Accessibility when developing for the web

There are two concerns when speaking about accessibility: accessibility for disabled people and accessibility for “disabled” browsers.

I’m talking, on this post, about the second category.

So, is this really something that is needed nowadays?  Even the browser on my mobile phone has JavaScript support (though it doesn’t have Flash support) and can run a complex dynamic website without any special change or anything designed to be used on mobile browsers (be them Blackberries, iPhones or other mobile phones).

Some recommendations are to degrade the design when JavaScript is disabled, but the vast majority of users has no idea how to disable JavaScript on their browsers.  If some IT guy from a company does that, should you really be worried with this person accessing your website?  Of course, you are reducing the market for your product, but how many people are excluded?  Are they people that would “buy” some service from the website?  Is this amount of people worth the extra investment that it takes to design a website to degrade and to be pleasant when used fully dynamically?

It is something different — and getting popular

One of my clients is amazed at how one of our websites is getting attention from people from all over the world.

He’s been mentioning about somebody in Atlanta that has been using the site for two weekends in a row, using the site for two consecutive days last weekend.  “And the website is only in Portuguese!” he says.

We have both created a tool to work with curve fitting using R as the software base for the curve fitting algorithm.  It is still just a concept, but we have already decided on expanding that to include other algorithms besides linear fitting.  And we’re willing to make adjustment to points according to the equation from the curve.

Simple, yet efficient.

Of course, for the next version I’ll sneak in an English translation.  I still haven’t tried implementing any I18Nization with R (I’ll probably need to convert from lazystrings to Unicode before passing the string to R, though…), but it should work fine.

Umbrello from KDE 4

It is much better. It was really unstable and unreliable. Up to the point of me questioning its use.

But, it has been more stable for a while. I don’t know if it was just my environment because I use Umbrello from KDE 4 while running a KDE 3.5 session here, but it crashed all the time it was open without specifying a filename. I even had updated my shortcut to have it opened with a dummy.xmi file so that it was usable.

“Did you file a bug report?” you ask. “Do what I say, not what I do”, I tell you. If the process was more automated or faster, I would have. But it was simply easier to not use the program by the time. I could go without having those diagrams by then and I had defined a target date where I would reinstall Umbrello from KDE 3.5, to have them done.

BUT, it is all right now. Thanks, Umbrello team, for the great tool!