Please create an account to participate in the Slashdot moderation system


Forgot your password?
The Internet

IETF vs. ICANN 146

Ian Lance Taylor writes: "Two IETF drafts were filed today which fire a shot across ICANN's bows. They say that anybody who introduces a new version of an existing TLD is destabilizing the DNS--even ICANN. These are still only drafts, mind, not standards. They both acknowledge input from Karl Auerbach, the member of ICANN's board who was elected by North America. The drafts are Alternative Roots and the Virtual Inclusive Root and Root Server Definitions." The IETF drafters are attempting to define a system where non-ICANN TLDs can easily be added. ICANN is set to push their one root concept of operations where ICANN gets absolute authority over internet naming. All ICANN PR is geared toward presenting the ICANN-only plan as being necessary for "internet stability".
This discussion has been archived. No new comments can be posted.


Comments Filter:
  • We'll have to add this to the list:

    "One world. One web. One browser." (Microsoft IE ad)

    "Ein. volk. Ein reich. Ein fuhrer." (Adolf Hitler)

  • by Anonymous Coward on Wednesday May 30, 2001 @09:41AM (#188554)
    ICANN's monopoly on Internet naming created artificial shortages that resulted in the same kind of asinine market conditions which prevail in cities that issue taxi medallions. Cypersquatting and speculation on dot-com names are prime examples. I'm going to register whatkindofidiotpaysamilliondollarsforadomainname.c om and offer to sell it for half a million dollars. But seriously - isn't there a better way to administer the Internet than the ICANN dictatorship?
  • Hey, actually this is a hell of a good idea!

    Instead of you'd have microsoft.sux. Problems of confusion would be directly addressed by saying 'No, of course microsoft.sux is not the Microsoft site- NONE of the .sux sites are the official sites, are you nuts?'

    The cost of registering every last tld is not prohibitive but perhaps it could be _forbidden_ to own every one. Personally I'd be in favor of a 'you only get to own one' rule, but that's very restrictive. Maybe 'business units' could be allowed to own their own name.

    Imagine the possibilities: pizza.hut,,, natalieportman.rok :) The sheer scope of it tends to defuse concerns about trademarked names. So gets to be the actress (most likely), but slashdot trolls of the old school would get to register the .rok because it would obviously be a different story, and most notably because the SCALE of the undertaking makes the idea of 'confusion among TLDs' unpracticable- it would be ridiculous to make claims that because you had .com you had rights to .net and .rok and .sux etc.

    Most of all, the idea would SCALE. Bigtime. Compare that to the 'There's 3/7/X TLDs so obviously we need to own them all' mentality so prevalent today. It's not even about the cost so much as it is a matter of expanding the TLD space to where it makes no logical sense to treat it as an insignificant addendum to the primary namespace. What we need is .sux, and total access to stuff like microsoft.sux, and support for the OBVIOUS fact that microsoft.sux is not You could argue that might be confused with MS because the part that changes is a part normally used to specify sub-categories:, etc. (there's another great one- .etc would rule. What would be better for mailboxes.etc?) But you can't argue that microsoft.sux is confused with MS because the whole TLD will rapidly fill with 'annoy' sites, that is its obvious purpose in life.

    17576 TLDs is a complete winner of an idea! Let's bear that in mind and keep raising it every time the namespace is forced to scale and add new TLDs. Because it is GOING to keep scaling, like it or not- why not go for the full deal immediately?

  • But collision of domains would be disastrous. It amounts to taking control of the internet from ICANN and giving it to AOL. AOL could easily decide to point somewhere else. Stink.

    I frequently see TV commercials that say at the end something like, ``visit us on the web at AOL keyword: foobar.''

    I note that the world has not ended.

    Disclaimer: I've never used AOL, so I don't really know what these ``keywords'' are, but I can guess.

  • Damn, two replies calling your idea stupid, but no one gave the reason. So, to shed a little light:

    EMAIL is going to be a problem for any scheme which puts the choice of root solvers in the users' hands. This is because you do not send email directly to the recipient's email client, or even directly to their mailbox. Email has to bounce around the network-in-a-network of SMTP servers before it arrives at its destination, and each of those SMTP servers needs to use the same namespace to figure out where mail is going.

    The real Threed's /. ID is lower than the real Bruce Perens'.

  • WTF have free market economics and freedom of speech got to do with DNS?

    You're advocating god damned chaos as a good way of managing networked systems on a global scale?


    I think ICANN is fucked up, the current DNS heirarchy is fucked up but making it more fucked up is *not* progress.


  • My take on the whole DNS fuckup that we're currently living with.

  • Let's say I start a company: PlushMicro, and market using the slogan "Soft, Fuzzy Computers".

    I register (yes, this is an absurd, long example) and get sued - why shouldn't I get to keep my domain, simply because the last syllable of my company name and the first word of my slogan, when combined, are the name of another company?

    What if USMicro and SoftToys merge to form USMicro/SoftToys LTD? Shouldn't they be able to get, being that you can't have a / in the domain name itself?

    Just rambling...
  • > and I've decided that for my Internet use
    > it would be acceptable for a Pop-up window
    > to appear

    ... but the internet is not only http, there are
    many servers working in the background which have
    no way to "pop up" a window to ask which you want to mail to...

    But maybe that's the first real use for instant-

    Message from What
    have you meant? Reply with
    1) for <ICANN>,
    2) for <SLASHNET>,
    3) for <KORU5HIN>


  • In other news today, ICANN has announced that it has been acquired by Microsoft for $5.6 billion in common stock and innovation vouchers. "Domain names are intellectual property," said Bill Gates, "and there isn't anyone better at intellectual property defense than Microsoft. Look at us innovate."

