5 Mar 2009 4 pm eastern

Web Standards Test: Top 100 Sites

While working on the third edition of Designing With Web Standards, I decided to visit Alexa’s Top 100 US Sites to see how many of the top 100 use valid markup, how many nearly validate (i.e. would validate if not for an error or two), and which DOCTYPEs predominate. Even with a fistful of porn sites in the mix, it was dull work: click a link, load the home page, run a validation bookmarklet, record the result.

I had no expectations. I made no assumptions. I just clicked and tested.

Such tests tell us little

I make no claims about what I found. If all the home pages of the top 100 sites were valid, it would not mean that the pages beneath the home page level were valid, nor would it prove that the sites were authored semantically. (An HTML 4.0 table layout with no semantics can validate; so can a site composed entirely of non-semantic divs with presentational labels.)

Validation is not the be-all of standards-based design; it merely indicates that the markup, whatever its semantic quality may be, complies with the requirements of a particular standard. Conversely, lack of validation does not prove lack of interest in web standards: ads and other third-party content can wreck a once-valid template, as can later third-party development work.

Moreover, nothing causal or predictive can be determined from these results. If 25% of the top 100 sites validated in my test, it would not mean that 25% of all sites on the web validate.

And I got nothing like 25%.

Enough disclaimers. On with the test.

Seven percent validate

On this day, in this test, seven out of 100 “top US” sites validated:

  1. MSN (#7 in Alexa’s list) validates as XHTML 1.0 Strict. Who’d a thunk it? (Validation link)
  2. Craigslist (#10) validates as HTML 4.01 Transitional. I’ll buy that! (Validation link)
  3. WordPress (#22) validates as XHTML 1.0 Transitional. The power of the press, baby! (Validation link)
  4. Time Warner RoadRunner (#39) validates as XHTML 1.0 Transitional. Meep-Meep! (Validation link)
  5. BBC Newsline Ticker (#50) validates as XHTML 1.0 Strict. Cheers, mates! (Validation link)
  6. The US Internal Revenue Service (#58) validates as HTML 4.01 Transitional. Our tax dollars at work! (Validation link)
  7. TinyPic (#73) (“Free Image Hosting”), coded by ZURB, validates as XHTML 1.0 Transitional. (Validation link)

Also-rans (one or two errors)

  1. Wikipedia (#8) almost validates as XHTML 1.0 Strict (two errors).
  2. Apple (#29) almost validates as HTML 4.01 Transitional (two errors).
  3. Linkedin.com (#45) almost validates as HTML 4.01 Transitional (one error).
  4. AWeber Communications (#83) almost validates as XHTML 1.0 Transitional (one error: an onClick element)

Suis generis

The Pirate Bay (#68), “the world’s largest BitTorrent tracker,” goes in and out of validation. When it validates, it’s a beautiful thing, and it belongs on the list. But when it goes out of validation, it can quickly stack up ten errors or more. (Validation Link)

No-shows

Google (#1) does not validate or declare a DOCTYPE.

Yahoo (#2) does not validate or declare a DOCTYPE.

YouTube (#3) does not validate but at least declares that it is HTML 4.01 Transitional. Progress!

A surprising number of sites that do not come close to validating declare a DOCTYPE of XHTML 1.0 Strict. For instance, Twitter (#93) is authored in XHTML 1.0 Strict, although it contains seven errors.

This preference for Strict among non-validating sites suggests that at one point these sites were made over by standards-aware developers; but that any standards improvements made to these sites were lost by subsequent developers. (It doesn’t prove this; it merely suggests.) Another possibility is that some developers use tools that are more standards-aware than they are. (For instance, a developer with little to no knowledge of web standards might use a tool that defaults to the XHTML 1.0 Strict DOCTYPE.)

Some sites that used to validate (such as Blogger.com, previously designed by Douglas Bowman, and Reference.com, previously designed by Happy Cog) no longer do so; maintaining standards or design compliance may not have been important to new owners or new directors.

[tags]validation, webstandards, alexa, test[/tags]

Filed under: Design, development, HTML, Web Design, Web Standards, Websites, XHTML

112 Responses to “Web Standards Test: Top 100 Sites”

  1. Andrew Mager said on

    So it looks like validation doesn’t matter yet. But I hope one day it will.

  2. Michel said on

    thepiratebay.org also validates? :-)

    (Well, it’s not a US website, but still, I believe it is in the World Top 100, anyway! :-)

    Interesting results you got… Most interesting is that Google still shows no interests in any doctypes or validators at all (but still is close to No.1) and MSN validates perfectly (who said M$ doesn’t care about Web Standards??)!

  3. Scott said on

    I wouldn’t say that they don’t “matter yet,” but I think it’s worth noting that thanks the amazing progress of the web standards movement a lot of popular sites are making the effort to make sure their sites validate.

