« previous: My Favorite Books | next: Merry Christmas ! »
Web Accessibility
Comments:
I think it’s a good way for Watchfire to make $300, and I think it’s broken for my site. It also reminds me of the little note on the bottom of every piece of anything released by PSU: "This publication is available in alternative media upon request."
But regarding its vision, I think it’s a good idea that’s somewhat misapplied. Sort of like braille on an ATM. The web is highly visual medium, and to make it more accessible to visually impaired users (which is 98% of the issue), I don’t think it’s a question of running your site against a checklist of rights and wrongs. I think it’s rather a much bigger picture issue that involves rethinking the purpose of any given website, and deciding whether it would be useful to redesign that site from the ground up in a way that completely avoids the use of tables for layout.
Misapplied may be a good word for it. I found the link while reading some pages that I stumbled upon while looking for instructions to access my university server space. After reading all that stuff, I didn’t really have time to read the Watchfire etc opinions about accessibility, so I quickly tried "Bobby" on my "letter" page. The results are rather difficult to interpret unless you’ve spent a lot of time reading up on the subject, in which case you can probably make an accessible site without "Bobby’s" help!
I’m not very good at writing html; I don’t make the effort to use tables and have never bothered to learn to use frames; I always include alternate text for pictures. I kinda assumed that would be good enough. I wonder what it’s like to use a program to read web pages out loud. I have a blind friend, so now I feel like a creep for not having considered this before!
Looking at the code on your page here, it’s hard to imagine how it could be more accessibility-friendly. It’s an ideally-designed page from an accessibility standpoint, in that it does not use tables or even DIV tags to divide the page up into sections (as does my page and many/most other pages, and which is a controversial practice). But I think I can explain a few of "Bobby’s" problems with the page:
"If you use color to convey information, make sure the information is also represented another way." -- I think this is because you’ve specified a page background color and/or link color (inside your BODY tag). The line numbers in the report all correspond to the links on your page (which aren’t even text, they’re images, but Bobby apparently couldn’t figure that out), so by specifying colors for them and for the background, you could hypothetically make them hard to read.
"If an image conveys important information beyond what is in its alternative text, provide an extended description." -- This again refers to the images. This one is just a guess, but maybe since your ALT text is more than a single word or two, Bobby assumes it’s a complex image that warrants an extended description in the text directly following the close of the IMG or A tag. Since it doesn’t find any more text before the next tag (H4), it flags it.
"Use a public text identifier in a DOCTYPE statement." -- If you click on this one, it provides a link to a w3c page full of extremely dry reading on the various SGML types. The idea is that each page should have on its first line (of source code, not of actual content) an identifier to describe the type of document that it is, or more to the point, the specific Document Type Definition to which the page adheres. This is to aid the user-agent (usually a web browser) in determining how to parse and present the information on the page. (I told you it was dry...) Anyway, the bottom line for most HTML-based pages on the internet is that you should just put this as the first line in every HTML document (even before the HTML tag):
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
If your page follows all the rules really well (as yours looks like it might), then you can use "strict" instead of "loose". In any case, you can put that line in place and visit the w3c validator to see whether your page is considered valid HTML under the "loose" or "strict" DTD.
"Is the user made aware that there will be pop-up windows or changes in the active window?" -- This is because you have some of your links set to open a new browser window by specifying target="_blank". Of course, it’s entirely subjective whether that is considered "wrong" and whether it’s necessary to put text next to every such link that says "WARNING: LINK WILL OPEN IN NEW WINDOW. STAND BACK, I SAY!"
"Avoid use of obsolete language features if possible." -- Boy, that sounds so rude! It’s because you use the CENTER tag. Use <P style="text-align: center;"> instead. (Or SPAN instead of P.)
"Make sure header elements are not used only for bold text." -- Ooh, they got you on this one. Bobby is angry because you’ve used header-elements out of order: H4s without any H1s, H2s, or H3s. If all you want is bold and/or bigger text, I guess they want you to use <P style="font-weight: bold; font-size: 150%;"> instead.
"Add a descriptive title to links when needed." -- You can use the TITLE field of the A tag, just like the ALT field of the IMG tag. Do it!, says Bobby.
"Identify the language of the text." -- This is handled by the "EN" in the doctype definition that should appear as the first line of the file.
"Separate adjacent links with more than whitespace." -- This one is because you don’t have text or other non-link images between your images that are links. I strongly disagree with this rule. They are saying that it’s never valid to have two consecutive words that are links to different places, and that’s obviously absurd.
The only thing I see that looks actually "wrong" on your page is that you’ve got a SPAN that starts right after BODY and doesn’t close till the end of the page. It would make more sense just move the style="foo" from the SPAN tag and put it inside the BODY tag. I think this is what’s causing the w3c validator to say that your Ps are all incorrect (which, of course, they aren’t).
To begin, I must stand up and take off my hat to you, because you were patient enough to go through all this junk, which I didn’t even bother to do for my own page!
>"Use a public text identifier in a DOCTYPE statement." -- If you click
>on this one, it provides a link to a w3c page full of extremely dry
>reading on the various SGML types.
Wow, you read all that?? I am in awe.
By the way, I don’t seem to have any text editors that show the line number. I thought I just had a flash of genius and tried opening the html file in vi, but my memory was at fault because vi doesn’t show line numbers either.
Regarding CENTER, H4 and TITLE, well that’s my fault because I don’t know much about html! But it’s interesting that my mistakes are all easier to write than the correct method.
But anyway, what do those mistakes have to do with accessibility? Will they have any negative affect? If not, why is Bobby interested in them?
None of this matters to me at all regarding my own "letter" page. However I’m thinking about it because the company for which I will be working wants me to begin a web page for them. This task is possibly well beyond my ability! Plus there will need to be identical pages in Chinese.
Well, it’s been interesting, and thanks a ton for interpreting it for me!
Hey, glad to help. And no, I didn’t read through all that doctype stuff; I already knew the general idea from having read about it before (in bits and pieces, not the whole huge boring lot of it).
About the line numbers, on most systems that I’ve used, vi is a shortcut to some lame editor like elvis. vim is the real vi, and I think that vim is usually configured to show line numbers by default. If not, just enter the set ruler command, or put that command in your ~/.vimrc file.
I think that CENTER is bad because according to the new rules, tags should only be used as structural elements, not for formatting. All formatting should go into CSS style="foo" statements, which go inside the other tags. I guess that aids accessibility by providing a more consistent approach to document authoring; more consistency and adherence to the rules means easier parseability. And using H4 without H1, 2, or 3 is bad because those are supposed to convey the outline of your document; some user-agents will in some cases just read the HXes, and present those to the user as a summary of the document. So used in that way, it doesn’t make sense for an H4 to appear by itself. And I think TITLE is analagous to ALT; in some cases the user may not be able to view the images, and in some cases the user may not want to / be able to follow links, so a description is helpful.
Reply to this message here:
[ Home – Create Post – Archives – Login – CMS by Encodable ]