Comments on: Podcast 27: Pragmatic accessibility

By: chain saw

chain saw — Tue, 11 Dec 2007 14:11:34 +0000

I think this is still an important topic, and deserves more exposure. Accessibility is becoming a highly important field of design, and it's important for designers to catch on. The problem is that most designers are not disabled themselves, and so only create what they think accessibility is... or what they think will work. It may end up that it is not practical at all for the users of the sites. Certain accessibility issues need to be adressed by the browsers themselves, as well. If a browser could find something that really works, it would probably skyrocket that browser in the browser wars.

By: Ed

Ed — Fri, 17 Mar 2006 16:24:10 +0000

For Mac users, there is now another option for testing on Windows: Install Windows on your Mac. This would mean obtaining a copy of Windows XP, and probably having to reinstall your Mac OS too, but it is something to think about!

By: Luis

Luis — Fri, 17 Mar 2006 15:09:28 +0000

“Right, and fair enough, although there is something to be said for the idea of a single file making changes, which can then be removed when no longer needed.”
True, and if that’s the method someone prefers, that’s fine. But then you can still use server side scripting to do the detection and select among the stylesheets. It would just be in the html instead of the style.
And I think the whole manual approval thing happened because I was submitting from a Palm device this morning instead of my desktop. Maybe something didn’t transfer over correctly. shrug
Oh, one more thing about dynamic css… It is technically more work for the server, but the browsers should cache the stylesheet anyway. So it’s more work once, but after it’s cached, it shouldn’t be an issue. I’m still looking into browser caching php information.

By: Ed

Ed — Fri, 17 Mar 2006 14:46:47 +0000

“It’s just as easy to update your CSS with new browser versions that deviate from expected interpretation. You would have to do that anyway with separate stylesheets to cover the variation, right?”
Right, and fair enough, although there is something to be said for the idea of a single file making changes, which can then be removed when no longer needed.
“manual approval?”
I think as long as you type the Security code, the post should show up.

By: Luis

Luis — Fri, 17 Mar 2006 14:36:22 +0000

I think the problems with Javascript based browser sniffing is that it relies too much on the client and whether or not their browser can do it if Javascript is enabled at all. Turning off Javascript is much easier to do than spoofing/cloaking your user-agent.
It’s just as easy to update your CSS with new browser versions that deviate from expected interpretation. You would have to do that anyway with separate stylesheets to cover the variation, right?
This may be restating some of what I typed in a reply earlier this morning (that didn’t post automatically because of manual approval ?), but I think if the fix for inconsistent browser style interpretation can be resolved at the server level and not rely on the client, you can get more reliable behaviour. Not to mention greater precision in the fix. “You can stitch the wound or amputate the limb”. You can fix the exact piece of css information that’s causing trouble, or you can create a completely different stylesheet that you have to maintain. That seems messy to me (please forgive the associated visual). Also, conditional commenting only detects if the client is IE or not.
The only drawback I can think of from using dynamic css is requiring the web server to support scripting, whether PHP, ASP, etc. But as professional developers, we should all realise the importance of server side scripting. If we maintain our own servers, PHP is free as is Apache server on Windows, OS X and Linux based servers. If we use a third party webhost, there are many inexpensive services that provide scripting as part of the package.
As for object detection, you’re right. I had to fix something of mine that had that issue last night. Something I forgot to test in IE7 for.

By: Ed

Ed — Fri, 17 Mar 2006 13:59:30 +0000

“If you user is spoofing their user agent, do you really care?”
Practically there’s not much you can do in that situation, but isn’t that what brought down the JavaScript browser sniffing?
Do you really want to be updating your CSS every time a new variation on a browser comes out that doesn’t render your site as you wish?
(By the way, I do actually quite like the php css idea, but just asking some questions to get some things clear about it.)
And in terms of object detection – I think if you are detecting an object, you should only do so to see if that object is there, rather than to infer something else about the browser – as shown with the XMLHttpRequest and IE7 situation.

By: Robert Hanson

Robert Hanson — Fri, 17 Mar 2006 12:33:27 +0000

“And there is no way of user-agent-spoofing with the PHP solution?”
If you user is spoofing their user agent, do you really care? Sure, I could use Opera, and have Opera pretend to be IE… but if I don’t get what I expected, I am the one to blame. I don’t think you can expect a developer to develop for that possibility.

By: Ed

Ed — Fri, 17 Mar 2006 11:59:21 +0000

“But with PHP reading the user-agent, you know what browser is being used directly instead of hacking it”
And there is no way of user-agent-spoofing with the PHP solution?
/////
There is a BBC article on an accessibility organisation, the Usability Exchange, which people might want to check out.

By: Paul Boag

Paul Boag — Thu, 16 Mar 2006 14:56:46 +0000

How I love making rude mac comments :) What a reaction I get every single time! Its only because I suffer mac envy ;)

By: davyG

davyG — Thu, 16 Mar 2006 14:37:19 +0000

Hey, mac users aren't stuffed when it comes to testing on "IEWin 6.crappy". Try out http://www.browsercam.com/ - you have to pay for it, but it's a great solution for those of us with the common sense to use macs!!! It's also good for conscientious PC users who want to test cross browser/platform AND!!! you can use their remote access facility to test functionality. I should be getting a commission for this... nevermind.