog and, as often happens, I assumed that it the readers would come from my usual echo chamber of fans who almost always agree with me. I thought that the article might be perceived as self serving as it sought to redefine my own place in the history of access technology and told the story entirely from my own perspective.
Within hours of posting the article, though, I realized that I had touched some sort of raw nerve among people who find screen readers interesting. Readers wrote comment after comment and, even more so, sent me private communications via the contact form on this site. By Wednesday of the week, a pointer to the article had been tweeted and retweeted repeatedly and the comments and emails kept coming. At some point during the week, someone posted a link to the article on the forums at AudioGames.net and the tone of the communications turned very dark. I realized at this point that, perhaps for the first time with this blog, I had reached the rank and file screen reader users – a population with whom I rarely have any contact.
I wrote “Death of Screen Reader Innovation” thinking, as I often do, of cool new ideas that have never made it into screen readers. Ideas that could improve the efficiency with which one interacts with a screen reader, ideas that could improve the understanding of information when using a screen reader, dominate my own thoughts and is what turns me on. This is my blog and I write about what interests me most.
At the same time, many of those commenting on the article, either publicly or privately, wrote that, while innovation is a nice thing and that JAWS releases were exciting for that reason years ago, the biggest problem with screen readers today is that they have deteriorated in quality, support fewer applications necessary at job sites and are often the barriers to employment and promotion.
On the day on which I resigned from my position as VP/Software Engineering at Freedom Scientific, JAWS 6.0 (the latest release at that time) supported applications required in many job sites, including professional applications like Microsoft Access and Microsoft Project among others. While I had no personal experience with Access (it just wasn’t important to my job at FS), I used Project daily. Some of the comments I received described situations in which a blind person was refused a job or a promotion entirely because they cannot use these two important pieces of software. These communications were heartbreaking, perfectly smart and talented individuals prevented from succeeding due to deterioration of their access technology wrote about the barriers they now face with great sadness.
I also heard a lot about the failures of JAWS in remote access solutions like Citrix and the Microsoft equivalent. Numerous people told me that they lost jobs when JAWS stopped working in these areas required for their continued employment. I also heard from students and professionals frustrated that software like Microsoft Word, Excel and Powerpoint no longer worked as well with JAWS as they once did. I heard all sorts of stories about all kinds of problems with JAWS in areas in which it once worked very well but, due to a refusal to fix bugs, now takes a blind user far longer to accomplish the same things as it did a number of years ago.
Quite a to of these people stated that they no longer maintain their JAWS Software Maintenance Agreement (SMA) as they found that continuing to pay for upgrades was pointless – the problems they experienced were ignored and the apps they need were no longer supported. Most of these people have moved to NVDA, VoiceOver or System Access which all have financial benefits over JAWS but, sadly, none provide the extensive support for professional applications needed in the workplace by these users who, due to such, are losing jobs and showing high levels of despair.
When I first started at Henter-Joyce (now called Freedom Scientific), I reported to a terrific guy and then General Manager of the company, Jerry Bowman. On my first day on the job, he told me, “Ted will screw up schedules for his pet projects. It’s your job to keep projects on track and to manage Ted.”
“I’m to ‘manage the company’s president?’ I’m to manage a legend?'” I asked myself.
A few weeks into my HJ tenure, I learned what Jerry meant. I entered the office of one of our full time software engineers to discuss a new feature. He told me that he wasn’t working on the assignment I had given him. When I asked for an explanation, the engineer told me that Ted had come to him and asked that he work on scripts for an application I had never heard of. Frustrated, I stomped down the hall into Ted Henter’s office and asked him what the hell he thought he was doing with my team.
In short, Ted explained that Joe Blow (a name we’ll use for a random user) had called him personally (at that point, anyone calling could just ask for the boss and be connected immediately by the receptionist), told him he was in jeopardy of losing his job as a trucking dispatcher due to the latest version of the software and asked that it be scripted so he could continue in his job. Ted Henter was willing to shut down the entire shop until we made the changes to save a single user’s position – a commitment that could only come from a visionary who believed deeply in the JAWS acronym: JOB Access With Speech.
