Monday June 28, 2004

Why I'm not attending JavaOne this year

Programatic accessibility - what I suggest is the third generation of techniques used by assistive technologies to obtain information about the applications running on someones system so it can be re-presented to users with disabilities - was first properly done in the Java Accessibility API. So, with the largest Java conference in the world this week in San Francisco, why am I not attending, and presenting, and staffing a booth?

Because, alas, GUADEC is also this week, in Norway. GUADEC is the GNOME Users and Developers European Conference, where many of the key GNOME developers will be meeting eachother and giving presentions. And GNOME is arguably the gravitational center of programatic accessibility currently, with the GNOME Accessibility Project having not only defined the GNOME Accessibility API (a superset of the Java Accessibility API), but implemented it on the stock GNOME GUI libraries (and also on Mozilla, and Evolution). Our own GNOME Accessibility Architect, Bill Haneman, ably assisted by Marc Mulcahy, will be giving a presentation on the current state of accessibility in GNOME 2.6 and GNOME 2.8, and will also talk about some future directions. Also giving an accessibility presentation at GUADEC is Janina Sajka, who will talk about efforts to standardize open source accessibility infrastructure as part of the Free Standards Group.

In addition to GUADEC in Norway, this time of year is also the Libre Software Meeting, this year in Bordeaux. I'm the co-chair of the Accessibility Track, and will also be giving a number of presentations (trip report to come...).

With any luck, next year JavaOne won't conflict with GUADEC or LSM, because two conferences in two weeks is simply not enough! (2004-06-28 03:45:43.0) Permalink

Sunday June 27, 2004

European Summer Tour 2004, part 1

If it is Sunday, it must be Frankfurt Flughofen...

This past week I've been in Heidelberg, visiting with the good folks at BAUM (makers of many fine products for the blind and visually impaired, including the open source Gnopernicus screen reader/magnifier for UNIX and GNU/Linux desktops running GNOME), in advance of the GUADEC conference in Kristiansand. We met in the delightful Schloss Langenzell, just outside of Heidelberg (not to be confused with Schloss Heidelberg). Bill Haneman and Marc Mulcahy, and I had several enjoyable and productive days of engineering meetings with the BAUM Gnopernicus engineers.

In our four days of meetings, we talked about some of the new technologies available in GNOME for working with Bonobo and AT-SPI from a high level language like Python, as well as debugging strategies for dealing with some of the nastier bugs that affect Gnopernicus. We also looked at some of the new magnification technologies that will soon be available in the Xorg server -> DAMAGE, XFIXES, and COMPOSITE. DAMAGE will allow the Gnopernicus magnifier (aka gnome-mag) to know when the source pixels are dirty and respond immediately (it's actually a much more sophisticated interface than that description suggests; but for the low-vision user, this is the key thing). XFIXES fires events when the mouse cursor changes, and also will provide the actual pixels that make up the mouse cursor image, so a magnifier can magnify them. Finally, COMPOSITE will allow a brand new way of doing magnification, and one that promises to be dramatically faster!

All UNIX and Windows full-screen magnifiers (as well as those for for OS/2, and for Macintosh OS 9 and prior) required two sets of bitmaps in order to do magnification: the unmagnified source screen bitmap (stuffed into a RAM buffer), and the physical screen memory bitmap. Virtually all also have a third bitmap: one containing the full magnified image. This means that to do 3x magnification of an entire 1280x1024 screen at 16-bits (minimum for GNOME to look good), 2621440 bytes for the original copy (2.5MB), and another 23592960 bytes (22.5MB) for the 3x magnified view, in addition to the 2.5MB allocated on the video card. And then for every pixel an application draws onto the screen that is also rendered to the magnified view, there is the first draw command (generally to system RAM, not video RAM), a scaled/copy to the 3x magnified view (again generally in system RAM), and then a final copy to the video display. There are additional steps if the magnifier is doing image smoothing (desired by many low vision users), and of course there are things like mouse cursor magnification and full-screen crosshairs.

All of this conspires to slow down the magnifier (especially as stock X isn't particularly good about copying pixels to/from system RAM). It would be nice if we could just stretch a magnified X window on top of the entire screen, but still be able to get the pixels that would be underneath that magnified window. It would also be nice if we could overlay things like crosshairs at will, rather than having to composite them ourselves...

These things, and more, are given to us directly by the new COMPOSITE extension! Now all we need to do is figure out exactly how to best make use of it to support magnification!

We also used the time in Germany to hook up with a number of blind and low vision users/testers, and helped them configure their systems to use Gnopernicus. After seeing experimental magnification with XFIXES and DAMAGE, one low vision Gnopernicus user got quite excited, and told us he was looking forward to retiring his Windows screen magnifier to move full time to GNU/Linux.

And so ends part 1. Next up: Kristiansand (2004-06-27 18:15:52.0) Permalink

Monday June 07, 2004

The CSUN Conference on Technology and Persons with Diabilities

Every year the Center on Disabilities at Cal State University Northridge hosts the Conference on Technology and Persons with Disabilities (though everybody just calls it the "CSUN conference"). What began 19 years ago as a small gathering of products for people with disabilities put on by the Center on Disabilities for the local community has become the premier North American conference on disability technology. I've been attending for the last 12 years, and while a lot of the people (and products) are the same year to year, there is always something new and interesting. And it's nice to see folks in person that you otherwise only get to connect with in phone and e-mail.