  4. Jesse Bilsten said on

    As much as I want to hate MSN I can’t help but be thankful they’re pushing the pack.

  5. Chris Jennings said on

    Its interesting that companies like Google and Yahoo who have offices full of brilliant minds have so little regard for standards in practice… no Doctype? really?

  6. Jeffrey Zeldman said on

    Michel:

    thepiratebay.org also validates? :-)

    The Pirate Bay was in my original “eight valid” sites. But then it stopped validating while I was writing this piece, so I removed it. The site appears to go in and out of validation.

    I’ll add it to the also-rans.

  7. Chris Casciano said on

    It would be quite interesting to chart the types or sources of errors encountered, though this would clearly take a great deal more effort. As you’ve stated “ads and other third-party content can wreck a once-valid template” and I’ve seen this happen far too often myself. and just checked one site I’ve recently worked on that I knew wouldn’t validate – and not by lack of trying – and sure enough the 3 screens can all be traced back to 3 chunks of third party code.

  8. Mikko Lauhakari said on

    did you find many/any site in HTML5 ??
    I discovered when Safari4 came out that the page for Safari @ Apple.com has a doctype of HTML5.
    Or at least it does if you run the W3C validator, if you look at the source code it only states “” …. ?!

  9. John Lascurettes said on

    Some sites that used to be validate (such as Blogger.com, previously designed by Douglas Bowman, and Reference.com, previously designed by Happy Cog) no longer do so;

    Bought by Google, so it should come as no surprise.

  10. jason said on

    Google makes no secret about their page not validating and have stated in the past that they are motivated by one single thing when it comes to their site – speed. If they can leave one line off that takes a thousandth of a second to load, they will.

  11. Stephen said on

    This validator is made of fail, sorry. A lot of these sites are filtering on user agent strings and feeding a very simple page to it. For example, yahoo.com sends a super-simple page to unknown/pre ie5.5 browsers. This page has no doctype deliberately to throw browsers into quirks mode. Yahoo.com actually has a doctype of HTML 4.01 strict. It doesn’t pass validation, but the markup is pretty good.

  12. Jeffrey Zeldman said on

    Bought by Google, so it should come as no surprise.

    John:

    Doug designed Blogger.com at the behest of Google (and in partnership with Adaptive Path) when Google bought Blogger. I was also involved, along with a handful of talented standards-based designers. Read more about Google’s Blogger redesign.

    So Google hired Doug and Adapative Path, and Doug and those of us who worked with him made standards compliance a priority. Google still owns Blogger, but whoever is in charge of it now either doesn’t know, or doesn’t care.

    did you find many/any site in HTML5 ??

    Mikko:

    Not in the Top 100 US Sites. But, uh, An Event Apart is authored in HTML 5. (More about that.)

    I can’t see the big corporate sites diving in yet, especially since HTML 5 is still very much in flux.

  13. Stephen said on

    The blogger.com that Jeffery validated is a 301 error page. Basically this validator is just loading whatever curl would load. Blogger.com DOES include a doctype.

  14. Matthew R. Ventre said on

    @Chris_Jennings,

    If I recall accurately, part of the explanation for Google’s missing doctype among other markup (inspect the source for missing quotation marks around attribute values) is a matter of scale, as explained by Doug Bowman at An Event Apart last June. When a site sustains the amount of traffic that these two behemoths do, those mere bytes of markup add up to gigabytes of throughput. While web standards are essential in my opinion, Google is under a different set of constraints when it comes to implementing them.

  15. Neal G said on

    Why on earth would a site not use a DOCTYPE and go into quirks mode? Certainly Yahoo! or Google don’t care about browsers such as IE5 etc.

    I can probably tell you how many XHTML sites are being served with the correct MIME type….0, which pretty much defeats the purpose of even using XHTML.

  16. Stephen said on

    @Neal I work at yahoo. I can’t speak for the company, but I can tell you that we serve the quirksmode page ONLY to ancient browsers. We want to make sure that every and any user gets something, and that the something they get is not broken. Go to http://www.yahoo.com. Look at the source. There is a doctype there–unless you are using an ancient browser. (try changing your doctype). Anyway, I don’t think this validator is valid, so to speak. Jeffery and commenters here are making some comments about google and yahoo that are based on a false premise.

    Blogger.com has only 16 minor errors from the w3c validator and does include a doctype. The 301 error page that Jeffery tried to validate does not include a doctype, but its hardly fair to say “maintaining standards or design compliance may not have been important to new owners or new directors.” where clearly an effort was made to do so.

  17. Top 100 US Websites im Standardstest | Webzeugkoffer Webdesign said on

    [...] Web Standards Test: Top 100 Sites [...]

  18. Jeffrey Zeldman said on

    Stephen, I validated Blogger.com which redirects to Blogger.com/home (i.e. to my Blogger account). I did not validate a 303 error page.

