Accessibility mashups: AxsJAX fun with XKCD Comics

Posted by T.V. Raman, Research Scientist

From time to time, our own T.V. Raman shares his tips on how to use Google from his perspective as a technologist who cannot see -- tips that sighted people, among others, may also find useful.

Earlier this year, I blogged about the potential presented by accessibility mashups with respect to delivering web interfaces that are optimized to a user's special needs. More recently, my office-mate Charles Chen and I blogged about our work on AxsJAX as a framework for leveraging Web-2.0 for injecting accessibility enhancements into web applications.

As we head into the holiday season, we decided it was time to have some fun and generate a few laughs based on what we've worked on during the year. As chance would have it, Randall Munroe, the creator of the XKCD comic strip, visited our Mountain View campus to give an extremely entertaining talk. He even made a reference to blind hacker geeks! So the temptation was too hard to resist. We had to speech-enable his comic strip.

The XKCD comics are highly visual, with a short comment from the author accompanying many of the episodes. Having a detailed written description that is visible to everyone would spoil the comic for the average user; part of the fun is to understand the jokes purely from the sketches. At the same time, notice that indexing and searching online comics runs into the same challenge that blind users face: to be able to locate past episodes, one needs access to textual transcripts that capture the essence of each sketch. To help with the latter, fans of online comics like XKCD have created a search engine devoted to indexing comic strips, replete with full text transcriptions. This is an example of a social Web application where fans can transcribe their favorite comics including XKCD.

In the Web 1.0 world, I would have to pull up an XKCD episode, then go to the site containing the transcripts, and finally find the associated transcript in order to make sense of the comic. But this is exactly where Web 2.0 mashups excel; mashups are all about bringing data from multiple Web sources into a single integrated view. Once we realized this, we were able to AxsJAX the XKCD site with a small amount of code. Now, I can browse to the XKCD comic site, and listen to each episode -- with the underlying AxsJAX-based mashup taking care of the minutiae of retrieving the relevant transcript and integrating it into the comic strip.

This approach leverages all that is powerful about web-based applications:

• Distributed accessibility --- the XKCD author does not need to create the transcripts.
• Transcripts can be integrated from across the web.
• The accessibility enhancements do not spoil the fun for XKCD readers in general.
• And with Open Source self-voicing plugins like Fire Vox, every XKCD user can listen to the strip when desired.

First year of Google WiFi

By Minnie Ingersoll, Chris Sacca & Larry Alder, Alternative Access Team

Our Mountain View WiFi network just celebrated its first anniversary, and we thought you'd appreciate a few data points. The network's 400+ mesh routers cover about 12 square miles and 25,000 homes to serve approximately 15,000 unique users each week month. Since the beginning of 2007, traffic has grown almost 10 percent each month, and the network now handles over 300 gigabytes of data each day, sent to over 100 distinct types of WiFi devices. Virtually the entire city has been taking advantage of the network, with 95 percent of the mesh routers being used on any given day.

Around the globe and across the U. S., many people are still not able to access the online services that are increasingly helpful, if not essential, tools for our daily lives. This is why we're committed to promoting alternative platforms for people to access the web, no matter where you are, what you're doing or what device you're using.

For those who have been following the effort to create a free wireless network in San Francisco, we continue to hope that EarthLink and The City will find a way to enable all San Franciscans to enjoy the free WiFi network they deserve. On a broader scale, we hope that the success of the Mountain View model will encourage others to think creatively about how to address access issues in many other communities.

Update: Corrected usage from "week" to "month."

Overview of our accessible services

Posted by T.V. Raman, Research Scientist

From time to time, our own T.V. Raman shares his tips on how to use Google from his perspective as a technologist who cannot see -- tips that sighted people, among others, may also find useful. - Ed.

We provide a wide variety of services that are mostly accessed with a web browser. People visit Google from a large number of browsers and platforms; in addition, we also understand that every user is special and may have special needs. Accessibility at Google is about making sure that our services work well for all our users -- independent of your needs and abilities at any given time.

Web search, our primary service, has a very simple interface and has always been accessible from a variety of user environments. Newer services that present highly interactive interfaces continue to present accessibility challenges when used with specialized adaptive technologies such as screenreaders. We are committed to finding accessibility solutions that make our services work better for everyone who visits.

