Tuesday, June 19, 2001 06:43 PM

Hi Ben:

First, I think it bad practice to mine a whois database and publish registrants' names on a website even though the information is available to you. In my opinion, it is as bad as spam to take the information and publish it in this manner, especially the number of domains held by any one entity (naming that entity) as part of a table. If this were done with the .com database, I think NSI might have a problem with it. As a registrant, I would have a problem with it, and would attribute any increase in spam to that publication. I can see now the importance of establishing privacy bureaus to avoid this type of publication.

It is a bit disturbing to see that you have assumed that registrations began with the formal launch of the automated registration system (POSSR) developed by The PacificRoot.

I think it is important to know that we took over management of the .BIZ tld in May of 2000 and any registrations at that time were done manually. When we homed .biz in the PacificRoot it was because the POSSR registration system was being developed. It went through the normal development processes and was formally launched after beta completion in October, 2000. Before and during beta we had requests for domain names and entered them manually. All were live at the time of registration and many were at no charge during the early phases, as we did not feel it proper to charge registrants while development was in progress. However, the registry was open to public registration at all times, regardless of methodology. We informed those who requested registrations that their domains were live, but that the online system was a few months away.

Your research results are based on the assumption that no registrations took place prior to the launch of the automated registration system, so all the dates you have listed are not related to actual registration dates. They were all date stamped after the formal launch of the online system. The oldest registrations are actually months prior to that launch date.

In addition, you noted that ARNI itself holds almost 200 registrations. While this is apparent (they are in one account), it does not take into consideration that these were requested by different entities and were entered in the ARNI account as a courtesy. Each entity must set up an account before the domains can be attached to that entity in the database. They are, indeed live, and available to those who requested them way back in May of last year.

Some of the domains registered in early summer were unitedservingafrica, unitedservingasia, bridgecompanies, bridge, business, @quasar... There are approximately 140 in the ARNI account registered between May and October 15, 2000.

While I realize that one would have to actually look at the domain names to realize they were not ARNI domains, one would not necessarily know who requested what names or when they were actually registered. Those results would change your findings considerably.

The second version of POSSR is now in beta and many of the domains held in the ARNI account will be separated into their own accounts when the new release is complete.

Your research correctly shows a steady increase in registrations over the next few months. We do not track who is registering domain names. The registration system simply recognizes accounts and character strings. It makes no value judgements. Since there is no limit to the number of domains that can be registered by any one entity, it is not something we monitor.

The tables also show the decrease in registrations due to the damage caused by ICANN's announcements of the duplication of .BIZ. That may not hold, however, with our plans for the future.

I think it necessary to point out that these statistics are indicative of very little with regard to the concerns over collisions. We have been managing .BIZ for a relatively short time, but well prior to any hint that .BIZ might be a consideration for ICANN. It is only due to ICANN's duplication of the TLD that we have had to alter some of our plans. We did not anticipate the acceptance of applications by ICANN for duplicate TLDs and really did not expect them to actually select a deliberate duplicate after having been notified many times of its existence.

The concerns raised regarding problems with email, hostnames, ftp, and http queries are quite valid. With an increase in registrations and content, this problem will, of course, increase. It really doesn't matter if it is one hundred or one thousand or several thousand duplicates, Ben. It will produce problems at any number. I don't think "relative" concerns are appropriate here. When a registrant's email is received by an unintended recipient, the damage is damage. When hostnames for servers have two different IP addresses and queries are made returning incorrect information, it is incorrect information. This does not have to happen and the market leader should not deliberately cause it to happen. Relative numbers have little or nothing to do with it.

ICANN has attempted to rationalize its actions by saying 1) that there are separate name spaces, then 2) they are responsible for only their root. Now both ICANN and Berkman are indicating that since the numbers are small, they are insignificant. Neither takes into account that future numbers may not be so small or that it is a problem regardless of root or numbers.

Had anyone from Berkman consulted with us regarding your research, we would have been happy to cooperate. Unfortunately, the results are subjective because of assumptions made of start dates and methods of registration. It is disappointing to see this come from Berkeman and presented as objective data when applied to the problems of duplication in the name space. I would expect it from ICANN, but not Berkman.

Leah Gallegos