    Blogger.com has only 16 minor errors

    The test was to see which sites validated, not which sites had 5 or 10 or 16 errors. It wasn’t a deep analysis of the quality of the semantics.

    If a site had 16 errors, or 5, or 500, I didn’t spend hours viewing source to determine whether, in my opinion, after careful analysis, there was evidence that the developers actually did care about web standards.

    As disclaimed at the top, it was a simple pass/fail test on validation.

    I’m glad to hear that Yahoo cares about web standards. If a different validation service shows that Yahoo’s home page is valid, please share the URL.

  19. Chris Heilmann said on

    Oh blast from the past. Validators are a nice idea, but if you want to test validity then test with a browser and see what sites give you back.

    This methodology is flawed as every web site on that scale does user agent testing and delivers the right experience for the current user agent. If your user agent is an unknown validator – you get bad code back.

  20. Jin said on

    When the new http://www.whitehouse.gov first launched, it was praised for passing the validation (very rare for a .gov site). These days, you can tell that they have people of different skill level to update the homepage.

  21. Federico Zoufaly said on

    Nice post. I did a similar survey on the Fortune 500 home pages a while ago and got the same type of results: http://www.aggiorno.com/blog/post/Benchmarking-DOCTYPE-validation-in-Fortune-500s-Web-sites.aspx

    Programming following standards should be no harder that not using them. The advantages are obvious (maybe not??) so why this is not happening faster?

    We the advent of IE8 Microsoft is pushing in the right direction. I hope it is not too late.

  22. Lori Packer said on

    Every year for the HighEdWeb conference, I do a similar survey of the homepages of Educause members (over 1,000 schools) and compare it with homepage of Fortune 500 companies. The results in a nutshell:

    88% of college homapages declare a DOCTYPE, compared to 79% of the Fortune 500.

    The most popular DOCTYPE for both is XHTML 1.0 Transitional (54% of college, 46% of Fortune 500)

    Of those sites that declare a DOCTYPE, only 11% of college homepages validate. This is the same as 2007, and up slightly (10%) from 2006. The Fortune 500 is even sadder; only 3% of sites with a DOCTYPE validate to that DOCTYPE (2% in 2007, and 3% in 2006).

    I attributed the difference in validation rates between the higher ed world and the corporate world to the theory that while the Fortune 500 may have more web development resources at its disposal, higher ed institutions may be more likely to have a standards evangelist in their midst.

    The 2007 survey is online, but my apologies, it looks like the 2008 survey wasn’t posted.

  23. Klaus said on

    Like Stephen, I work at Yahoo! as well. Whilst working at their London office (May 2007 until November 2008) I had the pleasure to work with the most skilled and knowledgeable front end web developers I have ever had the pleasure to work with. Unfortunately, like Steve says, we are individuals and can only speak for our work, but believe me, many here care about Standards.

    @Jeffrey: I agree with Stephen, I believe that if you would have used the standard W3C validator, it would not have complained about a missing DOCTYPE.

    In the midst of all those discussions and nearly flame wars, shouldn’t we look at them this way:
    So we have Standards, right? Put in place because we did not want to mess around and care about proprietary stuff developed by browser vendors but rather have a common ground agreed upon and to work with. So all that movement in the past, that Jeffrey and many other folks initiated, is a good thing, right? So why not do everything possible to stick to it?

    Correct me if I am wrong but standardizing stuff, in probably every other industry, brought nothing but good. I am no historian, so this is just a guess.

    IMO, it is also a different story whether you say that as soon as you put 3rd party code on a page, like adverts, the site is not going to validate, no matter how much time and effort you had put in the surrounding page. That I see as quite acceptable, cause you most likely can’t control the markup that someone else produced to include on your page.

    But doing something like putting in custom attributes in HTML elements because you either think: a. it is smart to do so, b. you think you are smarter than the W3C or c. you are just lazy and don’t do it the standards way, just defeats the reason why standards exist in the first place. If everybody thinks they are more clever than anybody else and pushes for “their own” standard, regardless of what the W3C standards are, then we are slowly just going back to where we were about 10 years ago.

  24. Erik said on

    @Klaus: Yahoo! source indicates (at least in Firefox 3) a DOCTYPE of HTML 4.01 Strict. But the W3C validator claims no DOCTYPE and so tested against HTML 4.01 Transitional.

  25. Rob Landry said on

    After Recovery.gov launched the other day, I ran it through the w3c validator for kicks and found it had 148 errors. I noticed that it doesn’t use semantic markup (no h1, h2, h3…) . It’s a Drupal site (not sure if that means anything), and I’m sure it was thrown together pretty quickly to hit a high profile deadline, to be fair to their development team.

    I also had a blind colleague navigate the site and he said it seems fairly good to him, even w/o the headings or a skipnav.

    While I agree with @Andrew Mager in that validation matters and designers should validate as part of their development process to reduce rendering bugs, I also agree with @Chris Heilmann in that the most important thing is that it renders in the browser, and that site owners will consider that the bottom line in terms of code quality.

