How to Drive Yourself Crazy


Too Many Details About HTML Validation

HTML Validation is an automated process by which a computerized device examines your finished HTML code, compares that code to all the rules it (the computer) knows about exactly how the HTML code is supposed to be written, and then tells you exactly where all the errors are in your code. You might think that this sounds like a very useful process, and for the most part it is. The big difficulty arises when you discover that most of the publically available validation services ( is the big one) don't really have a very good handle on telling you exactly why your errors are errors. This is compounded many times over if you don't have a very clear understanding of Document Type Definitions.

In order for a HTML document to be accepted by a validator, it must have a Formal Public Identifier (FPI) or a <!DOCTYPE> tag, which precedes the open <HTML> tag, at the very top of the document, before anything else. This identifier specifies exactly which version of HTML is being used in the document, and identifies the location of the Document Type Definition, which is used by the validator. The FPI for the documents in this site looks like this:

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "">

This identifies the "flavour" of HTML that is used in this site as "HTML 4.0 Transitional", and specifies the location of the HTML 4.0 transitional DTD. The Document Type Definition (DTD) is basically a set of rules which define exactly how HTML of that specific version is to be used, which attributes are valid, which values can be used with which attributes, and which order tags can appear. You don't have to understand exactly what the DTD says, but if you write code which does not follow the rules set down by the DTD, your code will not validate, and the validator will not do more than point out where your errors are.

A very good example of this can be seen in a web site I designed early in my career, which contained a <META> tag which identified the character set being used. The tag looked like this:

<META HTTP-EQUIV="Content-Type" CONTENT="text/html" CHARSET="iso-8859-1">

The response I got from the validation service when I tried to validate this document was "There is no such attribute: CHARSET". Now I know that it is possible to do this, I have seen META tags which indicate the character set being used in the document, but apparently I didn't look at such a tag closely enough before I tried to use it. In reality, the tag should look like this:

<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=iso-8859-1">

A very subtle difference, the quotation marks are in slightly different positions. The point is that the validation service was correct, there is no such attribute as CHARSET, but it didn't tell me that my quotation marks were in the wrong place, and it took me quite a while to search out and discover exactly the reason why my code was generating such an error.

What it comes down to is that, in general, validation isn't necessary. I don't want to minimize the value of validation, because it is, in fact, very useful and very educational if nothing else, but you don't need to be able to write 100% perfect code all the time, and while you're learning HTML, validation can be and often is more confusing than it is helpful, primarily because the error messages returned by validation services are so incomprehensible.

Here are some good sources of more information about HTML validation

  1. Why should you validate your HTML code?

  2. To understand HTML Validation, it is important to understand the different levels of HTML and what conformance to them means.

  3. HTML Validation is as much an understanding the correct use of HTML Document Type Definitions (DTDs) and the <!DOCTYPE ...> tags as it is an understanding of how to write correct code. Both are important, but if you lack an understanding of either one or the other, it's not likely that you'll have a very easy time writing pages that validate 100%.

  4. If you've already got a good idea of what to look for, here are some HTML Validation Services which are publicly accessible:

    • HalSoft - WebTechs HTML Validation Service. This was the Granddaddy of the HTML Validation Services, but it appears that the WebTechs domain name was sold to a pornographer, because the validation service that used to occupy that web space has been replaced by pictures of naked women. When it existed, it was one of the most comprehensive and the one which delivered the least understandable error reports.

    • W3C Validation Service is the latest version of the KGV, listed below. It was created by Gerald Oskoboiny and should be given preference over the KGV since the KGV is no longer being maintained.

    • The Kinder, Gentler, HTML Validation Service - Otherwise known as the KGV, it was created and is no longer maintained by Gerald Oskoboiny, who now works for W3C. The KGV runs from almost exactly the same code that HalSoft runs from, but the KGV delivers somewhat more understandable error reports. The KGV also allows you to validate your code using WebLint, which is an application similar to the "Lint" application used by C programmers to debug their code. The KGV is no longer being updated because Gerald got a job at W3C, so he's been providing up-to-date validation services there.

    • Dr. HTML is a pay-per-use service. You can validate one page of HTML if it's posted on an accessible server, but you can't do an interactive validation, like you can with the previous services. If you want to validate more than one page, you have to make arrangements to pay for the service with its creator, Thomas Tongue and Imagiware, Inc. Something to keep in mind if you use the Dr. HTML service. It doesn't really validate your HTML, but it will spell-check your document and verify that all the image tags, table tags and hyperlinks are constructed and functioning correctly, and it will alert you if you have included other code that is incorrect sometimes, but if it does, don't expect an explanation of what is wrong.

100% Valid HTML 4.0 - Thank you very much!