To validate or not to validate?

This post is intended to address the issue of HTML validation from my point of view. I will try to point out the advantages and the reasons a web page needs to validate (again from my point of view of course), but I will try to line out a few other opinions on the matter. Long story short, I will try to answer a simple question: "Should I care about validation or not ?"

What is HTML validation?

For those of you that still don't know what validation really is, here is the short explanation. Validation is the process by which an HTML or XHTML page, meaning it's source, is brought in accordance with world wide accepted standards. What this means is actually writing the source code of the page in such a manner that is respects all the guidelines implemented by the W3C (that's the World Wide Web Consortium). In order to make sure your web page is valid from an HTML or XHTML point of view, you can use a lot of different tools, but my favorite is the online one provided by W3C at http://validator.w3.org. Simply type in the URI of your page or copy/paste your code and hit the "Check" button. That's it.

Does it even matter?

Believe it or not, it does. There are people that say that validation is just something that developers thought up to brag about their skill. Others just don't care if it's not valid as long as it works. But here is where the funny thing comes in. For it to work it still needs to be coded after some specific guidelines, so unintentionally people do give a though to validation. After all, they do want their pages to look nice, don't they?

There are many situations when a page full of errors looks flawless, the layout is just the way it was planned. That doesn't mean it's right and that it doesn't need to be coded differently. It just means the browser is aware there are bad developers or that errors might slip by unnoticed and it just ignores the situations and renders things just like nothing happened, mostly "guessing" how things should have been.

So browsers fix things for me. Then why should I bother to validate?

The idea here is that browsers can't fix everything. There are situations when you do need a valid page and a well thought out structure. One of the most common situations i when you're using JavaScript, like the many wonderful frameworks available out there (jQuery, MooTools, Dojo to name just a few). These don't bother to check your code, they just use what they are given.

For example if you forget to close a paragraph (<p> tag), the browser will render just fine. When it finds the next paragraph tag or the next block element, it assumes you closed the previous one. But if you use some JavaScript code to work on your paragraph tags, let's say for example that you want to change their background when you hover over them, errors will appear. JavaScript will assume that the paragraph ends when it meets the closing tag (</p>), but that will cause the change to be applied to the next elements (maybe other paragraph or maybe lists), until that closing tag is met.

That's all? Closing tags?

No, that's not all. Because browsers continue to evolve and improve, a strict set of rules is required. And these rules must be respected by everyone, both website owners and browser developers. If they are ignored, progress will be just a long list of hacks and fixes from either side. Browser developer will try to support obsolete and sometimes wrong interpretations of standards just to please the users and not render the pages chaotically. And website owner and developers will try to work around browser incorrect renderings by various CSS, HTML or JavaScript hacks.

---

In closing, I would like to say that I try my best to make every page as valid from a W3 point of view as possible. There are situations when it's jut too much work and it's not worth rewriting the entire code just to shift a few pixels. But these days things are a lot more easy. Modern browsers even support standards not yet officially recognized in preparation for when they will be made generally accepted, like CSS3. The only thing developers need to do is focus on the new implementations (which are already pretty old by now) and leave behind the old hacks. And if a browser doesn't display the page the right way, it just means it's the browser's fault. Most likely it's an outdated version and the users should upgrade for their own sake.

On the same note, I would like to keep a moment of silence for the soon to be dead and hopefully forever forgotten, IE6. Your death's coming soon.

Source: StatCounter Global Stats - Browser Version Market Share