SEO and RIA: The Flex Directory Experiment
DIGG IT!
4
Comments
Published
Monday, January 21, 2008
at
2:05 PM
.
I have been running an experiment using XML/XSL/Flex and attempting to get search engines to spider the model of the application. This morning Google listed the "Flex Directory" under searches containing the body text of the model data. From what I am seeing the spiders have ignored the attribute data as expected. The lesson is to expose the most important elements of the model data and hide data in attributes that would otherwise hinder the results.
So the next step is to reorganize the data so that the directory gets far broader listings and add in the elements of cross-linking using <a href="http://www.blogger.com/url">COMPANY</a>. This will force the spiders to spider the resulting sites via cross-link and reinforce the network relationship via domains.
Proof: SEO and RIA work seamlessly together!
Here are some links for review:
Farata effectiveUI flex
iChameleon ASI
Cynergy Memorphic
Cynergy Flex Sanative
More to come!
Woot,
Ted :)
You recommend that developers "[..] hide data in attributes that would otherwise hinder the results."
I'm curious to know what sort of data would be hindering search results? Isn't the goal to tell Google as much as you can?
Perhaps the attributes would be more helpful (less of a hinderance) if they could be marked up in such a way that search engines will understand their relationships.
For instance, you could use the hCard microformat [1], which was specifically designed to represent information about people, companies, and organizations. It also has the advantage of being visible as plain text in a way that search engines will understand.
hCard is a step toward more semantic searches as well: for instance, Technorati [2] allows you to search for "contacts" across the entire web by recognizing the hCard microformat.
Very cool to see the results as this experiment moves forward Ted, you've clearly put a lot of effort into it!
[1] http://microformats.org/wiki/hcard
[2] http://technorati.com/weblog/2006/05/108.html
Great experiment there ted :)
I don't think seeding the search engines with meta data based around a RIA's model is the hard part. Telling the Search engine "I exist and this is what I potentially have inside me" is relatively simple (which you've indicated).
Having the said index reflect points in time within the RIA, now that's the where things get a bit tricky and where early planning in terms of state management apply.
I'd love to see a followup to this post where you've told the search engines you exist, now combine this with deep linking and you have officially solved the Flex SEO puzzle ;)
Great work ...
-
Scott Barnes
RIA Evangelist
Microsoft.
I don't see how this is better than stuffing content in the same page with the embeded flash object. In fact, it's worse because you have to load the model page twice. Maybe you could enlighten us with the motivation behind your design.
I'm looking forward to the next step. Using more open (and valid) HTML is going to do allot for the accessibility of this data, let alone it's search engine visibility.
At the moment the setup seems to be working against Google - following one of those searches presents you with a blank page (for those without Flash, JavaScript). Accessing it from my mobile actually asks me to download something (and fails).
It's certainly an interesting technique, I wonder how long it'll be before black-hat SEO practitioners pick up on it.