You can learn more about HTML simply by example - look at the code that other people have written and learn from it (Netscape allows you to look at the HTML code by selecting Document Source from the View menu, Internet Explorer uses View/Source). A disadvantage of this is that many web pages, though they look good, break many basic rules of HTML - use caution...
A few of the earliest extensions to HTML are listed on Netscape's page (the link is to a UK academic mirror of this document). These give some of the basic additions to the HTML specification, which have by now been accepted by the majority of browser developers.
Inevitably, there are a wide range of books available to help you learn HTML, some much better than others. We recommend 'HTML by Example' and 'HTML for Dummies', with the advice to browse before you buy and see which style of book best suits you.
If you're looking for a quick and easy way to test your web pages before you publish them, you can try weblint, a program on the Computer Society machines which thoroughly tests your pages to check that the HTML is logical and well-structured. Just type
weblint http://location...
replacing http://location... with the location of the page you wish to test (or just the filename, if it's on SUCS). You will receive a list of errors on the page. If the command prompt returns without printing any errors, it means your page is problem-free.
And finally, the golden rule: test your pages on as many browsers as you can. We particularly recommend Lynx, a text-only browser which is generally very intolerant of HTML errors and so is excellent for testing your pages. It's available by typing
lynx http://location...
from the command line on the Computer Society machines. Again, replace http://location... with the location of the document you want to view.
Happy web page writing!