Visually Impaired User Weighs In on Assistive Technology Debate

Disclaimer: I'm the lead programmer for Serotek Corporation, makers of FreedomBox and FreedomBox System Access. However, this comment is my own, and not an official statement from Serotek. Neither is this a plug for FreedomBox. That said, my involvement with Serotek has probably biased me in some ways. First, I'll address access to OpenDocument, since that issue served as the launching point for this article. In my opinion, we shouldn't expect anyone to refrain from using OpenDocument merely because current screen readers don't happen to support OpenOffice.org. Neither should Microsoft be expected to support it in Office so blind people can use it. Instead, adaptive technology developers should implement access to the applications that their users need. If OpenDocument becomes a widely used format, then these developers (including Serotek) will need to make access to OpenOffice.org a priority. I include the preceding prerequisite because we must prioritize our work. However, I don't think the status of current adaptive technology should be used to justify the Microsoft Office monopoly. I look forward to change in this area. I strongly disagree with the final paragraph of the article, which Peter quoted. If we (blind and visually impaired people) continue to ask state and federal governments to spend money on us, then it seems to me that people will continue to view the blind community as a charity case. Naturally, I don't think state governments should be spending so much on JAWS and MAGic either. Oddly, the second to last paragraph hits the mark; in a free enterprise system, adaptive technology companies should compete for the business of the users, not the state governments. Fortunately, newer companies are starting to develop products that are both useful and affordable. I'm not just talking about Serotek; another example is Code Factory, the developer of Mobile Speak Pocket, a screen reader for Pocket PC devices. Let's hope this trend continues. Even on Linux and Mac OS X, there's plenty of room for adaptive technology companies to fill in the gaps in the accessibility provided by the open-source desktops and Apple, respectively. At this point, I'd like to get on my soapbox and talk about what I think is missing in adaptive technology on these platforms -- at the risk of sparking some controversy. According to Scott, good adaptive technology for the blind "must be fully compatible with all available applications for the best readability and functionality." He also rightly notes that in some cases, Microsoft Narrator doesn't provide any access at all. So what do the current screen readers for Linux and Mac OS X have in common with Narrator? This may not seem to be a fair comparison given that these screen readers provide access to much more than Narrator does, but they are like Narrator in one important respect: they lack an off-screen model. I'm sure Peter could discuss off-screen models at length (and might in response to this comment), but I'll do my best to explain. An off-screen model (OSM) is a representation of the screen's contents which is designed to be easily queried. It consists primarily of text and images, though it may also contain rectangles, lines, and other shapes. An OSm is built by intercepting low-level graphics operations. This is a gross oversimplification, because the logic required to build and maintain an OSM is very complex. However, I think it will serve as a working definition for the sake of this discussion. With an OSM, a screen reader can provide at least limited access to apps which don't "do the right thing" with regard to accessibility. In particular, custom controls that don't implement any sort of programmatic access can be made accessible, even easy to use, by way of an OSM. These controls are often crucial to the operation of the applications that contain them. If the leading Windows screen readers didn't have an OSM, Windows wouldn't be nearly as accessible as it is. MSAA is woefully inadequate, and even Microsoft apps and OS utilities have important elements that can't be made accessible except through an OSM. I'm not being critical of efforts to improve programmatic access to desktops, toolkits, and applications. However, I don't think we should fool ourselves into thinking that programmatic access is enough. There's already evidence that even in the presence of a rich accessibility API and a rich widget set that implements it, app developers don't always do the right thing. For example, go to http://blindaccessjournal.com/ and enter "Java" in the search box at the top. This is the blog of a blind person who has been unable to perform certain job functions independently due to inaccessible Java applications. Even though he's using a Windows screen reader (JAWS), the screen reader can't build an OSM in this case due to the way Java 2D renders text. Read a few of his entries which discuss the subject, and you can tell that the complete lack of access is frustrating. Should we leave blind users without even a limited level of access while we wait on app developers to get their act together? I think not. I don't expect Sun, Apple, or any of the open-source desktop developers to develop a screen reader with an OSM; that would be expecting too much of them, in my opinion. It would help if clean, documented hooks were provided for watching low-level graphics operations, so adaptive technology developers wouldn't always have to resort to the hacks that are commonplace on Windows. Anyway, I think that Unix and Mac OS X need screen readers that use an off-screen model to provide more complete access. If companies develop such screen readers that are usable, reliable, and affordable, they will earn the business of blind users who need access to these platforms. To be sure, it will take a lot of work, but I believe the work is worthwhile. So is anyone willing to do it?

Posted by Matt Campbell on January 08, 2006 at 09:32 PM PST #