At that time, in 1997, I had spent most of my adult life in software engineering or the management thereof. I had delivered a ton of releases for companies like Microsoft, Wordperfect and Novell. I had shipped software on which I had worked to millions and millions of people. Shutting down a development cycle for a single bug wasn’t unheard of but doing so for a single user seemed absurd. I tried to argue that we had thousands of users waiting on the next release and its cool new features but Ted wouldn’t listen. To Ted, if a single user was going to lose a job, it was our commitment to the user to fix their problem. This attitude, pounded into my feeble mind by Henter, had to be pounded back out after the merger that formed Freedom Scientific so I would look at “the 80%” and ignore the individuals. Obviously, this attitude continues today and seems to have gotten worse.
The Accessibility Echo Chamber
On Thursday, I was on the phone with fellow Access Technology insider, Jim Tobias. Our conversation was centered in fairly obscure details of the AT industry, focussing mainly on the size of the market and the estimated marketshare of various screen readers. As an aside, though, Jim asked, “Are we always just talking to the same people?” I was unsure but as comments and emails continued to pour in from the rank and file, I realized that we so-called “leaders” in this field pretty much only talk to each other and focus our attention on innovation while missing out on hearing about the daily problems encountered by those we claim to serve.
To wit: On my Macbook Air, I have installed the following screen readers:
I can, therefore, use applications on three separate operating systems and find the one most convenient for me for any specific task. Most users have one or two screen readers, typically some older version of JAWS they got from a purchasing agent at their school or job and NVDA because it comes at no cost. A bunch are now using VoiceOver on a Macintosh and enjoying many of its features but complaining again that professional applications aren’t supported.
From the feedback I received, I’m certain that our echo chamber needs to expand substantially. There are a whole lot of people out there who cannot access employment because they cannot access the software they need to do these jobs. This is an abject failure of the AT industry to even maintain support for those who already have jobs to maintain such and to hopefully get a promotion or two int he future.
When I worked at FS, we would often start supporting an application just because I wanted to use it myself. I, like all blind managers at FS, used Microsoft Project daily. We used it to track progress on all of our software engineering tasks down to the hour. We used it to accessibly view Gantt charts and project budgets. I cannot imagine running such a complicated department with so many engineers, so many tasks and so many different projects and products without such a tool. I guess it would be impossible for a blind person to do the job I had at FS today and, worse, FS seems uninterested in making it possible.
The 508 Failure
If you don’t know about Section 508 of the Rehabilitation Act, you should look it up. It was one of the most important bits of legislation in the history of accessible technology. In short, 508 required that all electronic and information technology (E&IT) purchased by the US federal government be accessible by about a decade ago. At first, this law caused a whole lot of mainstream software companies, businesses like Microsoft, Apple, Oracle, Sun Microsystems, Adobe, MacroMedia, PeopleSoft, Citrix and other giants with very deep pockets to start taking accessibility seriously. Unfortunately, loopholes in the law and a lack of serious enforcement by the General Services Administration (GSA), the agency charged with enforcing the act, has resulted in software being used in the federal government being marginally accessible at best. The downstream effects of federal accessibility laws, therefore, have also not been seen as, if the feds don’t require accessibility, no one will.
Many people with vision impairment work in various government jobs. I have friends at Veterans Administration, Department of Education, Department of Labor, CIA, NSA and in most other government agencies. I’d bet that you, my loyal readers know blind people in these roles too. Just ask them, is all of the software they need to use accessible via their screen reader? They will say “no” but have no idea how to correct the situation.