As we have for the last few years, much of the Sun Accessibility team was there, including fellow blogger Rich Burridge, Malte Timmermann (who works on StarOffice/OpenOffice.org accessibility), and Marc Mulcahy (from Sun's GNOME accessibility team) among others. This year we again occupied the Washington room, and in addition to normal booth stuff we also had a bunch of "hands on" sessions where folks who are blind, or have low-vision, or have significant mobility impairments could spend an hour on the GNOME 2.6 desktop and play with the Gnopernicus screen reader/magnifier or the GOK Dynamic On-Screen Keyboard. It was a lot of fun to show these apps to people who until now had only ever used a Windows box because there wasn't a UNIX or GUN/Linux option available. Last year when we did this with very early editions of the code, one of the blind participants left my "guided tour", launched the then only vaguely accessible Mozilla, went to the LA Times web site, and started reading an article about the U.S. invading Iraq.

This year the CSUN conference keynote speaker was Vint Cerf gave the keynote talk (the transcript and Vint's slides are both available). Starting with a discussion about Internet technology in general, he nicely sequed to the topic of assistive technology, using his wife Sigrid as an example of how people use these technologies - and perhaps more interestingly, how those technologies change them and affect those around them. He ended his talk emphasizing the importance of standards - in disability technology realm as well as the more mainstream Internet world. I managed to catch up with him, and point him to work by the Accessibility Working Group of the Free Standards Group, which is explicitly trying to develop free accessibility standards for use on UNIX, GNU/Linux, and other computing environments.

My personal conference highlight was getting all of our bazilians of presentation demos working. In particular during the Gnopernicus session we managed an IRC chat session with Bill Haneman in Ireland, Marc Mulcahy (using Gnopernicus in speech and Braille), and myself using a single switch input device and GOK. Marc and I took turns relaying questions from the audience about the weather to Bill, and his replies were projected on the screen behind me, and also spoken aloud by Marc's computer. We also had some fun in the StarOffice accessibility presentation, first installing StarOffice with Gnopernicus speaking the entire installation, and then launching Mozilla to find a Powerpoint slide set on the Microsoft web site, downloading it to automatically open in StarOffice, and finally reading it with speech, Braille, and magnification in Gnopernicus. Finally, it was very nice to have Oracle Corporation join us in the last presentation of the day, in which Mike Pedersen showed off JDevelper 10g working with Gnopernicus under GNOME - a combination he uses every day to do his software enginering job at Oracle.

There were a number of other noteworty events and presentations for me this year:

• Apple Computer announced the Apple "Spoken Interface", their concept of a screen reader, to be included in a future release of OS X. The gave demos for a day in one of the hotel suites. This is only the third screen reader based entirely on an Accessibility programming interface (in this case Apple's Accessibility API (for historical note, the first was an unreleased effort by IBM written entirely in the Java programming languages based on the Java Accessibility API and the second is the Gnopernicus screen reader/magnifier mentioned above.
• Adobe demonstrated their most recent work in PDF accessibility. Their newest PDF viewer, Acrobat 6, shows a ton of good accessibility work. Acrobat 6 Viewer will self-voice and self-enlarge, or alternately it supports a "screen access compatibility mode" for improved interoperability with Windows-based assistive technologies. The Adobe session was packed! This is clearly a hot topic for a lot of folks.
• Owasys demonstrated ther model 22c talking GSM phone for the blind. It is sold only in Europe (for now). It uses Babel InfoVox text-to-speech, and uses Linux as its core OS. A very cool development! The Owasys 22c was developed in close cooperation with O.N.C.E. - the Spanish National Association of the Blind. Suggested retail price is €400.
• Janina Sajka, chair of the Accessibility Working Group of the Free Standards Group gave a talk titled: Developing Accessibility Standards for Free and Open Platforms. I also joined in, and gave a few demos to underscore the points she was making. In summary: there is an effort within the Free Standards Group to develop and promultage a core set of accessibility standards - things that all operating systems/platforms should provide (with an eye toward UNIX and GNU/Linux systems, but by no means exclusive to them). As Janina pointed out in the talk, several years ago she was involved in the quiet, early efforts to develop another open standard for the accessibility community - the DAISY book format for accessible print media; now that is fast becomming the most popular way to disseminate accessible print media and so she has moved on to the next quiet, early effort to develop standards for the next thing that will fast become very popular.
• The wonderful folks at Freeom Machines showed a preview of their upcoming PBS Documentary on assistive technology and its impact on people. I highly recommend this film, which should be coming to a public television station near you sometime later this year.
• Microsoft sponsored an evening reception for Vint Cerf the night before the formal opening of the conference and his keynote. As the sponsor, Microsoft brazenly used the opportunity of a captive audience to promulgate its typical sugary and otherwise unhealthy message and alas as is all too often the case, people simply ate it up.

Finally I want to end this entry with a new phrase I learned: "majority world" (instead of "third world" or the slightly less condescending "developing world"). When you think about it, the most interesting and important aspect of this collection of countries isn't that they are collectively the third region of the world to pass to and through industrial revolution and on to other phases of development; rather the most interesting and important thing is that they represent the vast majority of the population of our planet. Kudos to the folks at CTCNet for turning me on to this one.

(2004-06-07 15:15:11.0) Permalink Comments [2]