From: owner-wg-c-digest@dnso.org (WG-C-DIGEST) To: wg-c-digest@dnso.org Subject: WG-C-DIGEST V1 #12 Reply-To: Sender: owner-wg-c-digest@dnso.org Errors-To: owner-wg-c-digest@dnso.org Precedence: bulk WG-C-DIGEST Tuesday, February 15 2000 Volume 01 : Number 012 ---------------------------------------------------------------------- Date: Mon, 14 Feb 2000 10:24:56 -0800 From: "Christopher Ambler" Subject: Re: [wg-c] STRAW POLL > The winner of each is given the right to select a TLD name, in case of > conflicts, the earlier winner gets priority. Does that mean that if I win in the first 6, and AT&T doesn't, I can select .ATT?! Fantastic! Christopher ------------------------------ Date: Mon, 14 Feb 2000 10:45:44 -0800 (PST) From: Karl Auerbach Subject: Re: [wg-c] nine principles for domain names Regarding principles or rules: I perfectly expect that a legitimate and reasonable proposal for a new TLD will be some sequence of characters that make no sense in English or any European language but make complete sense in some Asian language. Since the suggested principles/rules all carry a strong aspect of linguistic and cultural predetermination, I suggest that they are not particularly useful principles/rules for a global, multi-cultural, multi-lingual, internet. As for the pricniple/rule of "Trust" - Why? That seems to be to be based on the presumption, an incorrect one in my mind, that TLDs are assumed to be marks of some good or service. I hardly see the DNS system as an appropriate vehicle for some sort of guarantee or impramateaur of quality or safety. To my mind these principles/rule are "chartered-TLDS-lite". --karl-- ------------------------------ Date: Mon, 14 Feb 2000 12:31:05 -0800 From: "Josh Elliott" Subject: RE: [wg-c] STRAW POLL Good points. Someone needs to define "charter". It is my view that all current TLDs have a charter that is implied in RFC 1591. Now, the charter may no longer be followed (as with .org or .net), but it was outlined there. While you may construe my point as borderline between options #3 and #4, I am sticking to option #3. Option #4 will never fly. Remember that the Board still has to vote on anything we come up with here whether we like it or not. #3 is the better alternative, and the beginning of a compromise among the variety of interests on this list. It is imperative that this group make some progress prior to Cairo. Progress means compromise. This goes for everyone. If we keep rattling off the same opinions over and over again, we are getting no where. This prolonged turmoil only leads to further delay in the creation of new gTLDs. Can't we all just get along? :-) Thanks again for your valuable input. Josh > -----Original Message----- > From: owner-wg-c@dnso.org [mailto:owner-wg-c@dnso.org]On Behalf Of > Milton Mueller > Sent: Monday, February 14, 2000 9:34 AM > To: Philip Sheppard > Cc: wg-c@dnso.org; Jonathan Weinberg > Subject: Re: [wg-c] STRAW POLL > > > These comments address points made by Josh as well as Philip > > Philip Sheppard wrote: > > > 1. All new gTLDs must have charters that meaningfully limit the > universe of > > people who can register in those gTLDs. > > > > BUT TAKE HEED Limit is not the right descriptive. A charter > need not be a > > restriction! The key is not limitation but differentiation. Dot > biz could be > > fine is it can differentiate itself from dot com. > > Philip:This is a crucial issue for many of us. As a matter of > market reality, no > business is going to run a registry if it cannot differentiate > the service from > its competitors. What has made this a stumbling block for many of > us is the > question whether ICANN defines the "charter" or the marketplace > defines it. If > the charter is imposed top-down as a kind of contract with ICANN, > then practical > issues concerning how the criteria are applied and enforced > arise. So I think > you need to clarify your position. When you say "all TLDs must > have charters" > what exactly do you mean, and where do those charters come from? > > Josh: > It may be a bit confusing terminologically to say that *any* criterion > constitutes a "charter," even one like: "totally open." Most uses > of the term in > this discussion have used "chartered" TLD to mean restrictions associated > semantically with the TLD string, such as .mil. There's nothing > wrong with your > usage, in fact I think it's probably better to use the term the > way you are than > the way it has been used. But again, it doesn't move the WG > forward in the way > Jon Weinberg is trying to do. Jon is trying to define positions > that are clearly > differentiated from each other and find out where people are on > that map. If you > believe that > ANY tld by definition has a charter, and you select Option #3 > under Question 1, > your actual position may not be that different from someone who > selected Option > #4. > > You (Josh) wrote: > "I wish the registries themsleves could take up these issues, I > don't see how it is possible. If ICANN approved registry A, B, C > and D, but > all three wanted to create .EXAMPLE, who gets it? " > > Any proponent of Option 4 knows that registries probably will > submit mutually > exclusive applications. Those can be resolved in a number of > ways: auctions, > lotteries, merit determinations, first come first served. I > support rationing > methods that are non-discretionary (auctions, lotteries, FCFS) > and oppose