How to make Flex RIA contents accessible to search engines like Google?

Question

How would you make the contents of Flex RIA applications accessible to Google, so that Google can index the content and shows links to the right items in your Flex RIA. Consider a online shop, created in Flex, where the offered items shall be indexed by Google. Then a link on Google should open the corresponding product in the RIA.

Answer 1

There is a massive thread available here:

http://tech.groups.yahoo.com/group/flexcoders/message/58926

But essentially, google already indexes .SWF files (you can test this out yourself by restricting search results to just .SWF files). It can search any text content within the SWF file.

However, if the text information in your site comes from a database / web server. Then it won't be able to access this information easily.

One example of getting this to work is using an XML file as your index page, then using an XSLT transform to render it using Flex. "Ted On Flex" has good information about this.

http://flex.org/consultants

Answer 2

Currently the best technique for making an RIA indexable by search engines is called progressive enhancement (or graceful degradation, depending on which way you see it). Basically you create a simple HTML version of the application using the same data as the application loads. This version should be dynamically generated by some kind of backend server technology. This HTML version can be indexed by Google, but each page also contains a check that determines if the visitor is capable of viewing the rich version, and if so replaces the HTML content with the Flash, Flex or Silverlight application, preferably in such a way that the application starts in a state where it shows the same data as the current page. "Replaces" can mean that it just embeds the application on top of the HTML content, or that it redirects the user to a page that embeds it. The former solution is preferable, because the latter can be considered cloaking.

One way of keeping the HTML and RIA versions of a shop synchronized is to decide on a URL scheme and make sure that RIA uses some kind of deep linking technique. If a visitor arrives to a specific item via a search engine, say /items/345 the corresponding pseudo-URL in the RIA should be the same, so that you can embed the RIA on top of the page and set that URL as a parameter to make the RIA display that same page as soon as it has loaded.

This summer, Google and Yahoo! announced that they would begin using a custom version of Flash Player to index Flash based applications by exploring them "in the same way that a person would". Now, two months later there is still no evidence that this is actually happening. Ryan Stweart had to cancel his Flex SEO competition because it became evident that no one could win. The problem seems to be that event though the technique may very well work (although I'm sceptical), the custom Flash Player needs some kind of network interface to be able to load any referenced resources, like XML data, other SWFs, etc., and this is currently not implemented by Google. This means that for an application that loads all it's data dynamically, like say, all that I can think of, Googlebot will not actually see anything relevant. Yahoo! ignores SWF based content altogether.

Oh, and it just so happens that I talk about Flex and SEO on the latest episode of the Flex show =)