Translations of this article are available.
This section isn’t really about web standards or accessibility, but it’s here because the way a URL is constructed can have a great effect on how well a website is indexed by search engines, and how usable it is to its visitors.
Some search engine robots don’t follow links to URLs that end in a query string. This kind of URL is very common on websites that dynamically get their content from a database, and may look like this:
http://example.com/products.asp?item=34627393474632&id=4344
The easiest way to construct a URL that is better for both search engine robots and humans, is to change it to look like it is pointing to a directory. The above example would then be changed to look like this:
http://example.com/products/item/34627393474632/id/4344/
The web server then interprets the new URL and internally converts it back to the original URL, complete with the query string. At the end of this section are some links to sites where more info on how to do this can be found.
An even better, but somewhat more complicated, way of changing URLs is to completely rewrite the visible URLs to make them human readable:
http://example.com/products/flowers/tulips/
The advantages of using this kind of URLs are that search engine robots will index the site better, it becomes easier for humans to read the URL, and you avoid revealing which server technology you’re using. Since the URLs don’t contain server specific file extensions, like .asp, .cf, .cgi or .jsp, this will also make it easier to change the technology used on the server, should that become necessary.
If you choose to use query strings in your URLs, it’s important to encode any ampersands, &, to their HTML entity, &. If you don’t, some web browsers will, as they should, see the ampersand as the start of an HTML entity, and if the text following immediately after the ampersand matches an HTML entity, the browser will convert the URL, and in most cases break the query string.
Another thing worth mentioning is that for most websites, using a www subdomain is unnecessary, and that http://example.com/ should be used instead of http://www.example.com/. More info on this can be found at no-www.org. Whether you use the www subdomain or not, it is a good idea to configure your web server to redirect any traffic to http://www.example.com/ and http://example.com/ to the same URI.
An article that explains the problems with certain kinds of URLs, and how to make your URLs better.
An article on why ending URLs with “/” is good.
A collection of links to articles and tutorials on URLs.
A longer explanation of the problems with unencoded ampersands in query strings, and a test case.
Comments, questions or suggestions? Please let me know.
© Copyright 2004–2006 Roger Johansson