    Interesting piece. Something I’d wondered about but never took the time to check.

  26. Josh Stodola said on

    Google’s blatant disregard for a DOCTYPE makes me question its importance (in general). Seriously, if was truly important, how could Google be so successful in the midst of omission?

  27. Narshada said on

    very web site on that scale does user agent testing and delivers the right experience for the current user agent. If your user agent is an unknown validator – you get bad code back.

    Sorry but wasn’t part of the point of web standards to eliminate sending different pages to different user agents and browser sniffing etc? Wasn’t the credo “Write once, publish everywhere”? Yeah I know IE still lags behind everyone else and is the main reason virtually no one serves XHTML with the correct MIME type, but this is stuff that should have been adopted universally years ago, with the possible exception of hobbyists and crazies and sorry, but I think behemoths like Google and Yahoo should be leading that charge by example. To say that they’re under a different set of constraints excuses them from doing the right thing, even if it’s true.

  28. John Lascurettes said on

    Sorry Jeffrey, I forgot about the sponsorship originally by Google for valid markup on Blogger. At the time, it appeared from the outside like the big ship was starting to turn at Google; however, to this day, they have made few significant improvements to the main Google assets. Without proper systemic adoption of web standards, it’s no surprise that Happy Cog’s, Doug’s and others’ original efforts have eventually been forgotten and dropped.

  29. Don Ulrich said on

    Does the information inside the document require it to be archived or cataloged or is it merely Googles default page? I guess maybe the line to cross is just how temporal
    the data is. So calling a db and returning it as HTML has a different temporal state than a HTML document that is the source of the data?

  30. Mark said on

    Google web search results pages include the HTML5 doctype. They don’t come close to validating, but my understanding is that the HTML doctype is the smallest (byte-wise) way to trigger absolutely-no-quirks mode in all browsers.

  31. Spiesser? | Code Candies said on

    [...] Wenn das alleinige validieren doch keine Aussage über die Qualität des Codes bringt, warum wird es dann immer wieder gemacht? [...]

  32. Todd said on

    In agreement with both Chris Heilmann and most everyone about validating, yes I want the site to look good in every browser and yet have clean code. So when is there going to be one service, and only one that will be the validator of choice and do the job correctly?

    I think it’s some of these other sites, the smaller ones I have seen and wrote of lately that need more help. Google may not burp out a DOCTYPE, but when I see a site, done by a “professional/pro firm” that not only omits a DOCTYPE, but scores flaw after flaw code-wise, it makes me turn my attention to the little sites. Maybe they’ll listen better than a mega-giant that doesn’t mind that their code isn’t to what the percentage of the masses wants.

  33. six03 » Web Standards Test: Top 100 Sites - Jeffrey Zeldman said on

    [...] Zeldman has an interesting article over at The Daily Report about the Alexa Top 100 Websites in the US and how they do/or do not [...]

  34. Validation is not the be-all of standards-based design ~ Sponge Project Blog ~ Web design in Tewkesbury said on

    [...] quoting Jeffrey Zeldman who’s latest post provides a useful reminder that semantic markup and web standards takes [...]

  35. Jeffrey Zeldman said on

    Chris:

    Respectfully, if you sniff before serving a DOCTYPE, you’re doing it wrong. Sites with millions of page views per hour operate under different constraints, and may not have the luxury to say no to any user agent, but they can still write valid markup against a DOCTYPE.

    Stephen and Klaus:

    I have friends and colleagues at Yahoo; they are some of the smartest and most talented designers and developers I know; many care about web standards. None of that is at issue.

    This was not a test of which companies have employees who care about web standards. As disclaimed at the beginning of my post, this was a test to see how many and which of Alexa’s Top 100 US Sites pass a simple, standard, binary (pass/fail) validation test. I did not and do not hold any company or its employees up to ridicule, nor is any reader encouraged to draw personal inferences from these humble statistics.

    These test results are simply one metric of the rate of standards adoption among large sites. They are not the only metric, and they are not the most important one, but I was not able to conduct a week-long interview of the employees of each of the Top 100 sites to evaluate their knowledge of and commitment to standards-based design. Sometimes you have access to a state-of-the-art meteorological lab, other times you just sniff the air and say, “It feels like rain.”

  36. Yellow SEO said on

    Its a hard balance between wanting to be compliant and wanting the web pages to work on almost all browsers. I think that once browsers are more compliant so will most major websites.

  37. Mark Wunsch said on

    Jeff Atwood (of Coding Horror) also just recently wrote a post about HTML Validation:
    http://www.codinghorror.com/blog/archives/001234.html

    The gist of his advice:

    1. Validate your HTML. Know what it means to have valid HTML markup. Understand the tooling. More information is always better than less information. Why fly blind?
    2. Nobody cares if your HTML is valid. Except you. If you want to. Don’t think for a second that producing perfectly valid HTML is more important than running your website, delivering features that delight your users, or getting the job done.

