One of the most important things a web developer can do to achieve the broadest accessibility possible is to ensure that the development code that they are using is being used according to specifications. Code validation ensures that this critical point is being met; think of it as spell-checking your HTML code.
XHTML 1.0 (www.w3.org/tr/xhtml1), has been stable since August of 2002. These standards specify exactly how the authoring (or mark up) language is to be used, and establishes a benchmark, not only for the content authors, but for the software manufacturers who are developing the next generation of user agents (browsers) as well. Authoring to these standards allows content developers a level of assurance that what they are posting to the web will be accessible not only to the popular browser of the day, but to any standards compliant user agent in the marketplace; including but not limited to adaptive technologies used by people with disabilities.
Code validation is really pretty simple. Though your page may look fine in your own browser, there may lay within the page's code various HTML tags that aren't "nested" properly or tags that are no longer considered valid (see deprecation ). Various online validators will identify coding errors in any web page, outlining where those mistakes are by line and character number so that a developer can go back and fix the errors. The different validators have slightly different methods of testing your work, but they essentially provide a means for either checking pages on your server, or by inputting your source code into the testing tool. The code is then reviewed (the technical term is "parsed") by the validator and a results page is returned to you. Generally these results will provide not only the errors found, but also links to resources so the developer may undertake that repair.