    Microsoft then went on to announce that the Internet's "legacy" DNS will be phased out and replaced with Microsoft's own Active Directory Services. This migration will begin immediately.
  • The ICANN document this article refers to in turn refers to RFC 2826, published by the Internet Architecture Board. It seems to me that that is about as close as you're going to get to an official statement from the IETF on this. So it looks to me like it's Simon Higgs (perhaps supported by famed namespace-abuser Richard Sexton []) vs the sane world.
  • by Kiwi ( 5214 ) on Wednesday May 30, 2001 @03:21PM (#188564) Homepage Journal
    The ICANN's fundamental assumption is that there is only one (set of) root name servers, all of which contain the same data set.

    This, however, is not how DNS itself works. The way DNS itself works is like this:

    • The client sends to a resolver a request like,
    • The DNS resolver looks in its cache. It sees if is in its cache. If not, then it looks for a name server for If a nameserver for does not exist, it next looks for a name server for in its local cache. If a nameserver for does not exist, it looks for a nameserver for shop in its local cache. If a name server for shop does not exist, the DNS resolver asks one of the root namservers (as listed in its local cache) what to do.

      To get around the idea of there being only one nameserver, one has to implement a nameserver to, in addition to setting the root namservers when creating a new cache, setting the name servers for alternate non-ICANN TLDs and placing those names and IP addresses in the cache.

      This way, when one goes to, they ask the ICANN nameservers where to go. When one goes to, they ask the DNS servers that the DNS administrator has specified where to go. Anyone with a reasonable amount of clue can set up their won DNS cache and be in control of what TLDs they wish to resolve.

      This is very similiar to how USENET works. To create a newsgroup in one of the "big seven" (comp., soc., talk., rec., news., sci., and misc.) hierarchys, you have to get approval from David Lawrence. To create a newsgroup under the "alt" hierarchy, there is much less red tape involved. Less news servers (traditionally) carry the alt. newsgroups, but there is more freedom under the alt. hierarchy.

      There are also a number of other Usenet hierarchys which have even less propergation than the "big seven" and the "alt" Usenet hierarchys.

      DNS can be set up in a similiar fashion. The amount of code that needs to be changed is fairly trivial (there is a technical concern about what to do if the "root servers" for shop. give you a referral to different name servers that actually serve the shop domain instead of a referral to an appropriate subdomain, but that is easily enough handled compared to the amount of effort involved in making a caching nameserver). The only thing that has stopped this is mainly Paul Vixie's notion that non-ICANN TLDs are somehow evil.

      - Sam

  • Who gets to decide? No individual or organization with a financial stake in the decision itself.
  • by jimhill ( 7277 ) on Wednesday May 30, 2001 @12:42PM (#188566) Homepage
    Read the memos. Their entire point is that when it comes to domain names, (say it with me now, all you "Highlander" fans) there can be only one.

    The point is that at the end of the day, a domain name has to resolve to an IP address. If the same name resolves to two addresses depending on where you are, that's a Bad Thing. Unfortunately, that's exactly what is going to happen now that ICANN has decided to issue a .biz TLD. Existing DNS servers that don't look to ICANN for guidance already issued a .biz TLD and it is reasonably well-populated. There's every reason to believe that problems will ensue and that the lawyers will get to make a lot of money quarreling over the rights to and others.

    The suggestion in the memo is a good one: abstract the existing DNS away one more layer. All the roots have to play nicely together. ICANN doesn't get to introduce .biz because someone else already has. The thus-far successful first-come, first-served policy would dictate who gets a new .TLD just as it has dictated who gets

    It's a simple solution. The memos (which are far from being official IETF positions) are needed because name issuance has become a business. A very, very big business. Tens if not hundreds of millions of dollars can be made -- if the artificial scarcity of domain names is maintained. ICANN has/had a public trust: to administer the root domain fairly. They have perverted that into a profit-making venture. They have, not to put too fine a point on it, forfeited their privilege as the ultimate arbiters of domain naming. The Virtual Inclusive Root proposal of these memos sidesteps ICANN.

    Mr. Higgs has written a very clear pair of documents that deserve to be taken seriously, and to have their content codified into one or more RFCs requiring root solvers to treat other root solvers as peers. This is a problem that will have to be engineered away, presenting the US government with a _fait accompli_. The moneyed interests that have hijacked ICANN will never permit foreigners and weirdos to horn in on their cash train through legislative action and the US government will never permit other nations to have an equal say in such legislative action. The IETF way has always been "rough consensus and running code". Now more than ever, we need that.
  • It's no longer strictly true that a URL points to a "single point" somewhere. The widespread use of content delivery networks and content modification systems means that in practice a URL refers to "any copy of a particular resource, possibly as modified based on factors including your location, browser type and language preferences".

    Many URL purists will protest that this violates the spirit of URLs, but the flip side is that this allows much more efficient and flexible distribution of content. Al

  • Unfortunately, this is not a joke - it happens in real life.
    In the list of possibly infringing domains on the NISSAN trademark, there was found TenNISSANdiego and ReNISSANceCruiseLines.
    The last one was probably a misspelling cybersquatter, but it was not Nissan he was aiming at...
  • by hta ( 7593 ) on Wednesday May 30, 2001 @09:42AM (#188569) Homepage Journal
    Note: Anyone at all can publish a document as internet-draft. These documents are the work of Simon Higgs, who has been a proponent of "alternative roots" for many years.
    Their publication says ABSOLUTELY NOTHING about how the IETF community as a whole views these matters.
  • How can you trust the applicant to accurately represent their business. Is a non-profit still? Should /. be forced to discontinue the org and only use the com?
  • A "supreme court" is made up of many judges who vote on the outcome.
  • > why anyone would want to go to the trouble
    > just to view a specific web site when the
    > alternative (use an existing domain name)
    > would be easier.