  38. Devon Persing said on

    May I ask what validation tool you used? Thanks for doing this and writing it up.

  39. Jeffrey Zeldman said on

    As the provided validation links suggest, I used the validation service at HTMLHelp.com.

  40. Devon Persing said on

    Oh jeez. I need more coffee. Thank you for humoring me.

  41. Leannekera said on

    Hi Jeffrey,

    First I am sorry to go off topic but this has worried me for some time…
    As an avid worshiper of your work I am a little surprised at your choice of header and link tags… The outlined text style you have used and colour hurt my eyes and make me squint to read. This is surly not your usual godlike ability for bringing the best in user experience?

    Please delete this post and fix so I may continue worshiping your infinite knowledge :)

  42. François Nonnenmacher said on

    maintaining standards or design compliance may not have been important to new owners or new directors.

    You’d be pleased to note that this site still validates, four years after we worked on its redesign. If you remember, I had that in mind before we started, and it cannot be emphasized enough that one needs to embed quality right within the tools that are used everyday by innocent editors (most importantly the content management system) in order to prevent things from wrecking.

  43. Damon Cool said on

    The typical answer at my company is “we don’t have time for that”.
    Sadly we’re still stuck in 1999.

  44. serdar said on

    It’s very interesting that Microsoft chooses to use XHTML on msn.com; a language which is not (and won’t be) supported in Internet Explorer.

  45. Jeffrey Zeldman said on

    Internet Explorer supports XHTML; it doesn’t support it with the correct *MIME type*, but it interprets XHTML markup correctly.

  46. Jason Friesen said on

    @François Nonnenmacher: You’re right; keeping the site valid is a neverending battle. My kingdom for an idiot-proof text editor! XStandard.com comes awfully darned close.

    For what it’s worth, Selkirk College in British Columbia usually validates XHTML 1.0 Strict. But it’s a battle to keep it there.

  47. Joseph Lowery said on

    You inspired me, Jeffrey, to look at the top sites with another microscope – the suspiciously eponymous Jeffrey Zeldman’s Web Standards Advisor. I posted my results on the company blog. It’ll be interesting to see how long it takes for the big ships to turn.

    Btw, thanks for the opportunity to use “eponymous” in a sentence!

  48. Joseph Lowery said on

    My bad. Find those results here.

  49. serdar said on

    @Jeffrey Zeldman; if you serve XHTML as text/html it is not parsed as XHTML but in quirks mode. Just as bad as not declaring a doctype at all. So Explorer does not support XHTML but just ignores ‘errors’ in tag soup.

    Why marking up a document in XHTML if you would serve it as tag soup anyway?

  50. Jeffrey Zeldman said on

    Joseph:

    Can you share the URL of your post?

  51. Joseph Lowery said on

    Messages passing in the late afternoon… it’s in my previous post, Jeffrey

  52. Casey Brown said on

    Thanks for the note about the Wikipedia.org portal. :-) It should be fixed now.

  53. Faruk Ates said on

    While I’m a continued firm believer that validation is nothing more than a tool to find mistakes in your site/code, it’s interesting to see these results. Interesting, but as you, Jeffrey, already pointed out, these results may suggest some things but they prove nothing.

    I’d be most curious to see how the Zeldman Web Standards Advisor would report on them, though that requires a hell of a lot more setup work.

  54. Nathanael Ritz said on

    Actually, Google doesn’t hate standards. They are just slow at getting to the full implementation of them. For example, the section they have about their website survey validates as HTML 5.

    http://code.google.com/webstats/index.html (link goes to W3C validation results) is the example I have for you. Their intro spiel to the survey is interesting. So is the fact that the page validates. And obviously, so is the actual survey results.

    Cheers!

  55. Jeffrey Zeldman said on

    Joseph:

    Your summary is interesting. More details, please!

  56. soulscratch said on

    How ironic that this is a validity test, whereas zeldman.com fails validation with 7 errors, and when sent as real XHTML it results in the yellow screen of death:

    XML Parsing Error: mismatched tag. Expected: .
    Location: http://localhost/zeldman.html
    Line Number 385, Column 79:Besides emceeing with Eric Meyer, I’ll cover the following topics:

  57. soulscratch said on

    Of course I didn’t mean to link to my localhost incase anyone should point that out, forgot that it would try to convert it to a link, hah! It was a copy of the source code on the homepage as of 3/07/09 10:00 AM EST.

  58. soulscratch said on

    “Internet Explorer supports XHTML; it doesn’t support it with the correct *MIME type*, but it interprets XHTML markup correctly.”

    Pretty much every graphical user agent is setup to be lenient to handle malformed HTML, which includes XHTML sent as text/html even if it is error-full.

  59. Jeffrey Zeldman said on

    How ironic is it that this site validates! That is so *ironic*!