One of the applications most requested by those who sent in comments and emails about “Death of Screen Reader Innovation,” [was Intuit’s popular QuickBooks. While I worked at FS, I had the opportunity to speak to William Campbell, then Intuit CEO, on the phone and he said, “I can’t get them to make a Macintosh version of QuickBooks, I really cannot promise an accessible one.” He then described the fierce independence each product team has at Intuit and that even the CEO can’t get things done.
We had a meeting with the QuickBooks development team. They didn’t support MSAA (UIA hadn’t been invented yet), they didn’t use any standard controls and they had no VBA scripting ability. In short, QuickBooks is a black box that a tool like a screen reader cannot access. In that meeting, I asked if they used automated testing tools and when they said they did, I proffered the idea that they let us access QuickBooks by the test tool API which would have solved the problem. The engineering team agreed and we thought we had a solution; sadly, the product managers at Intuit killed the notion and the project stopped there.
While I think FS and the other screen reader manufacturers and publishers have a lot of faults, I can say with certainty that all would support QuickBooks if it was technically possible. Don’t blame the screen readers for this one, poor accessibility lies entirely in the hands of Intuit.
Where DO Screen Readers End and Applications Begin?
As with QuickBooks, we cannot blame the screen reader authors for inaccessibility in many applications. I haven’t looked under the hood at Access and Project in the past nine years. It is possible, therefore, that Microsoft broke the API that JAWS uses to communicate with these important applications and that, therefore, it is impossible for JAWS to offer the full level of support that was possible in the past. I’m willing, however, to venture a wager that the truth lies somewhere in between, Microsoft changed its API and JAWS wasn’t maintained well enough to keep current with said changes.
We do, however, need to recognize that the problems with deteriorating support for applications may not lie entirely within the screen rear industry. It is likely that due to lack of 508 enforcement, companies like Microsoft have taken accessibility less seriously as, if they’re not losing sales, they have no motivation to remedy the problems.
What About Window-Eyes?
Some noted that I neglected to say much, if anything, about GW Micro or their Window-Eyes screen reader in the article on the death of innovation. This is absolutely true, of all of the screen readers out there, Window-Eyes is the one about which I know the least. When, in 1998, I first joined HJ, I was told that JAWS and Window-Eyes had approximately the same marketshare with about 35,000 users each. Then, Window-Eyes seemed like a formidable foe. They had the virtual buffer-like, MSAA mode which was certainly nicer than the JAWS “screen reformatting” approach to the web and they did a bunch of things better than we could with our product.
Over the following six years, though, the years I spent in the business, the GW Micro Window-Eyes strategy turned from one that could build a competitive screen reader into such a bizarre approach that it seemed simply incomprehensible. Doug Geoffray, as CEO of GW, would assail JAWS for simply having a scripting language, the core power of our software, by telling users that they “had to learn a complicated programming language,” a statement that was simply a lie.
GW then allowed Window-Eyes to fall far behind JAWS in virtually all professional applications. As we were adding support for Access and Project, as we were massively improving the experience in the MS Office applications, Window-Eyes did little to keep up. As we defined the professional experience, GW Micro left their users behind.
At the crux of the difference between the HJ/FS approach and that taken by GW was MSAA, the first standard accessibility API. GW Micro felt that MSAA was the future and banked heavily on it. We realized that MSAA was, indeed, a standard but it was an awful standard. The API simply didn’t allow for access to a whole lot of information deemed necessary for a user to have full access to these often complex applications so, we used it where it was useful but sought other approaches where available to provide essential augmentations to data presented to JAWS users. Largely with the blessing of developers at Microsoft, Wordperfect (Novell) and other companies, we used whatever we could get access to in order to make the experience as excellent for our users as possible.
