[imps] validator front end not talking to servlet

Michael[tm] Smith mike at w3.org
Thu Apr 18 15:34:45 PDT 2013

Rancid Iodine <rancid.iodine at gmail.com>, 2013-04-18 18:13 +0100:

> Michael once said to me to get rid of the legacy w3c validator altogether
> (considering the validator.nu does a better job validating any version of
> HTML), but we're in a corporate "situation" here wherein I'm told they will
> not accept any validator unless it has a "w3c" stamped on it's forehead ...
> which validator.nu does not have.

The W3C instance at http://validator.w3.org/nu/ does. Even the URL alone does.
http://validator.w3.org/nu/about.html makes it quite clear that it's an
official W3C product/service -- the "W3C Nu Markup Validator". I'm not sure
what more I could do to stamp "W3C" on it clearly.

> Is there any obvious solution that springs to anyone's mind? If not, does
> anyone have any stories to tell regarding an appropriate distro, networking
> issue or install context they came across?
> The server at http://validator.w3.org seems to have everything we need - I
> can only assume that's running validator.nu on the back end as well. Does
> anyone know what that's running, or any details about how it's set up?

It's running on a Debian 6 server, with the latest patches.

The code for the legacy validator is Perl code running through a CGI script
in a Perl 5.10 environment.

The validator.nu backend is Java code running through Jetty in a Java 6

When the legacy validator gets a request to validate an HTML5 document, it
does some pre-processing on the request and then sends it to the
validator.nu backend, through the validator.nu REST interface. The
validator.nu backend then processes it and sends the validation results
back in XML, which the legacy validator Perl code parses and formats into
HTML and then emits through its Web UI.


Michael[tm] Smith http://people.w3.org/mike

More information about the Implementors mailing list