Wildly Appropriate

Feeding Not Spidering: SEO for AJAX?

Feb 2, 09:38 AM

I was checking out Scott Abel's excellent The Content Wrangler blog this morning and came across a link to something called ROR - an unfortunate acronym that's always gonna be misunderstood as Ruby On Rails. ROR in the present context is supposed to stand for "Resource of a Resource  and is explained by its inventors at rorweb.com as:

[A] rapidly growing independant XML format for describing any object of your content in a generic fashion, so any search engine can better understand that content....<snip>

ROR promotes the concept of structured feeds, enabling search engines to complement text search with structured information to better understand meaning. ROR information is typically stored in a ROR feed called ror.xml placed in your website's main directory. Unlike Google Base, ROR feeds can be easily accessed by all search engines: at http://www.your-website-name.com/ror.xml

\You can think of your ROR feed as a powerful structured feed for describing all your objects to the search engines: products, services, reviews, discounts, images, events, schedule, podcasts, anything you want.

And get this: ROR says their implementations pre-date both Google Base and Google's Sitemaps.

I've been watching Base and Sitemaps for a while now as being portents of the Next Big Thing in search. I'm the proud owner of the as-yet-unused domain SEO-for-AJAX-BOOK and a big part of the work I'll be doing if I ever write when I write that book will likely be in this area where we create and serve-up structured and standards-based feeds into the search providers instead of or as a way of augmenting crawler-based indexing.

Elsewhere on this blog I talk about Web 2.0 and the evolution from single, discreetly-addressable .html pages toward interfaces that can exist in a nearly infinite number of "states." Probably, when these infini-state UIs use AJAX and XmlHTTPRequest to grab stuff from the server based on what the customer is doing, many of those states are going to be invisible to search engine crawlers. That's where the ROR or Google Base frameworks come in to save the day. Uncrawlable presentation layers are no problem when you can feed a discreetly addressable "landing state" of that otherwise uncrawlable flash site or AJAX-o-rama to the search engine in a standardized feed and mainline all of your product attributes and addresses into the indexes.

As many of us start prying open the Pandora's Box of heretofore forbidden web interactivity, I hope somebody will be able to do so with a budget and timeline that permits some testing of states-based UIs and search engine visibility and the degree to which approaches like ROR and Google Base mediate the enormous risks that AJAXy interfaces pose to spider-based search engine visibility.

Geeks Only:

I used the free tool at rorweb.com to create my own ROR file for this blog. It's linked here. I'd have to modify TXP to generate this file for me if I wanted to be ROR-happy on an ongoing basis... Update: there's a swell dynamically self-updating Google Sitemaps plugin for TXP that I'll be using instead of ROR until an ROR plugin gets made. And isn't it super annoying that the by-the-book method for hooking up the ROR means users will see the ROR link as a livebookmark in a way that invites inadvertant errors choosing between ROR and RSS?