From: owner-nanog@merit.edu [mailto:owner-nanog@merit.edu]On Behalf Of Rodney Joffe Sent: Monday, February 21, 2000 12:19 AM
William Allen Simpson wrote:
Would it be OK with the rest of us for Rodney Joffe to create a database of all the requests and answers made thru geektools?
Not sure what you mean here. GeekTools SuperWhois is merely a proxy. When answers are returned, they are still subject to the copyright under which GeekTools retrieves them. We don't cache or store them. We just go get the answers, and display them, copyright and all (NOTE: we have received permission to do so). But we do not have permission to store or capture the data. And, as we've already discussed, the data is wrong sometimes :-)
It is arguable whether or not we need the permission, for the data. Copyright covers implementations of works, not the data on which the implementation is founded. Reformatting the data into something like RIPE's format, or something wholly different, would be sufficient. The data itself is still paid for by US tax dollars. There have been other arguments put forth, that view it from a different direction entirely, that the data belongs to each registrant and not to NSI. Basically, you own all the data pertinent to you, as does each registrant. In either case, the purveyor of data is not automatically the owner of the data, only the form that they present that data in. There is substantial case history for this view, but BEWARE:IANAL.
A totally new from scratch database needs to be created. And it should *not* be me, GeekTools, or CenterGate. Or any individual. It needs to be controlled by a body trusted by all. What ICANN should have and could have been.
There are two points here; One, what to build and 2), whom shall build it. The first point begs an issue, database synchronization. It is the same issue that plagues SRS and is the prime reason that they are having problems in the first place, that is the prime inadequacy of the SRS architecture and design. Anyone whom has ever worked with Sybase Replication servers and Oracle Parallel Servers, knows first-hand what sort of nightmare synchronization can be, especially over high-latency links and high transaction rate end-nodes. That was my biggest concern with Crispin's original design, over three years ago(that was the process box which was labeled "a miracle occurs here" <grin>). Obviously, NSI couldn't/wouldn't fund the miracle. Most of the problems you see is data-dithering, a classic symptom of synchronization problems. Having yet another authoritative database only aggrivates the problem. I understand trying to loosen the coupling, but that's the wrong way to do it, IMNSHO. The second point is unrealistic, ICANN has no money, no technical competency, and no desire. All they want is control, as badly as the DOC/NTIA wants to relinquish it. ICANN was the ones that forced SRS down NSI's throat in the first-place (w/ USG twisting both arms), as a means for generating revenue for ICANN. They don't even know, or care, that SRS has problems. ICANN makes a healthy fee for accrediting registrars.
CenterGate can provide the repository infrastructure through UltraDNS. UltraDNS already has a whois component that allows a dns user to specify which fields may be seen by the public, and which may not. Some thought has to go in to how the 'net decides who should be able to read something, and who should not. The major cause of complaints, and bogus data (555-1212 phone numbers and no-valid-email@ addresses) is the spam issue. In fact, I believe that this is where the major effort has to be spent assuming that this community really gives a damn about the subject.
You are correct, there is a serious data integrity issue with the current whois. NSI is to be blamed for letting it get out of hand. That said, there are serious privacy issues in the way that whois works. It is certainly a violation of recent EU privacy standards and is the main reason that spammers were able to abuse the systems in the first place.
So, while this discussion is interesting to some, unless the majority of lurkers here care enough to open their mouths and comment, we're just creating noise. So sad.
Agreed, the time to ready a solution ... is BEFORE the problem shows up!
We would change the Open/Net/Free/*BSD/*nix whois distributions to point at geektools. (Especially as default whois is pretty useless right now.)
whois-servers.net may be more appropriate, and it is already in at least 2 of the BSD distributions for whois :-)
One of the other problems is finding the whois servers out there. However, I think that is another topic. cheers =========================================== R O E L A N D M. J. M E Y E R http://mhsc.com/~rmeyer/resume