Hi Ed;-)... what you say might be true in general, but the ORSC policy is that the composite root should be free of TLD collisions, just as lower levels must be free of collisions. The mechanism for avoiding collisions should be one of coordination and not one of dictation by single points of control which are at best very controversial and arbitrary, and which summarily create collisions when they know that they are doing so. ORSC simply provides a root of greater value just because the ORSC root inclusively contains all the known public accessible TLDs, while the ICANN root exclusively only contains those TLDs which ICANN has given sanction. And, at this point there is one extant collision which ORSC is not able to resolve for the simple reason that ICANN refuses to even acknowledge the fact of the collision that ICANN has knowingly created. If anyone here has any good ideas for ho anyone outside ICANN might successfully negotiate a resolution with ICANN, please provide your advice, and ORSC will give it due consideration, but if your suggestion is to take ICANN to court, please understand that you will be asked to help by sending money for legal fees. At $1.5+ million per year for its legal expenses, ICANN clearly has an advantage over ORSC which relies on voluntary support, while we all contribute our domain name taxes to ICANN for our ICANNIC DNS names;-)... Best...\Stef At 3:14 PM -0700 7/29/02, Ed Gerck wrote: >All: > >The DNS is indeed defined as a hierarchical tree and thus must >have only one root. It will not work with more than one root. > >What Richard and Stef are saying, in short, is in agreeement with >that in the sense that they recognized you would need an *additional* >control structure in order to use multiple roots AND the DNS. > >But, adding multiple roots will NOT do away with the single root -- >as Brian noted. That additional control structure for the "multiple >roots" must still be controlled by someone -- and here is where we >go back in a circle, Brian notes. > >However, what Richard is saying would create a delegation mechanism >to the root, for those who want to use it. Those who do not want to use >it, would still see only the DNS root. Richard is not solving the DNS >control problem, he is just proposing a way to create a delegation >mechanism that is entirely optional for the user to see. > >That said, this logic implies that operators of root A cannot complain if >operators of root B define a TLD that conflicts with a TLD in root A. >Thus, operators of an alternate root cannot complain if the DNS operators >add .BIZ to the DNS TLDs and control it. An alternate root operator could >also add .COM to their namespace and control it. Operators of intranets >do it all the time. > >Cheers, >Ed Gerck