I’ve turned on the Search function, powered by Google search.  I still see some of the old Boyink.com results but overall most of the content seems to be indexed.

But why use Google search when ExpressionEngine has it’s own search engine?

Because - as much as I truly love ExpressionEngine, the EE search engine is sorely in need of updating.  EE has it’s roots in the blogging world - in fact I think one of the things that adds to the EE learning curve for newcomers is the fact that there are lots of references to “weblogs” in the Control Panel and code.  In the support forums I often see (and have posted myself) that you need to relearn what a “weblog” is when it comes to EE, and think of a weblog as a “content storage container” that can be configured to hold content of any shape.  With unlimited custom fields and the ability to create pages that are combinations of weblogs, EE overcomes it’s weblog roots and serves up weblog and non-weblog content with aplomb.

However - when it comes to EE’s search engine, the weblog roots still show.  In order to configure the search results, you need to tell the search engine where the url to the weblog page is, where the weblog comment page is, or a specific “Search Results URL” for this weblog.  It’s a very “inside-out” paradigm - starting from your content you tell the search engine where it should be on the site.

The problem is, often times I don’t have those specific URLs for a particular weblog.  For example - I’ve setup weblogs that provided sidebar information throughout the main sections of a site.  Such a weblog has neither a URL, a comments page, or a particular URL that the search engine should point to. 

What I’d prefer (and have logged a feature request for) is a search function that’s more of an “outside-in” paradigm, where the search engine simply indexes your site as a collection of rendered pages - exactly like Google or any other public search engine does.

While I realize that this would require EE to build an index of rendered pages which would increase the site of an EE site, I wonder if it could be tied in with the existing caching feature.  Such a search engine would be much simpler to setup, would match users understanding of your site structure better, and would likely deliver much better results.