    Thanks for pointing out the error.

    P.S. The XHTML DOCTYPE throws Internet Explorer (and all other modern browsers) into STANDARDS mode, *not* into Quirks mode. Please get your facts straight while you’re ranting.

    P.P.S. Fringe rant idiocy turns everyone off to the whole idea of web standards.

  60. soulscratch said on

    “How ironic is it that this site validates! That is so *ironic*!”

    At the time of my post it was invalid, but you updated during that time, removing an incorrect p end tag in the paragraph with Eric Meyer and replaced it with a dd end tag.

    Nice cover-up though.

    “The XHTML DOCTYPE throws Internet Explorer (and all other modern browsers) into STANDARDS mode, *not* into Quirks mode. Please get your facts straight while you’re ranting.”

    When did I _ever_ mention quirks mode? I was referring to your invalid HTML which had a DTD ( standards mode ) at the time of the post, and because browsers are lenient that missing p tag when it should have been a dd was not broken visually as the browser accounted for it.

  61. soulscratch said on

    And this was the markup used at the time: http://rafb.net/p/9JV9cL70.html
    So it may have been that you edited exactly after I copied that bit.

  62. Jeffrey Zeldman said on

    You win! I corrected an error in markup.

    I generally validate my posts but I neglected to validate the one in which the nesting error occurred.

    As for the Quirks Mode, it was another poster who incorrectly said…

    if you serve XHTML as text/html it is not parsed as XHTML but in quirks mode.

    I lumped you guys together in my response. Sorry. He’s the guy who had his facts wrong. You had your facts right, but you have no social grace. Why be rude about anything, let alone about such small matters?

    because browsers are lenient that missing p tag when it should have been a dd was not broken visually as the browser accounted for it.

    It actually did affect the layout, however, because in my CSS, a P inside a DD is different than a P outside such a container.

    But, yes, browsers are lenient, always have been. And many people believe that lenience has retarded the battle for web standards. See, for instance, J. David Eisenberg’s Forgiving Browsers Considered Harmful in A List Apart No. 107.

    If we had unforgiving browsers, all web content would validate. Validation is not an end in itself, but it is an important step.

  63. soulscratch said on

    I apologize for being so blunt, hehe it’s all good but nobody’s perfect and because of that fact I really don’t condone the use of XHTML – even if IE supported it, we would get xml syntax errors all across the board and a good majority of sites that use its syntax would be a victim because of the draconian error handling ( kudos to Tommy Olsson for that term ).. which is why I stick with HTML 4 Strict ;)

  64. serdar said on

    @Jeffrey Zeldman, Yes you’re right about the quirks mode. I think it’s changed since IE7, right?

    Still it does not mean IE supports XHTML in my opinion because it is not parsed as XHTML at all when served as text/html.

    According to Wikipedia, Internet Explorer “does not support XHTML, though it can render XHTML documents authored with HTML compatibility principles and served with a text/html MIME-type.”
    http://en.wikipedia.org/wiki/Internet_Explorer#Features

  65. inimino said on

    Nathanael Ritz wrote:

    Actually, Google doesn’t hate standards. They are just slow at getting to the full implementation of them. For example, the section they have about their website survey validates as HTML 5.

    Note that the author of that survey is also the editor of HTML 5.

  66. Jeffrey Zeldman said on

    inimino: URLs?

  67. goodwitch said on

    I predict that we are at the tipping point on the adoption of web standards. Surveys like this are important to document the current state of the web. And while your results, on first glance, seem dismally low, I believe that if you were to conduct this survey again, one year from now, you would see significant improvement. Why am I so optimistic? Because for the first time in the history of the web we have a shot at the major browser vendors rewarding sites coded to web standards.

    With IE8 set to support web standards by default (while allowing non-standards based sites to degrade gracefully)…the old problem of “well, it works in this browser, why should I care about standards?” will become, “oh, I get it, web standards really do make a better web for designers, developers, search engines, mobile and all users regardless of disabilities.”

  68. inimino said on

    Jeffrey:

    Retrieved from http://ln.hixie.ch/?start=1138508345&count=1

    Then around midweek I ended up being Slashdotted because of some research I did that Google published. I love that Google let me do this research. I’ve been hoping to study authoring practices for years.

    (Thanks to Philip Taylor for finding this.)

  69. Eric said on

    It is a continuing struggle to keep markup valid. I finally achieved that in the latest build of our company’s site, and now I just double-checked it and found that an engineer who changed the search box last week caused five new errors! There will be a discussion Monday…

    Our site validates differently each day depending on if the content team has used an ampersand. The WYSIWYG text editor on our backend does not escape ampersands, and it is a source of constant annoyance to me, so the days the site validates are intermittent. Still, I congratulate myself that I have done everything possible to make it valid!

  70. Wade Jackman said on

    Very interesting study. Im curious to know what kind of errors are showing up in the rest of the list.