    Easier for the viewer, but definitely not easier for the person putting up the site.

    There is a shortage of domain names, in part due to squatting. Any dictionary word is probably already taken, and the available two word combinations are getting a bit obscure. Things like "" have been taken since 1997 by squatters, but never used. To get a decent name, they hope people will buy the one they've already squatted on.

    It would be TRIVIAL to fix the shortage by adding new domain names. Then you could get potato.tx or whatever if you're in texas.

    But to make the squatting problem go away, you have to add a LOT of domain names. Adding them one at a time just opens up a new gold rush. Adding a hundred makes squatting pointless, and would make the existing pressure back off a bit. (The recession is helping there, though. But that's a temporary thing.)

    As long as there's tight control, there will be squatting. Right now, domains aren't really considered part of the name by most people. they just assume it will be dot-com, and if that doesn't work they try dot org or dot net. (And if that doesn't work, they give up.) Domains should be more like area codes.

    > Now as for this TLD dipute, ICANN is certainly
    > acting in a high-handed manner, but they do
    > have a minor point in that the people who set
    > up and signed up for these alternative TLDs
    > did so under the knowledge that they were
    > non-standard and there were really just
    > speculating that they might succeed in the
    > future.

    Clue: most of the growth of the internet since late 1994 falls under that category. And the "standard" is only standard because they say so.

    Kind of like saying all these non-standard browser extensions like "tables" netscape threw into their browser should be banned from all future standards because people who go and implement non-standard extensions need to be punished. Standards should only be extended by committees who know what's good for us.

    Nevermind users who are already trying to make the new thing work on their own initiative. Where would we be if standards documented existing reality rather than creating utopian ideals?

    > Joining some of them to the main root
    > now could be seen to be rewarding this and
    > thereby encouraging all kinds of other people
    > to set up non standard roots or TLDs
    > themselves.

    And this is a bad thing because...?

    Because they're trying to lead/control the system rather than co-ordinate the activity of the users?

    The internet is a free software phenomenon. A bazaar. Collaborative development. That kind of thing is co-ordinated, not led. If you lead, nobody's going to follow. IBM noticed that in the late 80's...

  • by landley ( 9786 ) on Wednesday May 30, 2001 @09:59AM (#188573) Homepage
    Whatever happend to the decentralized nature of the internet? Isn't there a reason we have more than one root server in the first place? Hello? What ARE these people smoking?

    As for a single point of CONTROL, if we're not going to let Microsoft be it, and we're not too happy about the federal government trying that out with encryption exports or various "commnications indecency acts"... Why do they think we're going to let THEM do it?

    The main problem with alternic type schemes is they need an alternate search engine. But if Google went along with it, there would be NO problem...

    The argument "grandma doesn't know how to set her nameserver" is kind of bogus if you stop and think about it: five years ago grandma didn't know what the internet WAS. The web was just geeks creating value for other geeks, and then the rest of the world found it and wandered in to our party. If we're over creating value in one corner and the rest of the world isn't doing as much with their 95%, then the rest of the world will find us. Remember Napster? Geeks are being QUIET about AudioGalaxy, this time around...

    Besides, remember usenet BEFORE AOL found it? A lot of people would consider the exclusivity (while it lasts) to be a good thing. Brains being the price of admission, and all. (Not trying to be bigoted, just saying it's not a BAD thing.)

  • By invoking "Goodwin's Law" you lost first.

    Goodwin's Law: A humorous joke taken at the expense of those who, in the early days of USENET, often got a little too passionate in expressing their viewpoints.

    The Modern Invocation of Goodwin's Law: [1]The mistaken notion that pointing to lessons of history, in particular those that occurred in Europe around 1935 through 1945, and relating them to current discussions is somehow taboo, that in so doing one has, by violating some obscure social norm certain cretins on the internet seek to define and enforce, automatically acceeded or reversed the very point they were trying to make. [2] The notion that any mention of Nazi Germany, the Holocaust, or activities surrounding these people and events, automatically disqualifies one from taking part in a discussion. [3] A tool being actively exploited by historical revisionists in an effort to use social engineering to diminish, and eventually marginalize, awareness of the holocaust and lessons learned therefrom. [4] The final argument of one who cannot win a point of contention through reasoned discourse, or had no point to make in the first place, but instead clings to an old, stale joke as though it were a gem of internet wisdom.
  • Why not expand this to:


    URLs, or extended URLs, remain universal, but just like hostnames resolve to local search paths defined by /etc/resolv.conf, so to could shortened URLs resolve according to the local system or user's configuration.

    Let a systemwide configuration option define what is the default if the root domain identifier is left off, and allow the systemwide default to be overridden by each user for their own processes. Then, all of my ftp, ssh, web, mail, and other traffic would default to opennic when I send it to, whereas the systemwide default might well go to icann, or elsewhere.

    Astute system administrators would of course set the default to opennic, but individuals can and should be able to override it.

    This solution has the added advantage that alternatives such as opennic would in no way have to adhere to ICANNs standards, and indeed, if ICANN decides to play dirty and steal anternate tlds, opennic could respond by issuing its own .com's in direct conflict with ICANNs.

    We could choose to whome we wish to listen, and to whome we wish to defer, and may the most just and fair naming service take the bulk of the marketplace, but may no one own it outright.

  • I think AT&T should start putting in their DNS servers a new TLD called .att and then telling these self-ritious, domain name hogging, greed bound, morons that they can go and bite themselves. If a really bog telco or backbone company thumbs their nose at the idiots that sit on that board the better.

    they dont OWN the internet, why the hell do we tolerate the constant abuse and "claim to ownership" these idiots have?

