Titled Why i-names? I think they work for my people., Kaliya Hamlin makes a good case for i-names.
Personally, I think it will take some time before we all know whether i-names are going to fly, and if so, for whom and for what application. Good arguments are being raised against, as are being raised in favor. But unlike some others, I don’t think it would be a good idea for anybody to dismiss XRI/i-name technology out of hand, just because their syntax is unfamiliar, the documentation is a bit "obtuse" (to quote from a recent post on an OpenID mailing list) or because i-names have an actual business model that, surprise surprise, involves the actual exchange of actual money — what a concept ;-)
Also, I know for sure that at least I have learned a lot already from smart people like Drummond Reed and the technology they constructed: for one, we wouldn’t have the very elegant Yadis but instead a much more rudimentary version; something that should be acknowledged even by those who won’t ever touch an i-name.
After all, if you don’t like i-names, simply don’t use them! And by virtue of XRI proxy resolution, the only thing you need to do to support i-names in your own relying party software is to prefix them with http://xri.net/ and treat them as a regular OpenID URL. That doesn’t sound like something that’s too hard …