Comments

1. Derek says at 2005-10-03 13:10:

   I spot checked a few of these and it’s worth noting, I think, that while the NSA homepage does validate, it’s only a splash page. If you continue to the HTML index, the page does not actually validate.

2. Pete says at 2005-10-03 14:10:

   Derek: As mentioned in the test disclaimer, it is possible that some pages validated are only a splash page. The test, in my opinion, still gives an indication on the current state of web standards usage. NSA will be off the list soon. 14 valid, going downhill…

3. David says at 2005-10-03 16:10:

   I see from this post that you are an advocate of the DCMI. You point out that all of the sites tested don’t support the DCMI. I’d say that I have seen very few sites that support it and would have to say that most web developers have no clue what the DCMI is in the first place.

4. Justin says at 2005-10-04 23:10:

   One thing to note is that some of these failures may be due to the inclusion of some very ugly JavaScript from WebTrends. In the case of the sites that I have developed for the government, the syntax is completly valid once you remove the JavaScript. So don’t count out all of the sites just because your report shows a little red X next to the URL.

5. Pete says at 2005-10-05 09:10:

   Justin: If you read the test disclaimer I mention that “a site that does not validate may still be more accessible than one that does”. I am not “counting out” any site based on validation error count. The report shows the number of errors found and if you look at figure you will see that most sites have multiple errors.

   In my experience adding javscript for WebTrends does not impact the ability to have a valid site. It is a simple matter of getting the script tag right. Which web site are you refering to?

6. Justin says at 2005-10-05 16:10:

   In particular http://www.disabilityinfo.gov. It is listed on the FirstGov “Site Index” but somehow did not make it into your test (not sure which list of sites you pulled your data from since there are many).

7. Pete says at 2005-10-05 16:10:

   Justin: The list is based on all the sites found at A-Z Index of U.S. Government Departments and Agencies. I know this is not the complete list of all government web sites but it was what I could get in a decent list format. If you have a more comprehensive list of sites I would be very interested in gaining access to it.

8. Matthew says at 2005-10-12 22:10:

   State of Louisiana should be removed from the list, in my opinion. The validated page is simply a meta refresher to the actual non-validating front page (256 errors).

9. Steve Moore says at 2005-10-12 23:10:

   I would also find it interesting to see how local governments would fare in this test. As a local government webmaster, whose site does validate and uses CSS for layout, Larimer County Colorado, I think the percentage might be higher. Most of my local government peers are solely responsible for their site, because of the smaller size, and frequently hand code the content rather than using non-standards generating CMSes.

10. Pete says at 2005-10-12 23:10:

    Mathew: Thank you for the information. Louisiana has been dropped. 13 valid sites.

11. John Allsopp says at 2005-10-13 05:10:

    I recently did a big survey of Australian sites (not just government) for their adherence to best practices in development and accessibility. It was presented at WE05.

    Similarly disappointing.

    The slides are here

    http://we05.com/resources/john-allsopp.pdf

    the podcast of the presentation is here http://we05.com/podcast/mp3/we05-12-john-allsopp.mp3

    A full article and the data will be available soon

    john

12. Pete says at 2005-10-13 10:10:

    John: Interesting findings. It looks like you did all those web sites manually. If you have suggestions for tests you feel are lacking from my automated version I would be interested in hearing your suggstions.

    I am about to profile more countries (next up is your neighbour New Zealand). If you have a comprehensive list of URLs for Australian government web sites I will “do” Australia too.

13. Tim Swan says at 2005-10-13 16:10:

    As a part of a team that has worked *extremely* hard toward using valid xhtml and prizes both usability and accessibility it’s a little dissapointing to see these results.

    One little thing: I notice that the site that I work with (www.maine.gov) says that we have 2 validation errors - yet when I test it I only get one warning — an unencoded ampersand. We work hard to validate every single page within the state government portal.

    I often find that people who theorize about validity and accessiblity are not the ones that are responsible for trying to make it happen. The typical government web site has hundreds of thousands, if not millions, of pages all of which are created and maintained by hundreds of people within agencies whose management doesn’t necessarily understand how challenging getting content on the web can be, or the special considerations that need to be taken into account.

    I find value in this sort of blunt-tool assessment of where government site are, though I ever day I see talented, hard-working people that are fighting the odds to make sure that the promise of web accessibility becomes a reality and not just some geek theory divorced from the reality of budgets, training, education, and the needs of an incredibly diverse audience.

    No excuses; just a little clarification. And remember — automated tests are a pretty blunt tool.

14. Pete says at 2005-10-17 09:10:

    Tim, thank you for your comment. My motive with this test is not to “expose” individual web sites. My ambition is to compare how government policy in various countries affect web standards usage. Many countries have guidelines for government web sites that promote the W3C recommendations.

    I agree fully that this test is a blunt tool.

    Typically, the number of validation errors vary over time. When the automated test was run there may have been some information that created 2 errors. When I run the validator on http://www.maine.gov right now I get 7 errors.

    (If you want to find validation errors on your entire site, have a look at the article Validating an entire site or send me an email and I’ll run it for you.)

15. John Allsopp says at 2005-10-28 03:10:

    Hi,

    I presented something similar at WE05 and have just published the results. It’s a little more broad in terms of its criteria, covering 4 areas of best practices in web development, and looks at 83 major sites in Australia across several sectors in government and industry.

    Results are here

    john

16. Nena Moss says at 2006-03-07 22:03:

    Thank you for this effort toward validation. Our redesigned site was uploaded on 10/01/2006, just after your test. At least the link to the validator shows our improved results.

    Can you point me to some information on validating .jsp pages? Some of our Java-developed products don’t produce pages that validate. Please share any resources for validating dynamic pages. Thank you!

17. Peter Krantz says at 2006-03-08 09:03:

    Nena: Validating JSP pages is no different than validating pure HTML pages. JSP and other page frameworks create HTML and that is what the validator will look at. It is possible to create valid markup with JSP and other frameworks.

    A suggestion is to teach your development organization to use the validator for testing user interfaces during development.

18. Nena Moss says at 2006-03-08 17:03:

    Thanks Peter. I will attempt to do just that.