CSS Reboot participants far from standards-based

While doing research for an upcoming article I was browsing through a bunch of sites submitted to the CSS Reboot Spring 2006. Most of the time, the little HTML Validator icon in the lower right of my status bar did not display the green checkmark that indicates valid markup. Many sites made the poor thing display the yellow warning triangle, and for several sites it even had to bring out the red icon that means there are errors that Tidy cannot automatically fix.

I would have thought that most people who have invested the time to learn CSS well enough to participate in CSS Reboot would also have spent some time learning the basics of HTML. Or, if that proved too difficult, at least run their sites through The W3C Markup Validation Service. Seems like I thought wrong. Shouldn’t valid CSS and HTML be some sort of lowest level of quality control for any site that participates in CSS Reboot? Yes, that is a rhetorical question.

Sean Fraser has gone through all sites participating in the CSS Reboot of spring 2006 to validate their use of HTML and CSS as well as make note of the doctype they use. The details are provided in CSS Reboot as Web Standards Validation Indicator.

In summary: 71.8 percent of CSS Reboot participants use invalid HTML, CSS, or both. 71.8 percent! And that’s on a gallery site meant to showcase web standards-based redesigns:

May 1st 2006 Rebooters simultaneously launched their standards-based redesigns

Not quite.

Posted on June 28, 2006 in Quicklinks, Web Standards


  1. Well the question is not if there were any errors, the question is what the errors were.

    I’ve found for example my blog not validating with the tidy extension because of typos in comments (unclosed quotes) or warnings like “discarding empty LI”, which the commenters are to blame for and I can’t be bothered to fix all of these.

    Validation checking with automated tools is only a means to an end. The nature of the bug is what is important. I had many a complaint about non-encoded ampersands in Google generated code for example.

  2. 71.8% is staggering. When it comes to gallery submissions or design “reboots”, I think that the trend these days is to concentrate on the “wow” factor based on the initial appearance of a website. I have seen a number of occasions where web designers admit to code faults but proclaim that they’ve achieved an end result worth being proud of. If “reboots” required that code be valid, the number of submissions will decrease while maintaining the core importance of such events. Standards based web design.

    I remember doing my best to make sure that the source and stylesheet validated. I think that once code validation becomes more of a habit, it begins to turn into somewhat of an addiction. Well at least that’s my personal opinion as nothing brought more joy than seeing that validator turn green. ;)

  3. June 28, 2006 by Roger Johansson (Author comment)

    Chris: Yes, but unless I am mistaken Sean validated each site’s home page only, and those rarely contain comments. A quick look at the source of some participating sites reveals errors that you can hardly call minor or insignificant.

  4. June 28, 2006 by Eystein Alnæs

    I allways thought that would be a given, even if an unwritten rule. Myself I’d feel embarresed rebooting an unvalid site. But then again, I don’t even have a (live) site :/

    OT - thanks for pointing me to the html validator plugin! That’s up there with the web developer toolbar.

  5. I have actually found myself in the sh*t before due to the tidy plugin for firefox. I was relying on it too heavily whilst developing to pick up validation errors.

    I handed a project over to a client who then ran it through the validator and came back saying there were validation errors. It was quite embarrassing and only took a few minutes to sort but I guess this is a word of warning to those that may get caught out like I did.

    With regards to my reboot, I do have a validation error but as far as I can tell it is a server side issue with character encoding that the guy that hosts my site has been unable to resolve.

    Does this mean I am going to hell or does it mean I have become a statistic and have been tarred with the same brush as others?

    Just curious…

    Cheers :D

  6. If these sites run Google Ads, they will not validate despite being valid.

    I’ve also found that the Tidy plugin is very, let’s say buggy

  7. Wow! I cannot believe that so many of us (meaning CSS Rebooters) had validation errors. and I’m included!

    I just didn’t and I still actually don’t have the time to worry. I’m going with Mike Davidson on this one and… My site does not validate. :-)

  8. Roger, what have you done! CSSReboot is now for sale! ;-)


  9. June 28, 2006 by Roger Johansson (Author comment)

    THe HTML Validator plugin does not perform a true validation, so you need to be aware of that. Always use W3C’s HTML validator.

    Alan: I would think that you have just become a statistic ;-).

