Thursday, August 21, 2008

SEO - Nofollow links by default in Rails

I have been learning quit a bit about SEO these days. And one thing that I didn't realize is that "link juice" leaks from page to page. So in order to conserve you link juice you want to very carefully choose which pages should be the recipient of this precious juice. This actually went against my intuition which was to encourage the "bots" to follow every link that I have on my site.

Since we already have a pretty complex site with many call to link_to, I decided to tell the "bots" to not follow links by default. This is done by adding the "rel=nofollow" attribute to you anchor tags, in rails this is handled by the call to link_to. This can be accomplished by overriding link_to in application_helper:

    1  def link_to(name, options, html_options = {})  
2 if( rel = (html_options[:rel] || html_options["rel"]) )
3 foo if(RAILS_ENV == 'development')
4 elsif( !(follow_link?(options, html_options)))
5 html_options["rel"] = "nofollow"
6 end
7 super(name, options, html_options)
8 end

Ignore that 'foo' part, I just wanted to blow up if we are allready using the rel attribute somewhere in the codebase

Now you can define the method follow_link? to be whatever the rules are for you domain.

    1   def follow_link?(options, html_options)
2 return true if html_options.delete(:follow)
4 url = url_for(options)
5 if(url.match(/(.*?show.*?)|(news$)|(browse$)|(page=\d+)$/))
6 return true
7 end
8 return false
9 end

Now I specifically know which pages are important for the "bots" to index and the precious juice is being preserved.

Friday, August 8, 2008

Bonus Screencast for Rails on Facebook

Geoffery, from PeepCode, helped us put together a screencast to accompany the Rails On Facebook PeepCode PDF. It covers the specific details that you will need to setup the sample application Ruby Footprints. Check it out Rails on Facebook.