merit > determinations because it turns ICANN into a politically-driven > regulatory body. > > > ------------------------------ Date: Mon, 14 Feb 2000 12:33:08 -0800 From: Tim Vienneau Subject: RE: [wg-c] STRAW POLL Greetings, WG-C STRAW POLL QUESTION ONE Please select from the following possibilities, *as applied to the deployment of new gTLDs in the name space over the medium to long term*: 3. ICANN, in selecting new gTLDs, should approve some chartered gTLDs and some unchartered ones. (Alternatively, ICANN should require that all gTLDs have charters, but it should approve some gTLDs with charters that meaningfully limit the universe of people who can register in the gTLD, and some gTLDs with charters that do not impose any such limits.) QUESTION TWO Please select from the following possibilities, *as applied to that initial rollout*. 3. ICANN, in selecting new gTLDs in the initial rollout, should approve some chartered gTLDs and some unchartered ones. (Alternatively, ICANN should require that all gTLDs have charters, but it should approve some gTLDs with charters that meaningfully limit the universe of people who can register in the gTLD, and some gTLDs with charters that do not impose any such limits.) QUESTION THREE The issue of chartered gTLDs is tied up with the larger issue of how ICANN should select new gTLDs -- in particular, whether (a) ICANN itself should be the final arbiter of new gTLDs' names and charters, or (b) ICANN should simply select new registries and leave the choice of names and charters to them. 4. Each would-be registry proposing a new gTLD applies to the Names Council (or to ICANN directly) for approval; if the gTLD is to be bounded by a charter, the applicant must supply one. If the application is approved, the applicant becomes the new registry, subject to its proposed charter. ------------------------------ Date: Mon, 14 Feb 2000 21:20:58 -0500 From: Eric Brunner Subject: [wg-c] Sheppard/Kleiman in WG-B I'm glad that the proposal is over in WG-B, and not here in WG-C, as it is an awful mix of policies concerning ASCII labels in the root, ASCII labels in the namespace subordinate to the root, registry policy models, inter- registry policy models, and wishful thinking. I'd like to be more charitable, but quit simply issues like the model for the registry-registrant relationship ("trust" in Sheppard/Kleiman), the model for the inter-registry relationship ("differentiation", "competition" and "diversity" in Sheppard/Kleiman), the model for query resolution and character encodings ("findability" and "semantics" in Sheppard/Kleiman) and the model for the inter-registry registrant relationships ("honesty" in Sheppard/Kleiman) is at least four distinct types of bruised fruit stuffed into one torn and scuffled sack. I left out "multiplicity" as the number of persons in the marks areas of interest who've suggesting, even in private, that the "as needed" test can be met is exactly one -- and in the spirit of compromise I've compromised my sense of irony and humor. The "semantic" claim should only be advanced by those who knew what ".com" ment before picking up a browser in anger, and who understand that for the present, DNS labels are encoded in ASCII. The "findability" wishful thinking should take note of the work that is now in the Common Names Resolution Protocol WG of the IETF (this must be the Nth time I've mention this, but who in WG-B reads WG-C for technology?) The "trust" and "honesty" claims would carry more weight if they had to do with something the end-users actually care about -- public key infrastructure. "fraud" at the level of names is quasi-comic, compromise of the CA and all that can be strongly authenticated by derivation from the CA, now that is real money. Unfortunately, for marks people, the mark is the value to be transacted for, so credit card fraud in the attempted acquisition of marked goods is no crime -- people should be satisfied with marks -- those are the real goods. I suppose that the "simplicity" point is to be understood as a gTLD (registry operator) should not impose an overly bureaucratic procedure on a registr(ar, or registrant). Cheers, Eric ------------------------------ Date: Tue, 15 Feb 2000 09:47:50 +0100 From: "Philip Sheppard" Subject: Re: [wg-c] nine principles for domain names Karl, you comment on two of the principles: 2. Semantics – a gTLD should be meaningful in a language with a significant number of net users. Read this one again - we say nothing about English or a European language. 1. Trust – a gTLD should give the net user confidence that it stands for what it purports to stand for. You say why trust? Are you happy then to say to net users "I know it says X but it means Y. Didn't you know that dear? We all did." Philip ------------------------------ Date: Tue, 15 Feb 2000 09:52:13 +0100 From: "Philip Sheppard" Subject: Re: [wg-c] STRAW POLL Milton, I tried to define our position on how a gTLD differentiation (which you called charter) might work (i.e. is it top down or market led) in my answer to Question 3: "A registry proposes a new gTLD and the NC judges it against the above principles. If it passes it happens. This leads to the market proposing names it wants but there is a first-mover advantage as new names will exclude others. For example: If dot cars was accepted for "everything to do with automobiles" then another party wanting dot autos for " everything to