    Tim: You can (which I have done) configure the HTML Validator plugin to ignore the domain “googlesyndication.com”. Makes it a lot more useful.

    When you run a site through the W3C validator Google ads will not cause validation errors.

  10. One would think things would validate but I also see CSS Reboot being a marketing avenue for some designers—a way to generate traffic and potentially snare business.

    Just FYI, according to my Tidy FF plug-in, this page doesn’t validate, it has warnings. :)

    I’m guessing it has to do with comments.

  11. June 28, 2006 by d_fens

    I’ve noticed that the w3c actually validates some of the pages anyway even if they have according to html validator errors or character problems.

  12. June 28, 2006 by Roger Johansson (Author comment)

    Pop: No, the HTML Validator plugin displays warnings (which you can turn off) for this site because of the Google Ads. See my earlier comment on that. The extension really should be called something else since it isn’t an HTML validator.

  13. June 28, 2006 by Ryan

    My site doesn’t validate with the Tidy plugin, and thats because I have empty span’s for header and navigation image replacement. As far as I know this is the only way to have text show in place of the IR when images are off turned off.

    However it validates fine with the W3C validator.

  14. June 28, 2006 by Roger Johansson (Author comment)

    Yes, the Tidy plugin and the W3C validator are different things. Sean used the W3C validator to come up with the 71.8 percent, so things would probably look even worse if he had used the Tidy plugin instead.

  15. @ comment 13. Ryan, you could always put text in the spans and then negative indent it -1000em, that way your spans have the same meaning to screen readers as the images have to those who can read it from the image!

    @ comment 2. Derek, I agree, validation does become an addiction.

    I couldn’t release a site if it didn’t validate.

  16. It’s really something that you can’t believe. I really hoped it would have been the ads, but if it went through the validator…

    Shocking news.

  17. Tidy is a Lint Checker but the Validator is based on a SGML Parser/DTD both have their usages which slightly differ in one or two major aspects. Even so it’s a ridiculous amount especially the CSS failings.

  18. CSS Reboot is not May2stReboot : having a no-validating-website on the last is allowed, but I could not imagine having one in the CSS Reboot.

    In my mind, rebooting is not just a graphic experience as some may have wrote it here, but it should also be a matter of semantic and architecture. No validation does not mean no semantic but a poor case of its power.

    And it’s not fairplay for all those you realy play the game and make the effort to produce a valid website (hum, yes, like me ;))…

    Even if I’m disappointed with those percentage, I’ll probably reboot again last year, and I hope that the “validate” status will be take into account.


  19. I wonder what percentage clicked on your advert for SiteGrinder and used that for their reboot!? ;)

  20. The problem is that there isn’t any sort of severe penalty if your site doesn’t validate. Even though I agree with a lot of the arguments in favor of standards, I can’t help but notice that non-valid sites will still look good for the vast majority of viewers. Until a browser like Firefox or IE actually refuses to display an invalid page, this whole standards compliant thing is just going to remain more about bragging rights than anything else.

  21. Chris: Roger’s correct. I only validated home/index pages. And, for the most common HTML errors, see The Most Common HTML Markup Errors on my site. [Elementary Plug: The most common CSS errors will be posted in several days.]

    Alan: No, you wouldn’t be a statistic. There were numeous server-side inconsistencies regarding declared CharSets and DocTypes.

    Aran: You could always petition browser developers to remove the “Quirksmode” fall-back or to show non-compliant sites as “text/plain”. :-)

  22. “Well the question is not if there were any errors, the question is what the errors were.”

    I do wonder about this. If one were writing JavaScript, or C-sharp, the question would certainly be “are there any errors?”, because if there were, your program wouldn’t run (or would crash the user’s computer).

    Broken HTML doesn’t cause such problems for users reading the content - thanks to browser error-handling code. Broken XHTML, if served as such, does, but is faster to render because browsers don’t need error handling code.

    If every site in the world were valid XHTML, then RSS (for example) could, I think, be made redundant via an hRSS microformat. I wonder what current microformat-parsing tools would make of invalid hCards or hCalendars. Not much, I suspect.

    So, I guess for reading in browsers with error handling code, HTML with trivial vallidation errors is no problem. But if we want to move beyond that, then all HTML errors need to go, I think.