During that time, big companies like Adobe and MacroMedia would come to FS as we emerged as the de facto standard screen reader for help making their software accessible. At FS, we would point them to our consulting services group, a team with a price tag of $125 per hour or more and offer to do a proposal for a number of tens of thousands of dollars. Certainly, Adobe and their peers can afford to pay the comparatively small Freedom Scientific for our time, couldn’t they? Instead, in most cases, these big companies sought help from GW Micro who, at no charge to the big organization, would help them make an MSAA based solution, enforcing mediocre accessibility permanently. It boggled our minds at FS, why would GW give away hundreds of thousands of dollars in effort to help make billion dollar businesses make hundreds of millions of dollars in federal sales with no guaranteed return in units shipped of Window-Eyes? This was, plain and simply, bad business but it was a poor strategy for GW that worked very well for FS. GW would spend the dollars on research and development without compensation from the industry giants and, a few months later, after they did the work, we’d add support into JAWS benefitting from their work while growing our own marketshare.
GW’s intense commitment to the failing MSAA standard caused them to lose years of competition with JAWS and allowed FS to grow from a share roughly tied with Window-Eyes into a powerhouse with a more than 80% share around the world. We focussed on features and functionality for our users without regard to a poor standard; GW obsessed on MSAA seeing it as an easy answer to big problems, they got lazy and allowed their loyal users to fall far behind those who used JAWS, especially in the workplace.
Is the Future All Gloom and Despair?
A reader of this or my “Death Of Innovation” article may think, due to the various economic motives demonstrated for not improving or maintaining a screen reader that the future is quite bleak. The people who wrote to me after “Death Of” listed such a wide collection of professional humiliations caused by the failure of their access technology to keep up with developments in mainstream technology that I was brought to tears while reading more than one of their comments and emails. One went so far as to suggest that he contemplates suicide due to his lack of employment prospects. This sadness is real, very deep and, from what I could learn, is ignored by technical experts like myself and my friends and by the AT companies as well.
While I have no real numbers to support this, I think, based on the outpouring of frustration and despair by those who wrote to me, that most screen reader users experience a tremendous level of frustration, agony, fear, anger and hopelessness every day when they try to use these products. As I wrote at the top of this article, I have a bunch of different screen readers installed and have the technical wherewithal to handle a lot of problems. I am not representative of the community at large and, therefore, live in a state of privilege unknown to most other blind people.
The outlook, however, is not entirely bad. NVDA and Orca, in my mind, are the future. These projects are not run in the same way as screen readers made by either giants like Apple and Microsoft nor do they have the intense economic pressures faced by products from the AT industry. Free software solutions can be funded by contributions, corporate or government grants and, because they are driven by user communities, developers are usually also users and can write code to do the things they want and need.
The free software model isn’t perfect, we run into the problem of having too few hackers in our community to do the things mainstream free software efforts like Apache can do. Again, this comes to a problem of money but, if NVDA and Orca can find contributions and grants, they are better positioned to make the important features possible compared to a company that also needs to focus on the 80% rather than on actual usability.
How Can You Help?
A lot of people who wrote to me asked if they could do anything, as non-programming, blind individuals to help move things forward. Instead of answering them individually, I chose to include this section in this article.
Some things you can do to make screen readers better even if you can’t write code:
- Many of you wrote that you no longer maintain your JAWS SMA. This means you’re saving $125 per year. Please take a portion of these dollars, let’s ay $25 as a minimum and send it to either the NVDA or Orca projects depending upon whether you care more about Windows or GNU/Linux.
- If you have writing or teaching skills, maybe you could volunteer to help these projects write documentation and/or tutorials. These areas are cost centers for all software development and are harder for free software projects to accomplish so, even if your work isn’t great, it will be appreciated by some people out there. We need progress; not perfection.
- Try to convince others to support NVDA or Orca by using the software, sending contributions or participating in ways they feel comfortable that would also be useful to the projects.
- Think of your own ways of helping and make suggestions about such to the people involved in these projects. I certainly don’t have all of the ideas and you should feel free to add your own.
How Will I Work To Effect Change?