do with automobiles" might not be accepted unless they offered something to differentiate." Philip ------------------------------ Date: Tue, 15 Feb 2000 10:18:06 +0100 (MET) From: Elisabeth Porteneuve Subject: [wg-c] test test, please ignore ------------------------------ Date: Tue, 15 Feb 2000 01:21:07 -0800 (PST) From: "William X. Walsh" Subject: Re: [wg-c] STRAW POLL - -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 15-Feb-2000 Philip Sheppard wrote: > Milton, I tried to define our position on how a gTLD differentiation (which > you called charter) might work (i.e. is it top down or market led) in my > answer to Question 3: > > "A registry proposes a new gTLD and the NC judges it against the above > principles. If it passes it happens. > This leads to the market proposing names it wants but there is a first-mover > advantage as new names will exclude others. For example: If dot cars was > accepted for "everything to do with automobiles" then another party wanting > dot autos for " everything to do with automobiles" might not be accepted > unless they offered something to differentiate." Why give them what is essentially a monopoly on automotive related domain names? Force the .auto and .car TLDs to distinguish themselves competitively, like any other industry. Compete on price, service, etc. Don't get rid of one kind of monopoly only to create another. One other thing that bothers me is that you place the registry into the business of policing content in these cases. Please, think about this. I have a personal domain, one I own in my own name, that is used for a variety of purposes. It is used to provide email addresses to friends and family, to host an ad server for serving ads to various other websites I have, as a test bed domain for web applications I develop, and will probably have other uses as well in the next 6 months or so, some commercial some not so commercial. I think that people are really dropping the ball here, and tying a single threaded web content frame of thought into the domain name system. This is a common misconception some of the corporate interests involved in this process are promoting. And one that really should be revealed for what it is. If I run an automotive owners club from a .auto domain, and I also provide a commercial email service under it, would I be violating these holy charters? If so, that is simply overly restrictive without any basis before justifying the restriction. The DNS is not a certification authority or a content filter. - - -- William X. Walsh DSo Networks http://dso.net/ Fax: 877-860-5412 or +1-559-851-9192 GPG/PGP Key at http://dso.net/wwalsh.gpg - -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.1 (GNU/Linux) Comment: DSo Networks iD8DBQE4qRqD8zLmV94Pz+IRAvo2AJ9zzex5k3bAQWKBNHzJBghIMcFBNwCgtiud ShT25njauu1urzwtJxO6MiU= =FavB - -----END PGP SIGNATURE----- ------------------------------ Date: Tue, 15 Feb 2000 06:58:01 -0500 From: Eric Brunner Subject: [wg-c] Re: nine principles for domain names Philip Sheppard replies to Karl Auerback ... > Karl, you comment on two of the principles: > 2. Semantics <96> a gTLD should be meaningful in a language with a > significant number of net users. > Read this one again - we say nothing about English or a European > language. This answer is non-responsive, failing to address the character set encoding issue. Karl mentioned Asian languages, he could just as well have mentioned Hebew or Arabic, even Inuktitut or Braile. > You say why trust? Are you happy then to say to net users "I know > it says X but it means Y. Didn't you know that dear? We all did." If a FQDN resolves to a dotted quad, then the name-to-address resolution service works, if not, then it doesn't. If a trust mechanism is attached to some part of the resolution service model, then a trust model may be usefully available to end-users. Karl mentioned that the TLDs are not simply marks, or guarantors of some attribute of marks. A responsive answer would clarify the details of the trust model under discussion. Trust is important, so it shouldn't be defined poorly or defended as if it ment the same thing to all people at all times. Cheers, Eric ------------------------------ Date: Tue, 15 Feb 2000 07:56:23 -0800 From: "Roeland M.J. Meyer" Subject: [wg-c] Special characters in DNS names [For some this is a side-bar and this whole thing probably should include name-droppers] WRT native language character sets: I have looked and thought over Bill's ideac wrt NUBIND and other instances of NLS support in the DNS and other infrastructure tools. On its face, it seems to be a good idea becasue it will make the Internet more friendly. However, it will also make the Internet harder to manage and very much more unfriendly to those who do NOT use that character set. The envisioned use; someone in china, using a web-browser, to look up other web-servers in china, is probably a no-brainer. Of course, it will be a benefit. The problem is someone, in france, tracing a spammer, via IP address, using nslookup, get chinese characters as a response. Chances are real good that they won't be able to read the name. It will be a non-answer for them. Like it as not, the International character set seems to be Latin-1. Anything else would trigger a "Babel" syndrome. > From: owner-wg-c@dnso.org [mailto:owner-wg-c@dnso.org]On Behalf Of Karl > Auerbach > Sent: Monday, February 14, 2000 10:46 AM > > Regarding principles or rules: > > I perfectly expect that a legitimate and reasonable proposal > for a new TLD will be some sequence of characters that make no sense in > English or any European language but make complete sense in some Asian > language. > > Since the suggested principles/rules all carry a strong aspect of > linguistic and cultural predetermination, I suggest that they are not > particularly useful principles/rules for a global, multi-cultural, > multi-lingual, internet. ------------------------------ Date: Tue, 15 Feb 2000 08:22:36 -0800 From: "Bret A. Fausett" Subject: Re: [wg-c] nine principles for domain names Philip Sheppard wrote: > 1. Trust – a gTLD should give the net user confidence that it stands for > what it purports to stand for. > You say why trust? I understood Karl's point differently. "Trust" may be critical for a chartered/zoned domain (for example, you can't register a pornographic site in .kids), but it may be meaningless in a true generic TLD. What does trust mean in the context of .web or .123? As phrased, the statement assumes that TLDs stand for something, and some of them may simply be memorable strings, standing for nothing. -- Bret ------------------------------ Date: Tue, 15 Feb 2000 08:47:39 -0800 From: Dave Crocker Subject: Re: [wg-c] Special characters in DNS names Folks, >WRT native language character sets: > >I have looked and thought over Bill's ideac wrt NUBIND and other instances Changes to DNS technology are not part of this working groups's scope, or any other DNSO working group's scope. For that matter, there is extensive pre-IETF work that has been led by Tin Tan Wee and which is in the process of forming an IETF working group. DNSO working groups have a sufficiently problematic track record of non-accomplishment to strongly suggest that we need fewer distractions, not more. We should focus on the work we need to get done. d/ =-=-=-=-= Dave Crocker Brandenburg Consulting Tel: +1.408.246.8253, Fax: +1.408.273.6464 675 Spruce Drive, Sunnyvale, CA 94086 USA Gong Xi Fa Cai / Selamat Tahun Baru Cina ------------------------------ Date: Tue, 15 Feb 2000 09:39:24 -0800 From: Kent Crispin Subject: Re: [wg-c] nine principles for domain names On Tue, Feb 15, 2000 at 08:22:36AM -0800, Bret Fausett wrote: > Philip Sheppard wrote: > > 1. Trust – a gTLD should give the net user confidence that it stands for > > what it purports to stand for. > > You say why trust? > > I understood Karl's point differently. "Trust" may be critical for a > chartered/zoned domain (for example, you can't register a pornographic site > in .kids), but it may be meaningless in a true generic TLD. What does trust > mean in the context of .web or .123? > > As phrased, the statement assumes that TLDs stand for something, and some of > them may simply be memorable strings, standing for nothing. I think his point is that TLDs *should* stand for something, and that consumers should have some confidence in that. The term "trust" is a bit strong, perhaps, especially given that it has, at least to some, technical implications. Note that I don't necessarily agree that all TLDs should stand for something, but on the other hand, I take it quite seriously as a point of view. - -- Kent Crispin "Do good, and you'll be kent@songbird.com lonesome." -- Mark Twain ------------------------------ Date: Tue, 15 Feb 2000 10:58:31 -0800 From: "Bret A. Fausett" Subject: Re: [wg-c] nine principles for domain names >> Philip Sheppard wrote: >>> 1. Trust – a gTLD should give the net user confidence that it stands for >>> what it purports to stand for. > Bret Fausett wrote: >> As phrased, the statement assumes that TLDs stand for something, and some of >> them may simply be memorable strings, standing for nothing. Kent Crispin wrote: > I think his point is that TLDs *should* stand for something, and that > consumers should have some confidence in that. The term "trust" is a > bit strong, perhaps, especially given that it has, at least to some, > technical implications. > > Note that I don't necessarily agree that all TLDs should stand for > something, but on the other hand, I take it quite seriously as a point > of view. I take it quite seriously too. We can see the problems that arise simply by looking at .com, .org, and .net. Once upon a time, they stood for something. But innumerable circumstances changed that and now any distinctions are, for most purposes, meaningless. One unfortunate result is that corporate domain name owners are registering across all TLDs and attempting to enforce their trademarks across all TLDs. That expensive and annoying exercise may have been prevented if .com, .org, and .net had retained some meaningful distinctions. I also like the idea of seeing if consensus can be built around first principles, such as those that Philip and Kathy have floated. But that's going to be difficult if the "Trust" plank is meant to exclude true generic TLDs. I don't know that this was their intent, but we could clarify it by making a change in the wording. Something like: Trust: a chartered or zoned gTLD should give the net user confidence that it stands for what it purports to stand for. or Trust: a gTLD that purports to stand for something should give the net user confidence that the intended distinction between it and other gTLDs is meaningful. -- Bret ------------------------------ End of WG-C-DIGEST V1 #12 *************************