    Break ranks, insert your own TLD's in your routing tables!
  • I keep thinking that the solution needs to be NO TLDs, and some scheme for charging exponentially more for each domain name registered to a given legal entity. $35 for the first, $1225 for the second, $42,875 for the third and so on.

    This kind of sliding scale would effectively kill off speculation and hoarding by making it so expensive to do "honestly" and so time consuming to do dishonestly (shell corporations to hide ownership).

    The problem with inventing new TLDs is that the rights holders all buy/sue/bully their way into owning mytrademark.somenewdomain. You only mariginally increase the number of new domains by doing this, since the "lucrative" domains are already taken by people with the money to beat their way into ownership.
  • Wow, I never thought of that. I can have an email address that can only be spammed by spammers that use OpenNic? You're making this awefully tempting...
  • You have explained that the hospitals don't have a common TLD, but have not explained how that's a bad thing. What is the downside of some hospitals having .edu, some having .org, and some having .com
  • by sethg ( 15187 ) on Wednesday May 30, 2001 @11:07AM (#188580) Homepage
    Under the current (ICANN) set of TLDs, anyone who sets up a domain name must classify it as a ".com", ".net", ".org", ... or as one of the country-specific TLDs. Most people recognize that this bureaucratic classification scheme doesn't correspond with our mental classification scheme; for example, as someone else in this thread points out, should a hospital be a ".com" or an ".org"?

    Unfortunately, the commonly-proposed solution -- adding more gTLDs -- is not going to help. If health-care organizations get their ".med", then sooner or later, someone is bound to want separate gTLDs for doctors, dentists, and homeopaths. If a ".mp3" gTLD becomes widely used and another music format supplants MP3, then people distributing music in the new format will still set up ".mp3" sites for that purpose. And so on and so on, until users are confused by too many gTLDs, and companies afraid of cybersquatting register their names with 20 gTLDs, not just two or three.

    When people learn vocabulary, they learn the words for genuses first, and learn other levels of classification later. That's why a child, seeing a wolf, says "that's a dog", and not "that's a member of the species Canis lupus in the order Carnivora." That's why so many people set up personal domains under the ".com" TLD, even if they have no intention of making these domains commercial ventures -- they recognize ".com" as the default TLD and don't care about its alleged purpose.

    Back in the eighteenth century, a number of philosophers tried to construct languages to mirror (their views of) the natural order of things -- their dream was a language where a false statement would be ungrammatical and where related concepts would have similar-sounding words. The people who want to "improve" DNS by adding more gTLDs are falling into the same trap.

    We need fewer gTLDs, not more.

  • /.
    I don't know why someone modded you to "funny" - I think the idea is eminently practical.
    We could check with Paul Vixie to be sure (keeping in mind he is, um, "a man of strong opinions" :)) but I don't think it would be too tough to scale BIND up to 17576 TLDs. Theoretically, the number of actual domain names would remain nearly constant anyway, since the companies with multiple names could drop down to one truly descriptive one.
    As for the .tld domain, that's the perfect place to keep the index of what the other TLDs mean! In trademark law [] names are distinct within a sphere of influence or "class" - since Acme Machine Tools and Acme Supermarkets exist in different classes, there is no valid trademark infringement. The address would be a page defining the business ventures allowed to own unique names in the .sex TLD, for example.
    I think you are on to a good idea, and one that might eventually come to pass simply because of the strong evolutionary forces affecting the DNS. ICANN can't keep their stranglehold forever.
  • /.
    The .md domain serves admirably as an expensive "vanity tag" for physicians, and I hope Moldava continues to prosper from it.
    It has not (perhaps fortunately? Is a goverment less politically driven than a corporation? Is there any difference between the two anymore?) evolved into a TLD for medical establishments at this point.
  • /.
    Sorry, though it was obvious. Sometimes we lose prespective over time.
    Before NSI let the /org/com/net TLD fall into chaos, by letting anyone take any or all of the above regardless of function, it was easier to find stuff. You didn't have to use slow-loading, information-broadcasting search engines if you were familiar with the basic system.
    If you don't remember those days (and I'm not trying to pretend it worked perfectly, BTW, just that it was better organized than purposely disorganized) think of this - when you see a DNS name that has a .mil or .gov on it, you KNOW what that means. It's not a random or purposely deceptive tag, it carries meaning to the user.
    A .med TLD would indicate the owner provided some function associated with medicine. It's more important to the function of these institutions than their profit/non-profit status, after all... the .com and .org carry zero meaning other than differentation.
    If differentiation is all that is required, we don't need names, we can just use IP addresses. Names (rather than addresses) are designed to be humanly memorable, and humans are better at remembering things that carry meaning - for example, I can't remember in my head which of the 30 hospitals I deal with on a daily basis are .edus, and which are .coms or .nets. The TLDs for these are essentially random and not memorable. The individual names, on the other claw, are easy - for example, NMMC is North Memorial Medical Center, but is it a .com or .org? I honestly don't remember right now.
    And before anyone mentions it, bookmarks are a sub-optimal solution, because I don't use the same machines every day, or even the same operating system, and the existence of a kludge does not make a proper solution undesirable.
  • /.
    I know that's the official rationale, but I don't buy it. I mean, how is it a valid experiment to offer lengthy, unpopular, undesireable, cumbersome, language-specific names?
    You can't validly extend the response to a .museum TLD to predict what the response to a .sex or .xxx TLD would be. Is everyone at ICANN an idiot, or are they assuming we are all idiots?
    And there are no real technical issues here that ICANN has any control over. The people with the ability to influence the technology work for Vixie Labs.
  • by Medievalist ( 16032 ) on Wednesday May 30, 2001 @10:20AM (#188585)
    Disclaimer: I'm one of the people that've been petitioning for a .med TLD for years... so those readers who keep up on DNS issues have already heard this rant.

