At first blush, I don't see anything wrong with looking at it that way.

But then I'd have to ask - presume that a company passes this accreditation: what then? They propose a TLD and into the root it goes?

And how do you measure harm? And who gets to call foul? I can forsee a case where for every new proposal to come along, competitors will do their very best to find as much harm as possible.

And while I'm here, allow me to be a little selfish and remind you that there are still "applicants" from 2000 who were not turned down, but told that their applications would remain pending. Regardless of how we proceed (and I'm probably the most willing to try anything fair), there are still some of us waiting.

In my case, with .Web, I've been waiting for almost ten years now.

Anything I can do to help the process, you have but to ask. From the standpoint of the person who's been waiting the longest, I suspect I have a unique perspective. Feel free let me contribute in any way I can.

Christopher Ambler

I suppose I don't know what "semantic harm" means. I remember in 2000, SITA ended up with .AERO because some folks thought that ".AIR" was too big a term to belong solely to the airline industry. Is that what you mean? Or does this also include trademark issues and country name issues. It's a big world, and I think you'll be hard pressed to find a text string that doesn't strike someone, somewhere as inappropriate. Some Board members once even considered a nonsense string like .III inappropriate because it didn't mean anything and was difficult to pronounce.

I'm skeptical that you could (a) come up with an agreed definition of "semantic harm" and (b) apply it in an objective way.

-- Bret

I entirely agree that an accreditation along the lines of the way that CLECs are approved in the US makes a lot of sense.

But I also have to agree with Bret that you will never, ever, get any agreement on "semantic harm". The IP crowd finds harm in any new domain since they see it as yet another place for squatters. You'll have to drain the IDN swamp and decide what to do about a domain that uses unicode characters that look exactly like COM, or maybe not exactly but pretty close like C0M. Is it a problem if two domains mean exactly the same thing like COM and BIZ? Approximately the same thing like AERO and TRAVEL?
I don't know the answer to any of these questions and neither does anyone else.

I put some notes in my blog on the STLD renewals noting that STLDs to date have all gotten only 5% of projected registrations, so they're all failures, so I don't see any point in doing more of them. But do we need more fully generic TLDs?

Why ask if we "need" more? Do we need more fast food drive-throughs? Do we need more lamp stores? Do we need more banks?

Let those who wish to compete do so. Otherwise, all we've done is create a situation in which restrained trade gives a closed market to existing registries like Afilias and Verisign.

I agree with the statement that accreditation is preferable to approval. It would be a major step forward.

This comment is from Milton Mueller, who was unable to log into Blogware [sorry, Milton]

Susan:

First, you need to question your basic assumption. Why is "semantic harm" more of a problem at the top level than at the second level? We still seem to be laboring under the dot com fallacy - the idea that whoever gets a TLD is going to automatically inherit the gold rush of 1995-1999, and that millions of registrations will ensue as soon as a name is assigned. And yet we all know that isn't true, new TLDs are at a severe disadantage relative to established ones. In a second level DNS, anyone can get virtually any string and no one really cares unless it violates well-established, legal rights such as a trademark. And those concerns are always raised after registration and use, not ex ante.

That being said, here's a constructive proposal. Lee McKnight and I gave this a lot of thought when proposing a TLD process almost three years ago now. The solution is fairly simple. the first thing you need to do is set an upper limit on the number that will be added each period. This maintains the hierarchical strucure of the DNS while also ensuring that no artificial limits on supply will occur later. Then you let people propose any name they want and bid for it. If you want, you can segregate the process by commercial and noncommercial, assigning commercial ones via auction and noncommercial via lottery

Wrt to "semantic" issues, the solution involves two things. First, "semantic" problems should only be raised AFTER a TLD string has been won in an auction or lottery. I.e., the selection from among competing applicants should not be based on semantics at all, but on bidding or luck, both of which are quite objective, transparent and straightforward.

Second, any claims of semantic harm must be based on well-established criteria for maintaining exclusivity, NOT vague claims of "harm" based mainly on lobbying and noisemaking and well-poisening. This means, in effect, that claims of harm must be based entirely on the kind of criteria associated with trademark protection: consumer confusion, fraud and deception, passing off. Competition should explicitly NOT be recognized as a form of "harm."

ICANN should also show a little modesty (for a change) and recognize that it does not have to, and in fact cannot, solve all potential social problems that might arise indirectly from the assignment of a name. Let courts and other rules systems sort out the problems they were designed to sort out.

ICANN cannot and must not get into the business of deciding who is the worthiest person to get a name based on the semantics. ICANN shouldn't give a damn who gets .god, etc. It is a domain name assignment coordinator.

[MM]

Thanks to everyone who commented today. To sum up: There seems to be general approval (in this small group so far) of the idea of accreditation, but discomfort with defining and implementing a policy that defers to strongly-held views about particular strings as part of the accreditation process.

I'm sure there are other views out there, and I'm interested in hearing them.

I'll keep posting new TLDs inquiries from time to time.

Susan