  71. Web Standards Test: Top 100 Sites | Design Newz said on

    [...] Web Standards Test: Top 100 Sites [...]

  72. kikykuang said on

    Very useful

  73. Jeffrey Zeldman said on

    Im curious to know what kind of errors are showing up in the rest of the list.

    Easiest thing in the world. Go to Alexa’s Top 100 US Sites and run them through the validation service of your choice. validator.w3.org and htmlhelp.com/tools/validator/ are two good ones.

  74. Danny Ward said on

    makes me wonder if a similar reticence was their ‘x’ years ago about letting go of tables for markup. Namely the mindset is ‘We’re big and special, we don’t need to validate’. Maybe, the 200-300 guys would be more keen about being well-formed and compliant.

    Great work Jeffrey.
    Danny

  75. Jeff Croft said on

    Curious:

    What kind of breakdown did you find on DOCTYPES? As we are in a period now where cutting-edge standards-aware developers are migrating in droves away from XHTML and back to HTML 4 (or forward to HTML 5), I’m curious is that trend is also apparent among the big boys you surveyed.

    Personally, I think what DOCTYPE developers are aiming for these days is a lot more interesting than if they made it 100% of the way there or not.

  76. Jeffrey Zeldman said on

    I don’t know how many of the top 100 sites are authored by cutting-edge standards-aware developers. Yahoo is. Pornfingers may not be.

    Most of the top 100 sites declare XHTML 1.0, whether they are valid or not. Many sites declare strict, even when their content begs for a more lax DTD.

  77. goodwitch said on

    I think we are at a tipping point when it comes to the adoption of web standards. A year from now, I expect the results to be significantly improved if IE8 follows through and rewards web standards code (while degrading gracefully for non-compliant code).

    For web standards to really permeate the web, designers/developers will need immediate feedback from browsers. The less forgiving the browsers…the faster web standards will become a real standard.

  78. Max Design - standards based web design, development and training » Some links for light reading (10/3/09) said on

    [...] Web Standards Test: Top 100 Sites [...]

  79. gringaus said on

    We also at first got a large number of common and not very common errors, and worked them through. Don’t give up on the “whole validation thing”. You don’t have to shoot for the stars, just make the page semantic and accessible – easy. it’s quite hard to get a large product validated, but we’ve done it on a number of large sites, not only their homepages (bbc), but also all channels, subchannels, and all major pages are now valid. we provide our editorial with the table of special characters to avoid having content with the bunch of errors. So, here we go so far:
    http://southernliving.com/ passed
    http://coastalliving.com/ passed
    http://sunset.com/ passed
    http://realsimple.com/ passed
    on vast number of templates

  80. appleton.me // Only 7 of 100 Top Web Pages Validate said on

    [...] this in mind, it is a little worrying to read that when Jeffrey Zeldman tested the top 100 US sites listed at Alexia, only 7 validated. On this day, in this test, seven [...]

  81. Leannekera said on

    Jeffry, reading through these comments it must get tiring having to prove yourself to so many people. i find it sad :(

    You would think that these people had better things to do all for the sake of making themselves feel high and mighty!

  82. Thiago Cavalcanti said on

    Hi! The problem with the method you used to validate(through W3C online validator) could’ve been avoided if you had used instead this Firefox extension to perform the validation.
    Be noted that I still think your test valid, at least partially.

  83. Wait till I come! » Blog Archive » A few things the web development community can learn from The Green Movie said on

    [...] light of truth on the arguments provided. Another interesting “oh not again” moment was Jeffrey Zeldman doing the inaugural testing of the top 100 sites in a validator causing an avalanche of [...]

  84. Faruk Ates said on

    Here’s a thought in response to the debate over this site’s valid source code and what not:

    How many developers out there today test a site in Safari or Firefox, see it work, then try in IE and see it break, only to just blame IE and proceed to find workarounds and hacks, as opposed to validating their HTML and CSS to see if maybe they’re just authoring with errors?

    I suspect more people do the former than do the latter. And that’s a bad thing.

  85. Dave said on

    Self-professing XHTML sites all use type=text/html which is barely legal for lowest grade XHTML. XHTML should use application/xhtml+xml.

    The practical importance is sharing XML namespaces, as with MathML and SVG.

    All the compliance banners touting XHTML ignore this facet of the standard. Even the W3C pages are inconsistent, for example, the stock template defaults to text/html and fails to correct downstream for XHTML.

  86. Bookmarks for March 14th from 16:38 to 16:38 | Matthew Kempster said on

    [...] Web Standards Test: Top 100 Sites – Jeffrey Zeldman tests 100 top sites to see if they are standards friendly [...]

  87. Real Estate Raj said on

    Validation Novice here- Generally speaking, what are the seo benefits having a your site validated?

  88. Jeffrey Zeldman said on

    Validation Novice here- Generally speaking, what are the seo benefits having a your site validated?

