Just for fun, and because I couldn't make it to OpenIDDevCamp, I thought I'd go back and recount history of the XRDS/XRDschema/concept. The use of the XRD/XRDS for stuff outsideXRI was proposed at the IIW 2005 in Berkeley (in what became YADIS), butthere's a ton of earlier history...
Before that, I think we had been focusing more on using DNS records forresolution (specifically DDDS or SRV or other mechanisms). That documentformat described what we called at the time "Local Access" serviceendpoints. They were basically REST endpoints. This was basically thedocument format that ended up in XRI 1.0... very tailored to the XRIResolution mechanism and limited to the concepts within XRI (which wereadmittedly a little unique...) Trusted resolution soon followed (a chain ofdigitally signed XRD's to ensure the resolution results were authentic).
The XRI syntax and XRI resolution mechanisms were put into separatespecs, not because we thought the resolution spec would be used outside XRI,but rather because we anticipated multiple resolution mechanisms for XRIs and wantedto keep the two specs separate to allow them to evolve independently.
There was a long and (one of many many many) heated discussion about whetherXRI resolution services should be described in the same way as otherservices - that is, whether a Service element should be used to describeendpoint services as well as XRI resolution servers. We had, for a longtime, treated those as very different things described with separate elements. Thefinal big change to XRDS/XRD was to make XRI authority resolution "justanother service" from the point of view of the XRDS/XRD schema definition.It wasn't until Nov 2005 that the XRI authority was "demoted" to "justanother service" in working draft 09 of XRI Res 2.0: http://www.oasis-open.org/committees/download.php/15310/xri-resolution-V2.0-wd-09.pdf