Fisking Chris Hofstader

Update for Katieplayer readers

(2007.02.17)    I am not “on the wrong side of history” and I have nothing to apologize for when it comes to writing about Macintosh accessibility.

Chris Hofstader is a person living with blindness who used to be a senior vice-president at Freedom Scientific, makers of everyone’s favourite screen reader, Jaws (perverse official orthography: JAWS). He had the worst people skills of anyone I’d ever done business with bar none. I distinctly recall how many E-mails it took to be permitted to talk to him on the phone, and then the 45-minute screaming match that ensued as I attempted to persuade him to modify Jaws to work better with standards-compliant sites. The strange thing is that the screaming match was apparently a hazing ritual of some kind, as Hofstader spent the next year or so actually listening to me and answering my mail. I am not the only person who has had to negotiate with him by megaphone.

Anyway, Hofstader no longer works for the least responsive and helpful company in the industry since Quark. He is now doing what everyone else does, blogging. I had previously documented his “written direct testimony” in the Microsoft antitrust case. It was an apologia for Microsoft’s accessibility methods, which can be summarized as follows:

• Staff an entire department
• Claim that your APIs are sufficient
• Force adaptive-technology makers to reverse-engineer your software anyway because your APIs don’t work and your own software violates them
• Leave accessibility to the open market at high prices

Before I go on, let me ask a skill-testing question here: Do you believe your computer should be accessible by default or do you believe that accessibility should be something you pay extra for? Your answer will concisely illustrate your understanding of equality. At least we know where Hofstader stands. He’s on the wrong side of history, but puts up a good fight, particularly when the opponent, however he might define that, is Apple.

1. March 30:

   I really like some of the people working on [Apple’s] screen reader very much and don’t want to trash it, as I don’t want to continue [sic] to stamp on the toes of old friends. I will just suggest that anyone interested in it read Jay Leventhal’s article in Access World… and try to give it a whirl at an Apple salon shop at your local mall before committing to using it.

   In other words, read a review with factual errors before you decide to abandon his former company’s product.

   On the iPod:

   With a variety of different accessible portables… why does Apple remain so completely bigoted against us blinks?

   The rest of the piece is an overlong, if rather jovial, screed that the iPod isn’t accessible to blind people. It isn’t and could be. It is a major failing. But a comparable product from Microsoft, the Xbox, isn’t accessible, either. Neither is Microsoft Windows accessible as delivered to the blind user, and neither is any other competing music player not custom-made for “blinks,” not that any of those matter commercially.

2. March 31:

   The latest version of Rockbox apparently works on an iPod and has optional self-voicing menus as well…. Also, it comes as no surprise that the Rockbox developers build the software voluntarily and without any notable support from Apple.

   What might that mean? How would that be distinguishable from the case of having to write screen readers for Xbox? In neither case do they arrive as standard equipment with the product. How much help from a manufacturer do you need to redress a failure by that manufacturer? (How much help did Freedom Scientific get from Microsoft in reverse-engineering and hacking their nonstandard and undocumented programming methods?)

   If their accessibility solution falls into “let the community do it,” they don’t really have an actual accessibility strategy.

   Then how does he explain Jaws, which had to be created because Microsoft didn’t make Windows intrinsically accessible?

   [T]he Access World article slights the VoiceOver documentation[, which, with the help system,] is quite robust. As I have not seen or tried to read either the help system that comes with OS X or checked into the online documentation either, I cannot make an informed comment.

   That didn’t stop him from endorsing the original review. Why not follow his own advice and use VoiceOver himself before writing about it? (Hint, though: Trying it out in a bustling and noisy Apple Store will not suffice.)

   Big Blue… provided us with a completely open system; we could replace interrupts at will, easily disassemble the BIOS and make incredible hardware and software hacks which, ultimately, led to its dominance [sic]. Apple took the hackers out of the picture so only programmers who liked following rules could write software for it. Jaws for DOS, Vocal-Eyes and all of the other screen readers that blinks could use to do jobs, get educations and learn their own way to hack had become impossible on an Apple platform.

   So there you have it: A hackable system (like present-day Windows) is more accessible than one designed for accessibility from scratch (like OS X Tiger). And here as elsewhere, Hofstader accepts that “hacking” undocumented systems is a viable development strategy. To be consistent, shouldn’t he be calling for Microsoft and Windows software makers to hew tightly to the documentation? Wouldn’t that make his job, or his former job, easier? At that point, Windows accessibility development would not be structurally different from Macintosh or Linux accessibility development.

