Failed Commitments?
Author(s) and publish date
- By:
-
-
Karl Dubost
-
- Published:
Do you remember? it was just three years ago or so. There were parades and brass bands. Many large Web sites were, at long last, making the switch to Web standards. For example, the Web designer Douglas Bowman was announcing the launch of the XHTML, CSS, Validated Wired News Web site. That was great! That was a time of joy and feast. Groups like WASP or MACCAWS, Web professionals who believed in Web standards, who fought for them, were chanting the promises of better days, of a delightful rising sun on the horizon.
And where are we now?
Failed Redesign
Recently, the unparagoned Joe Clark has published an article about companies which have recently decided to redesign their Web sites and somehow failed.
A failed redesign is a Web page created from scratch, or substantially updated, during the era of Web standards that nonetheless ignores or misuses those standards. A failed redesign pretends that valid code and accessibility guidelines do not exist; it pretends that the 21st century is frozen in the amber of the year 1999. It indicates not merely unprofessional Web-development practices but outright incompetence. For if you are producing tag-soup code and using tables for layout in the 21st century, that’s what you are: Incompetent.
What this reminds us, is that redesigning a Web site is not only a question of using XHTML and CSS, but using them in the appropriate way thinking about semantics, accessibility, internationalization, etc. In Jazz, you could put together the best jazz cellist and pianist together and still end in a monstruous cacophony. They have to be in harmony, they have to get their improvisation right. In Web design, you can improvise, you can create, but you have to be in harmony if you don't want to turn your blue note cave into a business catastrophe.
But that's not all! It's not enough to get it right once, to give one enjoyable jam for the audience... You have to give a good performance every. single. night. Oh! Of course a little quack, sometimes might not be such a problem, it will just remind that we are all humans. But a continuous sequence of quacks only shows your impressive ability to impersonate a gaggle of ducks.
Failed Commitments
Jeffrey Zeldman, in his talk about Designing with Web Standards, had a list of Web sites which, a few years ago, had finally decided to go live with a new design, with an exciting redesign. They were examples of companies which had made the switch to Web standards. Trumpets in the sky, we could hear them.
| Web site | Validation |
|---|---|
| Fox SearchLight | Failed |
| K10K | Failed |
| Navy | Valid |
| Inc.com | Failed |
| ESPN | Failed |
| Wired | Failed |
| PGA | Failed |
| Quark Inc. | Failed |
| Note 1: These results have been collected during January 2006, and they may change in the future. Note 2: The homepage is a poor indicator of the validity of a whole Web site. | |
Is there hope?
Never Stops The Music
Making a valid, accessible Web site is only the beginning of the music. Most sites, good or bad, are living things: pages are added, content is updated, reorganized sometimes. It's unlikely that it will be kept as is, clean and polished, unless you put a continous quality process in place. Just keep the music flowing.
Unlike in Jazz, it doesn't really take greatness to always play Web Design right: it's okay to make mistakes, because it's always possible to come back and fix the mistakes. And with a little discipline and a few good tools, it gets really trivial.
The LogValidator is a tool which might help in this task. It's simple, it's extensible to one's own needs and it can be tweaked to meet the internal requirements of any company. How? Just read Making your website valid: a step by step guide and Web Standards Switch.
There are wonderful things in real jazz, the talent for improvisation, the liveliness, the being at one with the audience.
There are wonderful things in real Web design, the many ways to convey information, the being at one with the readers... and the fact that a failure can always be worked upon and fixed.
Wow, nice article I have to admit.
It should be noted that K10K did indeed validate when I tried it a few minutes ago - possibly just a bug in their CSS?
I for one love standards based design, am also seeing so many up-and-coming designers adopting it too. So are there a core of developers that're unwilling to change their ways (often hear the argument that there's no increase in profit or cost benefit)? Or is this slow change due to the time-cost of updating large systems, with little to gain from it?
One thing's definate though: Call these people incompetent and they'll entrench even further. The solution lies in understanding them, discover the reasons for their lack of motivation, then construct arguments meaningful to them.
Douglas Bowman, shurely?!
Joe: yes, we gave him one n too many, fixed. thanks.
We tried being nice and we tried the mechanism of losing a great deal of billable time by publishing free or low-cost guides to Web standards. I see nothing to lose by getting in their faces a little. Time, essentially, is up.
we got our site validated - feels good :)
thanks
Quoting myself:
"possibly just a bug in their CSS?"
Sorry, that should've read:
"possibly just a bug in their CMS?"