    Currently, the hospitals of the world are randomly scattered across the DNS. For example, Fox Chase Cancer Center is, and Holy Spirit Hospital is HolySpiritHospital.Com, and the American Hospital Directory is AHD.Org. All these are non-profits except possibly the last.

    The need for a .med domain is so strong that one of the alternative registries has already created the TLD. Unfortunately, ICANN's stranglehold on the industry prevents this alternative registry from seeing wide use (although adoption by hospitals through inclusion of alternative roots in the bind cache is ramping up).

    ICANN has refused to discuss the issue except to say that people who supply a solution to the problem (i.e. alternative registries) are the bad guys, destablizing the Internet (ha! I've been using all the major DNS roots simultaneously for years; just add the additional root entries in BIND).

    I expect that when the .med finally comes about, the people currently trying to serve the public good by providing the TLD will be marginalized if not driven out of business.

  • URLs are defined for all standardized Internet protocols.
    pop3:// is a valid URL, as something like telnet://
    Yes, those as 'resource locators' make little sense, but it's legit.
  • If that means that when I type in "" into a browser, I go to and when you type it in, you go to so what? It's my choice which listing service I use.

    Do you have an email address? Do you give it out to other people and expect that they can send mail to that address and it will get to you? Thank centralized namespace authority.


  • This has got to be ICANN fud. I haven't read the RFCs, but there's no reason you can't make the "root" system a p2p distributed database. This is not a power grid where economies of scale are proving to be a more efficient way of working than deregulated competition;

    It's not the infrastructure that has to be centralized, it's the authority. You can make a p2p distributed database, but who populates the database? How do i know that mail sent to my address will get to me when the address points to different places, depending on whom you ask?

    it's more like a telephone grid.

    Exactly. There is a centralized authority that doles out phone numbers.


  • by Mike Schiraldi ( 18296 ) on Wednesday May 30, 2001 @12:09PM (#188589) Homepage Journal
    Oh, please.

    Although the Internet allows a high degree of decentralized activities, coordination of the assignment function by a single authority is necessary where unique parameter values are technically required.

    The phrase "single authority" is never good.

    If not a single authority, then what? What happens when one authority says is and another says it's What happens when i register through Registry A but someone else beats me to it at Registry B? I can't put that domain in my ads, and i certainly can't use it as my email address.

    Imagine if you saw an ad for 1-800-FLOWERS, but when you went to call it, you got Joe's Crab Shack because you and the placer of the ad used different "telephone authorities". Imagine if you met some hot chick at the local bar and gave her your number, but when she went to call it, she got someone else.

    You need an "supreme court" of the namespace or else the namespace is useless.


  • by Mike Schiraldi ( 18296 ) on Wednesday May 30, 2001 @02:25PM (#188590) Homepage Journal
    Yeah, sorta like ICANN.


  • Check out theses nice ICANN cartoons:

    You might mention that these are not *just* humor cartoons; they are flash (and Quicktime and a half dozen other formats) cartoons that attempt to educate with humor.

    Certainly they are very slanted, but they give a pretty good overview of the situation (particularly the second one).


  • pop3:// is a valid URL, as something like telnet:// Yes, those as 'resource locators' make little sense, but it's legit.

    Heh... so is tv:// (RFC 2838), and audiocd:// (in KDE, this gives you a virtual directory tree full of wav, MP3 and OGG files that can be copied anywhere). My original comment regarding the proliferation of localized URLs is here []. I can't find the later essay that details several concepts like saving a file to printer://page.png to print a page, or reading cam://image.png or cam://video.mpeg to grab an image or stream from a webcam (and be able to do things like link it to /home/httpd/html/current.png to put it on the web with no code whatsoever). I've been a big proponent for awhile of using URLs and mimetypes at both the user and system level. It's a nice example of a simple, easy to use model for newbies that is also sufficient for most power users if implemented correctly.


  • For instance, if was resolved twice,

    Completely disregarding the technical side of that concept, and even disregarding how that would work for email, napster, automatic indexing spiders, etc, you wind up with a big problem:

    It's no longer a Universal Resource Locator. One of the great things about a URL is that it refers to a single, discrete point somewhere. It can be on your harddrive (file:), on a LAN (MyServer//), or on the internet (ftp:, http:).

    And as for the fellow up above who said "Just add the nameserver to the url", often DNS does not *have* a URL associated with it. For instance, setting up your POP3 or Napster server, you just enter a Domain name.

    Also, adding an extra bit of text to the domain name to get it to resolve correctly has already been written into the fundimental archetecture of DNS. It's called a TLD... that's what these things are FOR. Namespace collision is the problem, but a combination of politics and no clear orginzational responsibility is the cause.


  • Allow http://ns1/ to go to teh first name server, and http://ns2/ to resolve
    Except that this format changes existing documented (and standard) behaviour for <A HREF> tags.

    This doesn't belong in URL syntax. It should be done at the DNS level -- say, /etc/resolv.conf, or .../dnscache/env/... .


  • Forget e-mail, why? Email address are, and can be anything. While some people might fight for, fact is that 90% of the fight is for

    Namespace collision *is* the problem. Even if we add an infinit number of TLD people will still want

  • Actually I'm not. Why? Because the conflict exists in webspace, how many companies fight it out for a specific email address? I can only think of 1. McDonalds.

