I've tried to implement the approach here:
I've done two things:
- look for the _escaped_fragment_ query parameter and use it to generate a non-ajax view of the site
- added the meta tag in my index.html:
<meta name="fragment" content="!">
If I visit my site with _escaped_fragment_ as a query parameter, a snapshot is correctly delivered to my browser.
However, when I use the Google Webmaster tools to "Fetch as Googlebot", I just get the raw (and content free) index.html text: I never get the snapshot. Maybe that's what I should expect, I don't know. In any case, if I do a google search for some of the words on the first page, I don't get any results.
I've just added some counters to detect the Googlebot (by looking for that string in User-Agent) and the _escaped_fragment_ parameter to see if I'm getting either, but the Webmaster tools do say that my app is being crawled.
If I've just implemented it incorrectly somehow, or if there's a major step I'm missing, that would be much appreciated feedback.