Quoting Joe Clark:
"by publishing free or low-cost guides to Web standards."
Correct, and those guides are very good, as a convert I've enjoyed them. But how are these people going to even know guides exist when they don't know to look for them? Moreover, these poor misguided fools don't seem to realise standards exist, even if they do they fail to see what's wrong with their current techniques.
More guidance from the W3C would help too, I know the recommendations state implicitly that tables shouldn't be used for layout, but try telling that to a table evangelist. If you have five minutes read this 'heated debate': It seems what's needed here is some clear guidance from the W3C in the XHTML specifications.
I must also apologise for the forum in that link not validating, the creators of the software won't bear the cost of making their templates validate, another group of people that need convincing!
"Delightful rising sun on the horizon" - rightly said regarding web-standards and its future and I expect it to be happening soon. Log validator helps in finding errors easily, but to make a website semantic following web standard guidelines, requires extensive knowledge on the subject we are into, manual intervention, micro level work, patience and then results.
Too bad. I think I'm going to start a new Web. Because with that approach (that I can't feel is really any change) I believe you have already failed.
Here is my hope. I hope the kids think. I still advocate everyone to (just) think.
While is is disappointing that even years after the effort began to encourage and promote the adoption of standards and associated best practices, there is still a lot of work to be done, but it's important to recognize that this is not an either or proposition. The commitment the identified sites have made to standards and associated practices might not necessarily result in validating pages, but it doubtless results in better pages.
I think the use of "does it validate" as a test of a web site's commitment to standards is by itself essentially useless.
Putting my effort where my mouth is, a few months ago I conducted some reasonably large scale research into common practices in site development at major Australian sites. I developed a four part methodology, as close to objective as I could make it, covering HTML validation, CSS use and validity, use of structural and semantic HTML, and adherence to WCAG1 priority 1 and 2 accessibility guidelines.
The results were on the whole disappointing, I did expect on the whole sites to do better. But they were also encouraging - for example, one of the best ranked sites was the largest company on the Australian Stock Exchange (a resource sector company at that).
It was also encouraging because anecdotally, I suspect almost all sites only three or four years ago to have basically been ranked at 0/20.
I encourage others to adopt the methodology I used to look at other sites. I'd be interested in whether the results are reflected in different geographical areas, to whether significant differences might emerge. Similarly across sectors, and perhaps most importantly over time.
My write up of the results and process is here, where you can even listen to my presentation of the results (and follow along with the slides if you really want) at Web Essentials 05
I personally feel that the validators are too strict. For example the CSS validator throws a warning if you specify a color without a background, but sometimes it's ok to do so. It is also tricky to make sure every single page in a site validates (talking xhtml now), personally I only bother to check the index page.
Perhaps we should accept that certain companies or individuals are TRYING to meet the standards, which is good progress, and instead mock the websites that don't even bother.
Mike: Yes, but for some people this feature of CSS validator is an important one - think about contrast - specyfing just the color and not the background could make the text non-legible.
dusoft: Yeah absolutely, someone could be really stupid and put white text on a white background. But what if I want to change my text from black to #222 and leave the background as default white?
Granted, it is only a warning. The CSS will not fail validation over this, but the warning 'looks bad', when in reality the CSS may be very good - and simply leaves out un-necessary background specifications for optimization.
I propose that the CSS Validator inspect the colors used to determine whether or not it is appropriate to throw a warning.
I agree with Mike Whitehurst. The warning about color contrast is my nemesis in the validator. Links and headlines are the worst, since they often cross over onto different-colored backgrounds.
If we are dedicated to pure validation, a decision has to be made:
1) Use background: transparent; and accept the warnings.
2) Add gratuitous extensions to the style sheet to accommodate different background colors for different areas of a page.
All said, dusoft makes a good point. When working on an enormous website, a unique instance of an element could be overlooked as unreadable. I think the chances are slim, but it could certainly happen. Thus we find ourselves in somewhat of a conundrum.
I think this is a topic that warrants further discussion. Unless, of course, we're taking things a mite too serious. Which is also a possibility.
Nice post, folks. Thanks.
what's even worse?
http://validator.w3.org/check?uri=http%3A%2F%2Fwww.w3.org%2FQA%2F2006%2F01%2Ffailed_commitments.html
I guess someone fixed it dixie, it just validated for me.
I used Tidy to validate my site […]. This works most of the time, except that you need to fix Tidy presentation errors. Tidy is a real time saver.