There is life in the "white spaces"

A really thrilling posting from the Google blog and plenty of interesting opinions from those who've linked to it.


FreeAirWaves.com is an initiative to free up the "white spaces" for public use. As explained on the website, once the analogue TV signal is switched off (here in the UK this will be taking place from 2008 till 2012), wireless spectrum will be freed up. This can be utilised to bring the world closer to my idea of omnipresent wi-fi paradise! Minnie from Google called it wi-fi 2.0 (but she preferred the name "wi-fi on steroids"! =)

Just a dream...

Because I use my phone so much, I sometimes dream about it at night too! This time it was one of those dreams that make me wake up in cold sweat.

The first part of my dream was about having my mobile (N95) stolen. Needless to say, a great part of my life is in my phone! Obviously I don't store stuff like PIN codes, but if you get hold of my phone, you can easily get into my inbox, my contacts, my calendar entries, etc. One of the things I got out of the dream is that it's definitely time for a Backup day, and another is that I should invest into anti-theft software for Symbian. It would be also really cool to have some red button control over my phone so that if it gets stolen, I can send a message to my phone, and it destroys all of my data on it.

The second part of my dream was about me looking at my phone screen and being able to track the contacts from my phone book. So, whenever someone from your phone book is within a certain range, the phone detects them, and displays a map with their location. Really cool, but the privacy aspect would probably discourage me from ever using such application.

Brushing it off... hopefully

Today I've been going through my questionnaire using JAWS 7.10. If you go to the page where my questionnaire lives and navigate to Section 1 of it, you will see a subsection called "Contact details" (I'm providing a clickable snapshot below)


I've been trying to ensure that the questionnaire makes sense to those users who can read the page, and those who have to rely on screen reader software to do so. If you're using tab button to navigate running JAWS 7.10 (both on top of IE 7 and Mozilla Firefox 2), you will hear the following:

one to three Contact details, please enter your name edit [tab]
one to three Contact details, please enter your email address edit [tab]
one to three Contact details, please enter your telephone number edit [tab]

This is slightly different to what's visible on the page. The reason being is that "1 to 3 Contact details" is wrapped in the legend tag (resized and taken off the screen to make it invisible using inline CSS) and all the three questions form a fieldset. The screen reader thus repeats whatever's in the legend tag, followed by the question for every text input instance.

For radio button and checkbox questions it's different.



four can we contact you for further information yes radio button not checked one of two [tab]
four can we contact you for further information no radio button not checked two of two

In this case JAWS reads out what's in the legend tag for every radio button option, but it's ignoring the actual question. So, had I not put the question in the legend tag, users would only hear:

yes radio button not checked one of two [tab]
no radio button not checked two of two

This example illustrates the importance of using legend and fieldset tags when designing for diverse user agents.

Hiding Content using CSS

Just as I was on the brink of desperation, I asked Jon about hiding content. It's a simple idea, but it goes a long way. Here's an article on hiding content from view. It can come handy for screen reader accessible design.

Questionnaire again

Well, just a couple of things I ran into

• As I've said - every radio button/checkbox question has to be wrapped in fieldset and legend tags. Yes, it's in the UAAG 1.0 guidelines too. Check out this article on
  Fieldsets, Legends and Screen Readers (Joe kindly pointed me to it)
• This introduces a horizontal scrollbar when increasing the text size in Firefox. But it was useful to find out that in IE it never worked in the first place =)
• Using br tags to break the questions wrapped in legend tags into several lines makes the questions look quite bad
  
• Each question is duplicated (once in the legend, once in where it is placed by create survey's content management system), so watch that Cognitive Load!
• If I only leave one version of a question, the one in the legend tag, quite expectedly, data analysis becomes hell. All the fancy bars and all the stats need separate decyphering
  

Anyways, I'm not giving up and am yet to try a few workarounds (remember, I am working with accessibility gurus, who also happen to be very helpful).

Accessible forms

There's hardly been a day when I didn't bump into some sort of a form or questionnaire. They seem to be ubiquitous, bombarding people from all directions - by post, email, while visiting websites, filing a tax return, applying for a job. I have to admit - they're highly annoying, many badly designed, and time consuming. But what if it's you who needs to use a questionnaire to gather data?

Provide incentives to users, we were told. Make your questions short and simple. Don't lock people into few options, none of which applies to them.

Well, designing accessible forms is even harder. On top of all the usual considerations, you have to approach your forms from the perspective of users with disabilities. Is what you see on the screen the same as a blind user will hear through a screen reader? Is your text adequately formatted for people with dyslexia? Those are just a few questions to ask yourself.

And of course piloting is crucial as ever.

Some thoughts on bandwidth

There is a text-only version of Google Maps! As was rightly pointed out, apart from being accessible, it's also quite handy for the bandwidth-aware users (like me on my mobile).

Re bandwidth - I realised that it works both ways. Heavy pages can be a nuisance for the mobile users due to slow download times and being expensive. Likewise, heavy pages mean more cost in terms of storage and bandwidth for the owners of a website. Here I'm not referring to sites like YouTube, where such requirements are inevitable. But I truly believe that one million byte home page for a news site or an email service is excessive.

Strange as it is in 2008, every megabyte of mobile internet browsing costs me £3 with o2. I can go for the "Unlimited" o2 bolt on for £7.50. This is of course much better, but few people realise that "Unlimited" only means up to 200MB or so, in accordance with o2's "fair use" policy. I'm sure mobile companies wouldn't have been able to get away with such pricing had there been more free public open hotspots (visit Estonia if you want to experience a free wi-fi heaven).

There are some positive changes though. The new bbc website is mobile device friendly. Now I can finally afford to Load All Content =) Let's hope that the others will follow.