It is necessary to critically assess how successful the project was in achieving its goals. The following was asked in the research plan.
"From the research make conclusions how successful the project was in meeting a goal of creating an audio enhanced and engaging web site. Try to answer the following question. Is it possible to a redesign a site in the manner the project advocates on a mass scale? If not why?"
I believe that the project meets the goal of creating audio enhanced and
engaging web site. It is indeed possible, broadly speaking, to use the solutions
I advocated on a mass scale. There are a few qualifications however those
are outline below.
Below is an outline of the features incorporated into the final version of the proposed project. I will briefly comment on each feature and it possible application on a mass scale.
1. Meets W3C HTML 4.01 or XHTML 1.0 standards
This feature required a strict observation on good coding practice it was
simple enough to achieve. The one caveat for designers is not to assume that
commercial web design programs will create or implement the correct code,
often they don't. A flexible approach needs to be adopted. For instance
strict XHMTL does not allow the <embed> tag which is used to display
flash movies. Older browsers however will not display a flash movie without
using the <embed> tag. Problem either use the correct code and restrict
the audience or use the <embed> tag and not be XHTML compliant. Solution,
use an external javascript file to write in the flash movie with both the
<object> and <embed> tags. Simular issues arise when using Cold
Fusion.
2. Page layout using Cascading Style Sheets (CSS)
It was possible to rely on Cascading Style Sheet for page layout rather than use tables. In version 5 or later browsers pages looked pleasing and consistent. However on version 4 browsers problems arose. When a browser only half implements a style the results are often disastrous. Using Cascading Styles Sheets for layout and text formatting compounds the problems. The solution would be to use a "browser sniffer" and load a bullet proof CSS for older browsers.
3. Dynamic Control over visual presentation
The most successful achievement of the project was the coding solution that allowed the user to load different style sheets. It was possible to built and dynamically load a style sheet that catered for users with low vision; a default style that rendered text at 100% and a style that incorporated small fonts and low contrast. Other coding solutions exist to dynamically load CSS that use "cookies" to write and load an external variable. The value of the variable determines which CSS is loaded. The "cookies" solution uses rather verbose coding and an alternative was developed.
Word Up uses a more elegant solution. Javascript is used to rename the browser window name and then reload the page. Upon reloading the value of the browser window determines which CSS is loaded. As long as the browser supports the window.name property then the renaming solution will work. As with page layout a "browser sniffer" would be needed to load simple CSS for older browsers.
4. Dynamic Control over aural presentation
Theoretically Word Up achieved the goal of loading different Aural Style
Sheets (ACSS) based on user preference. It is possible to loading different
ACSS using Cold Fussion or Cookies. Two huge problems exist.
1. Only one browser supports ACSS, Emacspeak.
Emacspeak runs on Linux and only a tiny section of the public use the system.
2. Emacspeak interfaces the Web through the W3 browser, a browser that is
unlikely to recognise the coding required to switch between different ACSS.
It was incorrectly assumed when commencing the project that ACSS being an international coding standard would be implemented on most screen readers and audio browsers. This is not the case. Serious doubt exists whether ACSS will ever be implemented on screen readers. In response the W3C have released a new coding protocol for screen readers, Speech Synthesiser Mark-Up Language (SSML). The new code is less sophisticated than ACSS and time will tell if manufacturers of adaptive technology will even implement the more modest SSML.
Aural rendering of text seems a backwater on the Internet. Whilst visual rendering has progressed to stunning levels since the early nineties "text to voice" output is stuck at an embarrassing and crude level. One cannot help draw the conclusion that if visual impaired users were a more powerful economic and political force in the community then the situation would be vastly different. We would then have web pages that were as interesting to listen as to look at.
5. Flash Narration of each page
The word up site successfully used Flash for as an audio narration solution.
A console was constructed that allowed users to control the audio play back
and navigate through the site. The embedded Flash movies did not require a
mouse to operate and the tab key could be used to access and exit the audio
console. In the creative concept the use of cutting edge technologies or coding
was anticipated.
Due to the nature of the project many cutting edge and novel techniques may be employed. It is the purpose of the project to demonstrate the potential in using advanced techniques to aid accessibility. Therefore it is anticipated that the final project may use techniques and scripting not foreshadowed in the design document.
Since the research project commenced Macromedia released FlashMX, which boosted
improved accessibility features. Labels could be added to buttons that screen
reader could then access and output. Another feature of FlashMX was the ability
to stream and play external mp3 files. This function requires the latest Flash
player to work, r40 or r47.
Although the Word Up site used Flash MX to create streaming audio some issues
remain. There is a slight fault in the way the mp3 is initially loaded into
the browser. Another problem is that an earlier version of the Flash 6 player
(r29) will try and load the same audio file twice and probably fail. To overcome
these issues further development needs to be undertaken.
A version tracker in a dummy flash movie could check the version of the browsers
Flash player. Embedded of the Flash movies throughout the site would be conditional
on having the correct r40 or later player. An alternative would be to convert
the mp3 files into sound only SWF files and stream those files from a control
console. The appearance and function would be identical to the Word Up console
but the application would be accessible to a flash 4 and upwards players,
a much greater audience.
Before considering which option to choose for mass use a greater question needs to be asked. How viable is record audio narration? Depending on length narrating a whole web site could prove to be extremely expensive. I would only be comfortable in advocating recorded narration on small to medium sites where the content is fairly stable.
Daryl Allan Croke : darylallan@netspace.net.au
Electronic Design and Interactive Media
Victoria University 2002.