    Mostly they fight it out over the address so they can have And I do mean the popup window should appear for ANY request to that domain. (even to send e-mail).

    The point is that if the functionality is added conflicts are not a major issue, and instead you can have a proliferation of TLDs and less concern over conflicts. (Which, preferably, would never happen).

  • by topham ( 32406 ) on Wednesday May 30, 2001 @09:37AM (#188597) Homepage
    I'm sitting here thinking about the problem, and I've decided that for my Internet use it would be acceptable for a Pop-up window to appear asking me which Root server should be primary for a given site (only to appear when a conflict exists between the root servers).

    For instance, if was resolved twice, once by ICANN, and once by otherDNS (ficticious) then I could simply choose from a popup which site I want, and depending on it's importants it could cache the results for the session, or flag it permanently.

  • Thanks for the comment. Interesting stuff. If I had moderator points, I'd mark you up.
  • I think the issue of whether ICANN control the whole root DNS, or whether there are alternatives is really missing the point.

    DNS is designed to make human friendly names out of computer friendly IP addresses. This is all it should do. The name should be informative, not a tool for advertising.

    So, and (to pick two random examples) are not particularly informative, except to promote the concept of a "brand". IMHO, the current DNS system needs to be deprecated and a new hierarchical system developed in it's place. Slashdot would then be something like comp.culture.slashdot and Freshmeat would be something like Sound like Usenet? Good, it's supposed to.

    Such a system would get around the problems with domain name ownership, domain squatting etc, and only present one problem: How to search for sites. I think Netscape had a technology out a few years ago designed to "replace" URLs as the first point of contact for users (was it called RealNames?). IIRC, this was proprietery so wouldn't be an Internet success, but some form of heavily distributed metadatabase, separated from the DNS system could provide a wrapper for the new DNS name, while the experienced user could navigate the hierarchy to get to their site. As an analogy, consider the difference between a search engine like Google, and a directory like DMOZ.

    It's not a particularly well thought out idea, but it could be implemented "in place" (ie. sites could map their names to the new hierarchy while continuing to use their current DNS).
  • I think the idea of wresting control from ICANN is great... Permitting new TLDs would be a great idea, and the new TLDs could be run by anyone, as long as they can prove they meet certain criteria. In this way, no collisions are necessary. It would, however, create a legal nightmare as players scrambled to open new TLDs and new rounds of squatting began. But ICANN refuses to be open about their process, so we'll never *really* know what standards are being met.

    But collision of domains would be disastrous. It amounts to taking control of the internet from ICANN and giving it to AOL. AOL could easily decide to point somewhere else. Stink.

    Just because there are shades of gray, it doesn't mean we can't tell black from white

    But it can keep you from getting the web sites you want. If there were a second that pointed to microsoft's site, how would you get the slashdot you wanted? DNS would be worthless.

    Boss of nothin. Big deal.
    Son, go get daddy's hard plastic eyes.

  • Score -5, Incredibly Stupid.

  • Up for bids is a personal DNS server -- a one of a kind application, even though every Linux box that exists is capable of performing this function.

    The product that for which you are bidding is the DNS server for such inventive domains as,, and even the speculative

    Oh well, trust the media to explode this non-issue into Gatesian proportions.

  • ICANN doesn't get to introduce .biz because someone else already has.

    And there's the root of the problem. Who gets to decide what it means to say that "someone else already has"? For any possible TLD, there could be someone (or several someones) who's already set up a machine under his bed to handle DNS. Somewhere there has to be authority that decides who counts, if DNS is to work at all.

  • No. Why subject those half of the IETF to abuse from the usual suspects?
  • Huh? What makes hospitals so important than they need their own TLD, like .museum [] and other new too-narrow loser TLDs []? Why not .car for automobile dealers? Why not .pub for the local pub? Does ICANN also hate car dealerships and pubs, too? I don't know about you, but there are lot more car dealerships and pubs in my town than hospitals, so of course it makes sense for them to have their own TLDs.

    Devolving into rant...

    Why don't people bitch about the dewey decimal system, insisting that their favorite topic has a top-level number? I mean, hell, it's obvious discrimination, since we no scarcity: we have an unlimited supply of numbers, right?

    Adding TLDs won't do squat to solve the problem: that idiot registrars don't enforce some semblance of rules in an organizational system. Imagine if, as a publisher, you could spend a couple of extra bucks, so that your book appears in five or six different dewey-decimal numbers in the library, just to increase your "hits". Or they register lots close-enough numbers in case you made a typo in your number query.

    Of course this is preposterous, as it obviously destroys the point of having an organizational system. But that's what happens today when company X decides it has to have X.* where * is all available TLDs that can be had.

  • I agree completely.

    However, the poster wasn't asking about that. He stated that .med is already in an alternative root, but that's not good enough. He wants ICANN to put it in the "official" root. That's different.

    Alternative roots are fine 'n dandy with me, but sadly, impotent. Email and URLs are only useful if we have one naming scheme. I'm not pointing my setup to Bubba's Root Server and Pork Rind Shop anytime soon. I'd rather communicate with people, get work done, and get information when I need it, than worry about people dicking around with pedantic naming issues. I don't care about the domain name of hospital down the street as long as I can communicate with it when I need to.

    Sure, you're free to set up your own root, or use an alternative one, just as I'm free to ignore it.

  • Mr connerbd,

    You are hereby notified to cease and desist from any further such posts.

    As you must be aware, any more discussion of distributed root servers would violate your Windows XP NDA.

    Without going into particulars, your public disclosure could jeopardize critical intellectual property that would subject you to immediate and severe litigation that would cause your molars to disintegrate.