    There are SEO benefits to authoring your content with semantic markup. Authoring lean, meaningful content with web standards helps people find your content by wrapping it in appropriate semantics that machines and search engines like Google understand and respond to.

    There are no SEO benefits to having your site validate. But authoring valid content helps you check your site for errors that may cause display or function problems in one or more browsers or devices. And striving for validation is a best practice of standards-based design.

    Sometimes you can’t achieve validation—for instance, when a third-party ad stream inserts invalid markup on your site.

    And sometimes you can’t maintain validation—for instance, when a person charged with updating the site’s content has the ability to insert malformed HTML into a page.

    But achieving validation, at least during the initial template development phase, helps assure that your web content will display and function correctly in as many browsers and devices as possible, thus getting your message (and branding) in front of people once they find your content.

  89. Andrew Gregory said on

    One of the best reasons for validation is to allow (X)HTML to be developed and expanded into the future, without causing problems. Invalid HTML is already causing problems with the enhanced form capabilities of HTML5. Opera has implemented many of these features, and every now and again falls foul of code that used to be invalid, but now has meaning.

  90. Sexta-feira dos Web Standards #17 · Portfolio e blog sobre Web Standards - project.47 said on

    [...] Web Standards Test: Top 100 Sites – Zeldman resolveu fazer uma pesquisa dos 100 sites mais populares da Internet (via Alexa) e concluiu que apenas 25% deles usam os Web Standards. Veja uma pequena análise dos sites mais bem colocados. [...]

  91. Designing With Web Standards | Facebook said on

    [...] IE-8 http://microsoft.com/windows/internet-explorer/05-Mar-2009 – Web Standards Test: Top 100 Siteshttp://www.zeldman.com/2009/03/05/web-standards-test-top-100-sites/26-Feb-2009 – Future is Bright for Web [...]

  92. JeffCroft.com: Jeffrey Zeldman: Web Standards Test: Top 100 Sites said on
  93. Understanding Web Standards | DL Web Design said on

    [...] Web Standards Test: Top 100 Sites Dew Lilly Website and Graphic Design: Professional . Simple . Design I am a professional freelance web designer located in Phoenix AZ.  My expertise is front-end W3C Standards Compliant Web Design.  I have over 8 years of experience in web design and web maintenance and typically I build websites for small to medium sized businesses and non-profit organizations. var addthis_pub = ‘dslilly'; var addthis_language = ‘en';var addthis_options = ‘email, favorites, google, digg, delicious, twitter, facebook, reddit, live, more'; Tweet This Post  [...]

  94. Web Standards Test: Top 100 Sites | Internet Microsiervos said on

    [...] Web Standards Test: Top 100 Sites Sobre la validación según los estándares web 100 sitios web importante. El dato: sólo 7 de los 100 validan correctamente. Ordenadores / Software Recibe i Microsiervos por correo. Un envío al día: [...]

  95. JW » Here’s the Thing About Validation said on
  96. Carl - Web Courses Bangkok said on

    Very interesting, especially that MSN validated! Wow

  97. Web Design References: Standards, Guidelines and Patterns said on

    [...] Web Standards Test: Top 100 Sites – Jeffrey Zeldman [...]

  98. Jeffrey Zeldman tests the top 100 websites… | What else is new? said on

    [...] read more at: “Web Standards Test: Top 100 Sites” [...]

  99. Twitter Trackbacks for Jeffrey Zeldman Presents [zeldman.com] on Topsy.com said on

    [...] Jeffrey Zeldman Presents http://www.zeldman.com/2009/03/05/web-standards-test-top-100-sites – view page – cached Web design insights since 1995. Personal site of Jeffrey Zeldman, publisher of A List Apart Magazine, founder of Happy Cog Studios, co-founder of The Web Standards Project, co-founder of the Event Apart design conference, author of Designing With Web Standards. — From the page [...]

  100. Web Development Nerdy Daily Links For 3/06/2009 – HUGE POST! | AndySowards.com :: Professional Web Design, Development, Programming Freelancer, Hacks, Downloads, Math and being a Web 2.0 Hipster? said on

    [...] Jeffrey Zeldman Presents : Web Standards Test: Top 100 Sites Great list of the top 100 sites for web standards (tags: webdesign accessibility webstandards articles validation web-standards) [...]

  101. Web Standards Test: Top 100 Sites said on

    [...] can wreck a once-valid template, as can later third-party development work."  Read more at: http://www.zeldman.com/2009/03/05/web-standards-test-top-100-sites/Published by: Zeldman.comDate: March, 2009Author: Jeffrey Zeldman [...]

  102. Website review – IFA Talk said on

    [...] is acceptable in experiential or cutting edge sites, but valid code is an indication of quality. As Jeffrey Zeldman says: Validation is not the be-all of standards-based design; it merely indicates that the markup, [...]

Comments off.