A bunch of people asked, “If things are so bad, what are you going to do about it?” My initial answer was to remind people of what it says in my blog and Twitter profiles, “I’m a crank, a crackpot and a stoner with a blog. I don’t do anything worthwhile beyond being a social critic, a public intellectual and a writer who discusses these problems. My role is to write these articles and hope they effect change in some indirect manner.”
Somehow, this didn’t feel right to me. I can state the problem but I can’t offer a solution. Some of this comes from my skills having eroded over time. I am pretty out of date. I know little of web programming and nothing beyond the standards for such. I’m an old C and assembly language hacker with no skills for working on accessibility software anymore.
Then, my friend Gordon, a super web and database hacker and one who has worked on some accessibility contracts over the years told me to stop crying and get back on the horse. He reminded me that I had pretty much taught myself everything else I know and that, just because I’m 53 years old, I’ve no excuse for learning a few new skills if, indeed, the work was so fucking important.
Over the past week or so, I decided that I would try to help by:
Working on WordPress accessibility by hacking some plugins and the like to make WordPress administration simpler for people who use access technology. I’ve never spent any time using the PHP programming language nor have I studied how WordPress itself works so, thus far, progress has been slow. After about a week of reading a lot of code, it’s starting to make sense to me and, perhaps, I’ll actually be doing interesting things soon. I think making the tools accessible is really important as, given that 70% of all new web sites going online today will use WordPress, making it accessible out-of-the-box will cascade forward making lots of automatically accessible web sites in the future.
As per a request by my good friend Pratik Patel, founder and CEO of EZ Fire, I will also be working on NVDA scripts to better its support in Microsoft and Open Office applications. As I did a lot of the design for JAWS once heralded support in these programs and that I had prototyped some of those scripts, I have a solid understanding of how the screen reader side of this works. I don’t, however, know the Python programming language at all so there will be lots of learning involved in this effort as well.
I’m going to try to learn a whole lot in order to do a few cool things. I don’t know if I’ll be good at these tasks but given that I’ve been programming since I was eleven years old (42 summers ago) I’m fairly confident I can figure it out. At the very least, I’m going to try as that’s all I can do.
A Few Heroes
Before I end, I want to mention a few people out there doing the really heavy lifting to support accessibility for this community. Friends of the community Mick Curran and Jamie The who lead the NVDA project come first to mind. These guys, ignoring offers for high paying software engineering jobs, spend their days and nights making the best free screen reader for Windows out there. They are approaching the JAWS level of support in many professional apps and do an even better job in some areas. Please, send these guys your thanks and some money so they can continue in their heroic efforts.
Joanmarie Diggs is the heroine of the Orca screen reader. She has a full time day job that involves Orca work but she has lots of unrelated tasks as well. She spends hours upon hours of otherwise free time ensuring that her users have her best efforts in hand. Please send Joanie your thanks and send a few bucks to her project if you like using the GNU/Linux OS or think you may like it in the future.
Having now read a couple of hundred emails from desperate users struggling to find employment or general satisfaction in their lives through computing, I conclude:
- Many blind people out in the real world are suffering badly and unnecessarily due to the failures of the AT industry, large mainstream OS companies like Apple, Microsoft and Google, thinkers like me who put cool new ideas ahead of solid, stable and usable technology, the economy, the government and the vocational/rehabilitation community.
- The people who are suffering aren’t lazy and aren’t avoiding work. Quite the contrary, these people seem deeply committed to contributing to our society but simply do not have opportunities nor the technology to get such in the current landscape.
- Section 508 and similar state regulations have helped in the past but it seems today that mainstream companies can get away with a lot of poor accessibility due to lack of serious enforcement by GSA.
- Our community can take ownership of free software screen readers like Orca and NVDA, getting involved in the projects and take control of our own technological destiny.
It is my never humble opinion that we can create an excellent technological future if we do, indeed, take ownership of our own problems. To date, we have let AT companies and mainstream OS authors drive the technology we need to use and the results have been mediocre at best and horrible for many.