    I am referring, of course, to the new Windows XP name server cache, which is meant to enhance the end user experience with increases in efficiency by caching frequently desired URLs, including advanced aliasing, such as ->
    that provides a richer experience and the innovation that our customers have come to expect. Of course the same product includes our advanced Pr0nKiller/anti-terrorist MShopping Cart that will be pre-announced by our Chief Software Architect.

    Good day.

    Geoffrey P. Foggbottom, JD

  • by zpengo ( 99887 ) on Wednesday May 30, 2001 @09:50AM (#188615) Homepage
    Here are some more informative RFCs regarding TLDs and related servers: RFC2795 [], RFC3092 [], RFC2551 [], RFC2100 [].
  • by zpengo ( 99887 ) on Wednesday May 30, 2001 @09:40AM (#188616) Homepage
    How dare anyone challenge the monopoly? Don't they realize that by bothering ICANN, they're only suppressing innovation?

    From the ICANN "one root" doc:

    This document reaffirms ICANN's commitment to a single, authoritative public root for the Internet Domain Name System (DNS) and to the management of that unique root in the public interest according to policies developed through community processes.

    I bet they'd change their story if it were decided that the "single, authoritative public root" out to be someone other than them.

    Although the Internet allows a high degree of decentralized activities, coordination of the assignment function by a single authority is necessary where unique parameter values are technically required.

    The phrase "single authority" is never good.

    Over the past several years, some private organizations have established DNS roots as alternatives to the authoritative root. Frequently, these "alternative" roots have been established to support for-profit top-level domain registries that have been unable to gain entry into the authoritative root as managed in the public interest by the IANA or ICANN.

    'Don't listen to the "other" guys. We're looking out for you.' Yeah, right.

    Because these alternative roots substitute insular motives for the community-based processes that govern the management of the authoritative root, their decisions to include particular top-level domains have not been subjected to the same tests of community support and conformance with the public trust.

    Sound anything like Microsoft's "Open Source is unsafe" theory?

  • by Christianfreak ( 100697 ) on Wednesday May 30, 2001 @12:31PM (#188617) Homepage Journal

    IBM buys all rights to all acronyms that contain the letter 'I'. The company then proceeded to sue the ICANN, the IETF, SGI and the RIAA unless they removed the 'I' from their acronym within 24 hours.

    ICANN promptly issued a statement that the would be switching to AYTBTU (All your TLD belong to us) and the RIAA says that they will fight the move in court. Their spokesperson stated

    "They can't do that because we along with the MPAA control everything in the would that we can trick the government into believing is intelectual property".

    SGI also joined in the response saying: "We got 'Open' and 'GL' but no one here ever thought of just the letter 'I'.

    Market analysts are predicting that other companies would soon follow this lead and begin copyrighting various letters of the alphabet.

    Stated one:
    This is a dangerous precident. Imagine if someone copyright's the letter 'M'. That person would take out IBM and Microsoft in one sweep!"

    Several bystanders who heard this quote rushed to the courts in order to claim the letter 'M'

    More information as it becomes available

    "One World, one Web, one Program" - Microsoft promotional ad

  • What about adding something to the end of the URL, or the beginning, that tells the browser what NS to resolve with?
    ex. is in 2 servers. Allow http://ns1/ to go to teh first name server, and http://ns2/ to resolve from the second?
    =\=\=\=\=\=\=\=\=\=\=\=\=\=\=\=\=\=\=\=\= \=\=\=\=\
  • I'd once suggested that migration to this could be done as follows:
    • A new TLD, "www" is created.
    • All .com names are copied over to ".www".
    • All .net names not conflicting are then copied over.
    • All .org names not conflicting are then copied over.
    • New releases of browsers convert "foo" to "foo.www" by default.
    • No new registrations are accepted in ".com". After 5 years, .com is phased out. ".org" and ".net" remain.

    This made more sense when things like "internet keywords" were being taken seriously. (Whatever happened to "RealNames", anyway?) It's probably not worth the trouble now.

  • by mybecq ( 131456 ) on Wednesday May 30, 2001 @10:14AM (#188624)
    Why don't they just make a TLD for every possible three-letter combination? That way, people/companies will put their website in the correct category, instead of registering their name under every possible com/org/net/biz/web TLD. It would become too cost-prohibitive otherwise (considering current pricing).

    This will solve the TLD problem once and for all -- plenty of supply for TLDs and new domain names for everyone...

    (Of course then ultimate TLD will be .tld itself.)
  • Intel trademarked the letter 'i' a decade or so ago, didn't you hear?

    Rate me [] on
  • Insightful? +1 freakin' Hilarious is more like it, even if he couldn't spell my freakin' email address...

  • Even then all you have to do is divide up the space appropriately. A p2p registration system would simply make sure that the name being registered doesn't exist in any other primary nameserver. The easiest way is to have separate tlds, but that's not all that practical.

  • This has got to be ICANN fud. I haven't read the RFCs, but there's no reason you can't make the "root" system a p2p distributed database. This is not a power grid where economies of scale are proving to be a more efficient way of working than deregulated competition; it's more like a telephone grid.

    I guess there's not too much else to say, except for acknowledging that politics will make a mess of things...

  • I bet you were just trying to see if you could get modded up with the mods reading them, right?

    Moderators, check the links before you moderate. Yes, they're RFCs, but if you look at the titles and the dates, you'll see he's not exactly "informative".
  • I don't know about you, but I don't want to explain to my mom why the "" link I sent her doesn't work, and what she has to do to fix it, and how that fix will wreck her links to her other sites because she has changed her nickname authority. It sounds like a worse mess than what we have now, where not even the most basic rule holds anymore: this URL takes you to this specific page. (how is that an improvement??)

