When search engines process queries from users, they rely on large keyword indexes that are generated using their special software - web crawlers, often referred to as spiders or robots. These pieces of software run all the time, and they crawl web pages following all the links and cataloguing the content. Search engine spiders learn about your site one of two ways. Either they follow a link from a known site to your site, or you have submitted your site to be entered into the search engine catalog. Here is the Google page for site submission. The details of the processing are kept secret but there are well-know and widely used techniques for helping these spiders classify pages correctly.
One simple yet important way of providing search engines with information about the page is via the <meta> tags of the <head> element. These were first mentioned in the Simple website templates article. There are many uses for meta tags; for this discussion there are only a few of interest. The example below has a partial sample of what the contents of the <head> element might look like. Notably, it includes three types of <meta> tags.
The description <meta> tag should contain the page description and is often used by search engines as the content of the summary displayed along with the link on the search results page. In addition, having the page description agree with the page content will likely increase the page rank in the pool of all pages on a particular topic.
The keywords <meta> tag is meant to contain a comma separated list of keywords relevant to the page. The list items may be individual words of groups of words. A page about SEO may have these keywords set: "search engine optimization, meta tags, seo".
Lastly the robots <meta> tag is there to indicate to the spider what it should do with the page. There are six valid values: "index", "noindex", 'follow", "nofollow", "all", "none". The last two are synonyms for "index, follow" and "noindex, nofollow" respectively. To tell the robot to index this pages, used the "index" value, and "noindex" to prevent it from doing so. To follow the links on the page use the "follow" value, "nofollow" otherwise. Combinations of the first four values are allowed as example 1 demonstrates. Search engines are under no obligation to use the hints in the robots <meta> tag. Setting it to "noindex, nofollow" or "none" will in no way guarantee that the information contained on the page will not be index and accessible via a search engine, this cannot be used as a security measure.
In the days long ago, when life on the web was simple and people were not obsessed with ranking highly on search engines, <meta> tags were all that was needed. In an attempt to produce higher quality results, search engine companies are relying on more advanced techniques for page processing that render meta tags alone pretty much useless.
Proper page organization methods were discussed earlier, SEO is one of many things that will benefit from it. Having major titles for sections that contain some of the keywords defined in the meta tags increases the page rank. This can be done with the use of headline tags such as <h1>, <h2>, etc. It also helps if the actual content of the page contains the same keywords sprinkled through the text. For the page appearance it is not necessary to use headline tags since all the formatting can be accomplished using CSS and accomplished in a cleaner way. However this will not help the search engine categorize the page. Keep in mind that CSS provides methods for applying styles to built-in html tags so no control over the page will be lost if headline tags are used instead of <span> tags.
Along side the meta tags inside the head element is the <title> tag. Often overlooked, it is good practice to make sure that it contains something relevant.
Another SEO method involves directory and file naming and organization. Sorting website content into directories named after the keywords relevant to the content will likely increase page ranking. For example, when trying to rank a photo album of images from Paris, placing these pictures into /photo/paris directory is better than just putting them into the same directory as all other images. Naming the image files well is important as well, eiffel_tower.jpg is much better than img_1002.jpg.
All the methods described above are no longer enough. The main reason is spam. Many popular sites that do not renew their domain registrations often end up in the hands of spammers. After that the pages uploaded to the site try to exploit various security holes in current web browsers (mostly IE) in an attempt to use this machine as a spam bot. Malicious intent is not always criminal. On many occasions web pages will have an overzealous amount of keywords to lure users who would usually not go to these sites. Porn sites used to be notorious for this mischief.
To combat the above problem, search engines now rank pages higher based on page popularity. Meaning, if a page is linked to from twenty other sites related by topic, it will rank higher in the result set than a page that stands alone.
Related to the robots meta tag is the robots.txt file. It should be placed in the root of the website. Search engines indexing the site will attempt to read it, if available they will parse the contents of the file which contain directives on which pages and directories to include or exclude from indexing. There is no guarantee however that these hints will be respected by any particular crawler. Many resources exist covering the robots.txt standard, a good place to look is this page.
A great way to learn about SEO is to look at how other sites do it. There was a recent competition on the subject, the objective was to gain number one ranking on Google for a particular nonsensical phrase. Press coverage is also available.