W3C HTML Validation Service: Source Code

The source code for the W3C HTML Validation Service is available under the terms of the W3C Software Copyright.

The purpose of making the source code available is partly to allow others to set up mirrors of the service (either publicly or within an intranet behind a firewall), but also to allow us to collaborate on making the service better -- there are many ways the service could be improved but I only have a small amount of time to work on it myself.

Retrieving the Code

You can retrieve the code a number of ways:

• If you just want to glance at the code, or see its revision history, you can browse it on the web.

  The most interesting files are currently a CGI script called "check" that does pretty much everything, and possibly also the httpd.conf. (select the topmost revision numbers on these pages to see the most recent revision of each file.)

• If you want a copy of all the files that make up the service, you can grab a tar ball (~1.5M, updated every day at 06:00 ET.)

• If you intend to actively mirror the code and/or contribute patches to the code, you should install and become familiar with CVS; this is the tool we use for revision control (it is also used by the Apache and Mozilla developers, and is generally a good thing to get to know.) More information on CVS is available courtesy of Pascal Molli.

  Our CVS base is available read-only, using CVS pserver authentication a la:

  bash$ export CVSROOT=":pserver:anonymous@dev.w3.org:/sources/public"
  bash$ cvs login
  (Logging in to anonymous@dev.w3.org)
  CVS password: anonymous
  bash$ cvs get validator
  cvs server: Updating validator
  cvs server: Updating validator/htdocs
  U validator/htdocs/about.html
  ...

Prerequisites

Before you will be able to get the code to run, you will need a few things already installed on your system:

• A Unix-like operating system. It may work with Windows NT or other systems, but I haven't tried it yet. It has worked on Linux, Solaris, FreeBSD, HP-UX, and Digital UNIX. If anyone tries it on other systems and gets it to work, please let me know (and send me patches, if they are needed to get it to work eleswhere.)
• SP version 1.2.1 or higher. SP is the SGML parser used by the service. More recent versions than 1.2.1 will also work; I am currently using version 1.3.
• A collection of DTDs and other SGML files to validate against. You don't strictly need these on your system since SP will retrieve them off the Web if you use URIs in your doctypes, but you probably want them locally for efficiency. (You don't need to download this tar ball if you mirror everything using CVS.)
• The LWP Perl module
• A web server: I am currently running Apache version 1.3.6, but likely any version will work, and other httpd's may work as well.
• Perl: the main CGI script that does everything is written in Perl.
• weblint (optional): the service provides an option to run weblint on the page being validated. Currently this services uses version 1.017.

If anyone tries to use the code, and succeeds (or fails), please let me know!

Development

For information on current development activity, see the www-validator mailing list archives. Any changes to the code or documentation related to the service cause a notification message to be sent to the www-validator-cvs mailing list.

Any changes to the service will attempt to maintain compatibility with a list of test cases.

Gerald Oskoboiny
$Date: 2000/04/28 07:17:59 $