Since there seems to be no document documenting the pros and cons of server-based approximate match in the DNS, I have caused one to be written. In order to be specific about it, the draft proposes a specific scheme on how to do this, but the author's list of problems with this approach should be relevant to many ways to do server-side approximate matching with the DNS. The draft is draft-vrang-idn-cname-00.txt Comments welcome. Note: the author knows that there may be intellectual property issues with the technology described in this document.