  23. June 28, 2006 by Kanashii

    That’s just plain sad…

    And just in case people didn’t know, the 0.8x Version of the HTML Validator Firefox extension does use the exact same parser as the w3c validator and gives the exact same errors so you may want to upgrade to it for more reliable error checking.

    HTML Validator 0.8x Beta

  24. Sean, any chance you would be up for elaborating so that I can look into it again?

    Roger, sorry for using your comments for trying to resolve this but I do genuinely care about my code and if I can get it resolved then I will be a happy man.


  25. June 29, 2006 by Scott

    I beleive that this is why they says “standards-based” rather than “standards-compliant.” Many people feel that coming close to validation is good enough for them and their visitors.

  26. i am getting a few errors from this page? not that it aint validating but it has errors

    get the point you are making and basically you are totally right, these people should feel ashamed.

    maybe this aint relevant but maybe your page should be perfect in html tidy (errors are to do with some external script as I can see it).

  27. June 29, 2006 by Alexander

    Although I think valid websites are important, not for others as much as yourself. Writing valid websites is better for loads of reasons. But I dont know if just pushing them through the validator is enough, there are bugs in the validator, stuff to sort out browser differences get caught (although I usually avoid them sometimes its necessary), new technologies fails (css 3 for example). Also I think that the valid website should be not a goal, but more of a way-pointer. Sometimes its just not productive to make 100% semantical and valid websites,

  28. Just a note. Standards for standards sakes is really becoming annoying. People think once it’s been given the green tick that’s all that’s needed. Far from it - that’s the first step. As has been noted, semantics is the goal, standards the tool to enable it.

    I think a key problem is hype here. Not enough focus is being placed on why validate and the background of exactly why the particular standards came into being. Rather we have a new generation of people who forget semantics, don’t understand the standards and aim to kick the standards goal so that they can add it to their cv - but are far from the reason why one should validate and the approach to take in developing the page framework for a site.

  29. One should make everything valid, and apart from problems with ads and comments that shouldn’t be too hard to achieve. Whether a page/site is valid or not doesn’t tell all that much in itself though.

    One may create “pure HTML nonsense” and make it pass the validator. How about the old evilml that validates as HTML 4.01 Strict? Something to build on?

    I use valid CSS - copied directly from the CSS standards, and the validator chokes on it. That’s ok with me, since the browsers that understands it handles it just perfect :-)

    Adding a few non-valid CSS hacks for IE, so the validator can help me find them if I want to change some styles later on, is probably not “kosher”, but it sure works well.

  30. June 29, 2006 by Roger Johansson (Author comment)

    Kanashii: Thanks, that sounds like a great improvement to the extension!

    Alan: No problem. But I think contacting Sean directly will be faster ;-).

    Scott: Maybe so, but several of the sites I randomly viewed source on were not even close. No trivial mistakes either.

    ben: This page validates when you use the W3C validator. Likewise if you view source with the HTML Validator plugin. The warnings are caused by the external scripts for Google Adsense. Nothing I can do about that unfortunately.

    Alexander: When is using valid and semantic markup not productive? I can’t think of any cases.

    Joel: Who has mentioned standards for standards’ sake here? I haven’t. Sean (who wrote the original article) hasn’t. But I totally agree that it is very annoying to open up a valid site and find nothing but divitis, classitis, a complete lack of semantics, and no attempts at being accessible: Validity does not equal best practices.

  31. June 29, 2006 by Ryan

    @15. Nathan

    But if I do that and images are disabled then they get no text ;)

    With the method I employed the span covers over the text, the span has the background image so when images are disabled they get a text alternative.

  32. Chris Heilmann is correct. Validation is just a means to an end. The end bieng the use of proper (and Valid when possible) code.

  33. Is the lack of validity for the sites posted on the Reboot really that surprising? As Sean points out on his site: “The only showcase gallery site that doesn’t allow failed Markup and/or CSS is css Zen Garden” but at least some of the gallery sites make a request that markup be valid.

    That appears to be missing altogether from CSS Reboot. If the site doesn’t request rebooters to meet any necessary criteria with regards validity then maybe people’s expectations of what the entries should be like are too high.

    It seems to me that CSSReboot is exactly what it says on the tin: a showcase of sites built with CSS, not necessarily a showcase of sites with valid markup.

