Thomas Bruederli wrote:
- For visually impaired people, have a selector for font size that
simply changes the css or a high contrast theme. I think it would also be neat to use flite (http://www.speech.cs.cmu.edu/flite/doc/index.html) for text to speech. There would be a link that could be enabled which when clicked would let them save the message as a wav file, then play it back in audio form. I already do this for captchas on my websites. I'm sure most visually impaired people already have software on their computer do this, but they are probably all windows people. this could allow visually impaired people in linux or other less supported OSs to get an audio version of their email. Some web applications sport a logo indicating compliance with http://www.section508.gov/.
It is very hard to create output for screen readers from within an AJAX application. To achieve this we have to make a non-js/non-ajax version of RoundCube first. After that we can start talking about this topic again. In general I suggest to use (or build) another web-based IMAP client that relies on simple and valid (X)HTML output. Same statement as for mobile clients.
I'm just a lurker on this list, but I think I can offer a valuable bit of advice on this topic.
I agree with your sentiment and your instinct on the technical approach. However, keep in mind that many universities (?=and corporations) are strictly bound by section 508, and that providing an alternate client is not a legal option (I think it has something to do with segregation and classes of service as a pathway to discrimination...) So, it's in your best interest to consider accessibility in your development if you want to get a large segment of universities to deploy RoundCube.
We (University of Wisconsin-Madison) are currently using Sun's Messenger Express webmail, which is one of the original AJAX applications (created years before the buzzword came about). It consists entirely of javascript talking to back-end C which communicate data back and forth in javascript data structures. The code violates most w3c standards and it's buggy with many browsers, but somehow Sun has maintained compliance with section 508 (as most screen readers, e.g. JAWS, today are compatible with javascript.)
I wouldn't be surprised if RoundCube already works with most screen readers. It could just be a matter of making a few tweaks here and there to make it fully compliant with section 508.
Jesse Thompson Division of Information Technology University of Wisconsin-Madison