Validation101

The Question

I'd like to redo the HTML on my website and am unable to vallidate the html from your site. I'm wondering if you have any tips to help me get started.

In three points, the basics are

1- learn proper HTML

Making your web pages valid means, basically, making them "proper" HTML. If you don't know what proper HTML looks like (as opposed to tag soup), then cleaning up your site is going to be very frustrating. Of course you will learn through practice, and validating a site is an excellent exercise for that, but a decent base level helps.

Get a good book. There are plenty of these, and there are also horrid ones. If you can, check the credentials of who built the books, notice if it mentions standards, and of course flip through it to see if it looks serious enough. That should help distinguish real HTML learning books from snake oil.

2- take your time

So you've used the validation on one page for the first time, and, ugh, you get an obscene amount of errors. Forget about that number. Focus on, say, the first 2 or 3 errors.

Most likely, the first error will be about a doctype thing. Fix that first, it means that your documents does not say what flavor of HTML it is, so the validator guesses. Go to Valid DTD list, pick one flavor (try HTML 4.01 strict or transitional, that's probably a good one to start with if you have no idea which to choose), and copy-paste the <DOCTYPE ... > at the top of your document, before the <html>.

Alternatively, the first error will be about character encoding. Maybe your document does not declare its encoding, maybe it declares the wrong one, maybe the server is badly configured. This sounds awful, but usually, it can be fixed by:

using the "encoding" drop down to try a few encodings. Usually, UTF-8 or iso-8859-1 will do. Find one that does not cause the validator to panic.
Then check the FAQ about Charset and follow the instructions to declare the encoding you found was the right one for your page.

Now save the document and re-run the validator on it.

Again, only focus on the first few errors. Trying to fix them all at the same time is difficult, even risky: sometimes an error actually cascades into a flow of other errors, so fixing the first one may make many others disappear. So just check the first few lines of error, sometimes they will give you context or information on what the first error is about.

Don't forget to read the error message explanations. The error themselves tend to be, well, obscure. The explanations below are here to translate the error into something intelligible, and often even provide instructions to fix the problem.

A bit of searching around may be necessary, but usually you should manage to fix that error easily. Just remain patient, and don't assume that the validator is wrong because you can't figure what's causing the error. The validator does have bugs, but for some edge cases you're not likely to fall upon.

3- use collective wisdom

If you have not managed to fix the first error by reading the error message explanations and flipping through your book, use the collective wisdom of the Web to help you.

Here are a few leads to find your answers:

have you read through the validator's Help and FAQ? What are you waiting for?
There is a large archive of the www-validator list where lots of people before you may have struggled with the same problems. Look at the right of the error message, there's a strange little "envelope" character. Click on it and you'll be on the "feedback" page. Don't rush to send a plea for help just yet! There is a form field that can help you search the list's archive for messages on the error you are facing. Give them a look, chances are you'll find your answer there. Then, *only* when you have checked these mail search results, you can send mail to the list asking for help on the error. Be clear, be precise, be detailed. The better your call for help is, the easier it will be for people to help you. And don't forget that the people on the mailing-list are almost all volunteers, behave accordingly...

Now you should be able to fix the first error. Good. Rinse, lather,repeat until the document is clean.

Congratulations, your first page is valid. Now what? You still have plenty other pages on your site to take care of. How? Just like you did for the first one, step by step.

There are a few articles on the W3C QA site that explain strategies and tools to improve the quality of a site, small or big, step by step. Among them, check the Web Kit and Step by Step Guide

I hope this guide can be useful to you. If it's not… Tell us why -> public-evangelist@w3.org (publicly archived mailing-list)