  34. Seems it’s across the entire web.

    Didn’t John Allsopp come up with similar figures (with similar rates of failure) for at least the Australian sites he analysed, and even then they weren’t aiming for any semblance of web accessibility.

  35. I just compared your findings with our own when we code-reviewed a couple of major German websites, and the results were horrible! It could be summed up as:

    In our elite circles we speak about a “new professionalism,” but HTML in the wild is still a very dark and ugly jungle.

  36. July 19, 2006 by Dissenter

    OK, so 70%+ of the Reboot sites didn’t validate. Fair enough.

    However, 100% of those sites are beautiful, fast-loading, and behave perfectly in every browser I happen to load them in.

    This is just more evidence that supports the idea that validity and web standards only matter to web geeks, not to the real world.

    Perhaps if we quit inventing things to be self-righteous about, the world might take the profession more seriously.

  37. July 19, 2006 by Roger Johansson (Author comment)

    Dissenter: Beautiful is a matter of personal taste, but there are plenty of CSS reboots that don’t look good to me. Fast-loading? Hardly. Some sites are way too slow to be usable.

    Web standards only matter to web geeks? Well, yeah, we’re the ones who need to know about them. I don’t see your point in arguing that web standards aren’t important on a site meant to showcase standards-based web design.

    Validity is not about being self-righteous. It’s about doing your job properly. Out of curiosity, can you give a few examples of CSS Reboot participants that are actually better because they use invalid code than if their author had bothered to learn how to write HTML and CSS?

    The lazy, ignorant, and widespread attitude that validity and best practices don’t matter is what has made people treat the web profession with little or no respect.

  38. I see this crap all over the place, total BS.

    I finally started doing something about it; usually I leave a comment to the site, but in this case, I had to make a screenshot!

    And I’m not talking about the accessibility issues that seem to be the most prevelant either.

    itgarage.com / node / 778

    Glad to see you are green!

  39. August 5, 2006 by Hoopyfrood

    Your observation says far less about the state of web developers than it does the nature of web standards.

    One would be wise to keep in mind that web standards are the target that browsers are trying to chase. It is unrealistic to assume that browsers will “get it just right” in every instance as that assumes that the standards themselves have anticipated every scenario, have been articulated clearly (this is complex stuff) and have fully anticipated development of the craft.

    We build websites to work in browsers. We are the better for the progress that the web standards movement has brought. Yet it takes but a single deviation to effectively invalidate.

    71.8% of entries — entries by professionals which are likely pushing the craft in areas that could not be anticipated a couple of years ago — don’t mesh perfectly with web standards? But look great in the most widely used browsers?

    Sounds reasonable to me.

  40. August 5, 2006 by TomFunk

    Ok, i understand that there are accessability issues around designing things properly, but my god! there is a hell of alot of ‘Snobbery’ around in web design these days. I have only been in the web deesign game for 3 years, and as a normal bloke, dad of 2 from yorkshire i laugh quite alot about how anal some “designers” can get over everything being Just so. A massive percentage of people who use the web are like my wife, as long as they can find what they want, they really dont give a monkeys if it is in a table or a list, frankly they would think it sad to peek at the source and find out. Just my 2 penneth guys.

  41. August 5, 2006 by TomFunk

    By the way, my wife DOES think I am a bit sad since I turned to the “geek side”, and thanks for for bringing sites like this (and others) to the unlearned noobs. My last comment wasn’t a pop at doing things right( i wish i was that good), it was an observation that - until I am clever enough to make sense of it all - will comfort me in my hours of coding frustration. (added in fear of being banned from the intorweb in disgrace)

  42. August 6, 2006 by Roger Johansson (Author comment)

    Hoopyfrood: It may sound reasonable but it isn’t once you start looking at the kinds of errors the sites contain. We’re not talking minor mistakes or insignificant typos here. Besides, the whole point of CSS Reboot is to push standards based design, so requiring participants to be that is definitely not asking too much.

    TomFunk: It isn’t about snobbery. It’s about learning your trade and doing things right so the largest number of people can use the sites you build.

Comments are disabled for this post (read why), but if you have spotted an error or have additional info that you think should be in this post, feel free to contact me.