Since you are using XPath, you can try with something like
links = page.search('/html/body//a[@href=http://www.joe.net]')
Explanation: The brakets let you specify a filter for your search and
you can use @ to access an attribute inside an element.
Also, for more complicated stuff you can just traverse the array that
the method search returns like so:
links.each do |link|
if LINK MEETS CRITERIA do stuff