Here's a list of our accessibility-related services and a few solutions to some accessibility challenges.

• Web Search: Result pages include headers to delineate logical sections.
  
• Accessible Search: Promotes results that are accessible.
  
• Book Search: Full-text access to public-domain works.
  
• Gmail: A simple yet functional HTML mode that works well with screenreaders.
  
• Gmail Mobile: A lightweight user interface that is also speech-friendly.
  
• Google Maps: Easy-to-use textual directions.
  
• Calendar: A functional, yet speech-friendly user interface.
  
• Audio Captchas: All services that use Google Accounts provide an audio alternative for the visual challenge-response tests that are used to distinguish humans from machines.
  
• Mobile Transcoder: A mobile lens for viewing the web that produces accessible views.
  
• Google Video: Allows uploaded videos to contain captions/subtitles in multiple languages for viewers who are hearing-impaired or unfamiliar with the original language.
  
• Google Talk: IM clients inside a web browser can pose accessibility challenges, but the use of the open Jabber API means that Google users can choose from a variety of Jabber clients, many of which work well with adaptive technologies.
  
• Web APIs: Many Google services offer high-level web APIs that aid in authoring mashups; this provides a means for creating highly customized accessible views.
• 1-800-GOOG-411: Here's an exception to the rule that we deliver most things through a web browser. Our experimental Voice Local Search service lets anyone who can speak into a phone search for a local business by name or category; get connected to the business free of charge; get the details by SMS if you’re using a mobile phone. (Just say "text message".)

Finally, many Google services such as Google Scholar, Google News, Blogger and Google Product Search work out of the box. While today's screenreaders can hit some bumps on the road when using more advanced features in these products, these web interfaces degrade gracefully to provide a functional interface.

If any of this interests you, we invite you to participate in our user community. Please tell us what works well, share your own tips on using Google services, and make sure to tell us what could be made even better.

Update: Added info on 1-800-GOOG-411.

Ever more books to read

Posted by T.V. Raman, Research Scientist

From time to time, our own T.V. Raman shares his tips on how to use Google from his perspective as a technologist who cannot see -- tips that sighted people, among others, may also find useful.

As a book lover, I couldn't have been more excited by the advent of electronic books in the early 90s. And with the rise of the Internet, the possibility of being able to discover books online was really exciting.

I work on a project at Google called Google Accessible Search, which helps promote results that are more accessible to visually impaired users. Building on that work is today's release of accessible public domain works through Google Book Search. It's opening up hundreds of thousands of books to people who use adaptive technologies such as speech output, screen readers, and Braille displays.

I'm extremely enthusiastic about many of our efforts at Google, which ultimately have the potential to make the experience of accessing information for visually impaired users just as easy as it is for those with sight. My reading used to be determined by what was available in a form I could read. But today it's a question of using Google effectively so that I can find the right thing to read. Today's Book Search launch is an excellent example of how technology is truly democratizing access to the world's information.

If you have feedback about Google's accessibility services, join our group to share tips on what works well and what could be made better.

Google and open source OCR

Posted by T.V. Raman, Research Scientist

From time to time, our own T.V. Raman shares his tips on how to use Google from his perspective as a technologist who cannot see -- tips that sighted people, among others, may also find useful. - Ed.

As someone who cannot see, I prefer to live in a mostly paperless world. This means ruthlessly turning every piece of paper that enters my life into a set of bits that I can process digitally. I scan in everything. Until now, I have relied on commercial OCR packages to convert these images into readable text. OCR is perhaps one of the areas where the benefits of Moore's Law are most evident; today, OCR can do remarkably well when handed a page image. Until now, my only dissatisfaction with the status quo in this area has been that commercial OCR engines afford me little flexibility with respect to training them to do better on documents that are specific to me.

The advent of our own open source OCR initiative, OCRopus (source code: Ocropus Sources) is a welcome change in this regard. I introduced support for OCRopus in Emacspeak recently, and the HTML output this produces compares favorably with output from commercial OCR engines, provided you place the page at the right orientation on the scanner. OCRopus' extensibility, and the ability to express the OCR as a structured HTML document makes it an ideal starting point for producing rich spoken output. The possibilities are enormous for people being able to collectively train, customize and improve an OCR engine.