   And indeed “Jaws for DOS” might well be impossible “on an Apple platform.” OutSpoken for Macintosh wasn’t impossible, though, was it? (And there was even a Windows version.)

3. April 3:

   MSAA is still the best game in town [for Vista]. This is of course also true for the Macintosh and Gnome accessibility APIs which, if a program is not written specifically for Cocoa or using the Gnome API, it is also not ready for prime time with a screen reader. I apologize to my friends at Microsoft and to anyone else who may have been misled by this error.

   It’s a limitation that programmers have to follow operating-system conventions on Mac and Linux, but a strength when they have to do it on Windows (via MSAA)? And why apologize to one’s friends at Microsoft?

   I’ve fought too many religious wars over technology: Windows v. Mac v. GNU, proprietary v. open or free, patent v. discovery, Jaws v. all other blindness products and many others.

   Why the use of the past tense, I wonder?

   I’ve also been involved in the document accessibility wars having sent hundreds… [of] E-mails to Webmasters with pointers to W3C/WAI and Deque Systems so they can fix their sites

   Yet there is actually very little payoff to a Jaws user when a Web developer codes to standards. (In what sense is Deque Systems a standards body?) Jaws has to handle compliant and noncompliant sites, and it adds little or nothing in the way of functions or ease of use when browsing a compliant site. This is a program that still can’t tell you what the title attribute says on nearly any element, including hyperlinks.

   Finally, I want to state for the official record that I have no technological religion. I do not pray at the altars of Microsoft, Apple, Sun or any other technology company.

   It isn’t a question of praying to altars. He’s completely embedded in, and a tireless defender of, Windows, i.e., Microsoft. (He later writes: “I can’t go out and tell the hundreds of thousands of blinks who use Windows-based screen readers to stop working until Apple gets around to filling their requirements…. Microsoft Windows XP… is the most robust platform on which a blind person can perform their job…. Maybe, someday, Apple or Gnome will surpass the Windows platform… but the overall package still tips heavily on the Microsoft side.”)

   All three of these companies as well as Dell, Sony, HP and others, including Apple, have, from time to time, sold me tools with which I can do my job.

   He’s really only discussing three platforms, and has not explained what he’s bought lately that came from Apple or Sun. (His anecdotes about the Apple II are a bit stale at this late stage.)

   This morning, I received an E-mail from a very smart guy who said that because a particular task was impossible to perform with Safari and the Macintosh VoiceOver screen reader that he would forego performing that task until the Apple guys got it right. He points to unethical business practices carried out by Microsoft to build their enormous market share. I don’t follow corporate law and don’t understand antitrust laws very well so I won’t comment on Microsoft’s history.

   That didn’t stop him from testifying in Microsoft’s antitrust trial.

4. April 7:

   Many people with vision impairments lament the high prices of AT products and scream about the unfair notion that screen reader vendors live high on the hog of windfall profits…. I can already hear some of you arguing that this proves that the current model of making assistive technology products is broken. I can hear the cries of the open-source people and of those who believe that screen readers should be built into operating systems.

   Well, they should. And we know the Jaws model is broken; Apple and Sun already broke it. All that matters now is improvement.

   Unfortunately, VoiceOver by Apple isn’t mature enough to support all of the programs blind people need at work.

   Of course it isn’t. It’s a Version 1.0 product. But that does not invalidate the concept.

   The GNU/Linux Gnome-based screen readers are hardly more than demos.

   I don’t know about that. I have indeed seen a demo of the Sun accessibility suite (by the almost-equally-irascible Peter Korn, who derided me as “my shill in the audience”) and it seemed to work. Linux tools also seemed to be quite functional for one of the blind people in the room, who used open-source screen reading every day.

   If all this is code for “These products don’t work with Microsoft Office,” well, of course they don’t, since it doesn’t exist for Linux and is a different program altogether on Macintosh. I assume Hofstader understands that a declaration along the lines of “Non-Windows screen readers don’t work with Windows software” is tautological and nonsensical.

   So, for now, the model we got is the only game in town, and if a trade war hurts the screen reader vendors dramatically, any kind of innovation, any kind of progress, and any kind of future for many blinks will immediately go on hold.

   “The model we got” is great for well-to-do blind people who must use Microsoft Office, but that says nothing about anybody else. Hofstader seems to be precluding the “future” improvement to Mac and Linux platforms he had mentioned before. Does he also think that Windows screen reading will not improve in the future?