    There is a lot of value in non-colliding nicknames. Geeks may not mind a few extra steps, but... Hey, wait a minute. I would mind too, actually.

    this is a public trust, then I'd like to see a public audit of the books.

    That I can definitely agree with.
  • The idea that we need a central authority to dictate nicknames is ludicrous. The idea that if nicknames collide, the internet is "destabolized" is equally silly.

    A large problem with this (which is mentioned in the linked articles, but who has time to read those?) is that a lot of existing software assumes that the hostname-to-IP mapping is global to the internet; email handlers, search engines, and transparent web proxies are three examples. There's also the madness that results when non-authoratative DNS records manage to cross over into machines that aren't "loyal" to the agency that defined them.

  • Ok, but ask yourself this; is motor parts joeblow registering just, and NOT You can be if he has the chance he's going to register all tld's that he can.

    When a company currently registers a domain it jumps on as many tld's it can get, making the point of tld's completely useless. I'm just suggesting that typing the extra tld in my url to actually be useful again, instead of just being useful to the registers to make money off of.
  • What good it would be is to add a hopefully "descriptive" tag to the URL. If it's got a .net then it's most likeley a ISP, since a commercial company is most likely going to use .com

    I'm not going to go much back over in what I've replied to the statement about flattening of the namespace... but essentially nobody is doing what you're saying, when a company registers a domain they are registering every possible combination of tld's they can get... effectively flattening the namespace, making the tld non-descriptive, and just giving money to the registers for no good reason.

    Once people actually *have* to make a choice and decide where they should be, would adding additional truely descriptive tld's be effective. Since there would actualy be purpose to the tld.

  • Look at the way it is today, if everybody was forced to use only *.com whoever had a .net, .org, etc. just lost their domain; would the internet only have a third of it's servers??? No, most likely 90% of the servers would still be there, because companies are not just registering one tld they are registering all they can get their hands on.

    Using your example, look at slashdot, there's a .com & .org, going to the same place, slashdot is stepping on two-thirds of the space already. Right now tld's are completely useless: they add confusion to the consumer (or websurfer), make me type extra characters for no good reason, currently they're only really good for lining registers pockets with money by having the same company register the one domain 50 billion different ways.

    If we did this there would actually be a point to adding additional tld's since you'd sit in whichever one made the most sense; instead of getting every one you can get your hands on, no matter if it made any sense at all.

    And just be a bastard, why don't you just include that tld information into your domain??? there you have your two-thirds of domains back :)
  • My opinion on the whole TLD mess is that there should only be one domain.[net|org|com|.*] Instead of having & that whoever registers it has to pick a tld and the rest of the tld's with that domain are unavailable.

    What this would get me is the confidence that I can say a .net is most likely a network provider, a .org is an organization, etc. Since you'd pick which tld is going to be most appropriate for you business.

    Also we don't have lawsuits between &, etc. since they couldn't exist because they'd have a domain clash.

    While I'm on a roll (rant), I'd even like it better if you actually had to show you truley belonged in a certain tld (you have to provide network services to be in the .net tld).

    I have no idea how to reverse the mess with all the different organization that are in the mess of having conflicts, but I know it could be implented for any new domains.

    Of course that's me being an old curmudgeon
  • should a hospital be a ".com" or an ".org"?

    .org, why would it be .com?

    they recognize ".com" as the default TLD and don't care about its alleged purpose.

    Yes, and now we're having trouble with squeezing everyone into .com. You have correctly identified the problem; do you have a suggestion for a solution?

    The people who want to "improve" DNS by adding more gTLDs are falling into the same trap.

    No, they're trying to solve the problem you identified above, the polution of .com has rendered it meaningless and absorbed a large number of domain names which other people would like to use. Since the "default" namespace is packed-out they need a new namespace. I assume that your answer would to be to progress on to polluting .org, .net, .mil, .int etc too. "Unfortunately" ICANN wouldn't go for that either as it would offend their friends who are passing them big bucks in bribes to prevent the namespace being extended in any way which, for example, would allow Mr Arnold O'Leary to buy or or

    A better system would be to allow AOL.person (or dot something) and simplly not allow bastards like ICANN to take it off Mr O'Leary when they get that call from a well known network with limited Internet facilities.


  • That's all fine and all but actually finding a replacement solution is a non-trivial exercise.

    We all know what the problem is with central authority and the DNS. What we don't know is what to do about it.


  • Why can't they handle that with a couple phone calls and maybe a 20 minute teleconference??

    They can but they don't want the MONEY from .biz to go to someone else. They know as well as anyone that .biz will be a goldmine once it's "blessed" as a TLD because every company in the world with a .com will have to register the equivilent .biz domain to avoid cyber-squatters.

    Every company except ICANN's special friends, that is. They can just wait for someone to buy a .biz domain they want and then get ICANN to give it over to them, which ICANN couldn't do if it did not control .biz.


  • Except why should a user have to know whether a hospital is for-profit or non-profit when trying to type in a URL to visit their site?

    Why should I have to know if St Mary's Hospital is on Main Street or Castle Street? Because they are two different addresses. Do you want to make it illegal for the second Hospital to set up just because we've already got one with the same name somewhere else?

    On the web, if you go to and don't find what you want it's not hard to try instead.

    It's moot anyway since most people use a search engine to find things like that so it could be a random number for all they care.


  • 1) ICANN, AlterNIC, and other alternate root servers make up the "Root Server Pool". New entities may join in as long as they meet some QA requirements.

    2) Each of these entities, in round robin fashion is asked every N months (12?) for a NEW TLD which isn't part of the system yet (less than N letters -- 8, maybe?)

    3) This turn's entity gets control over the chosen TLD and administers it according to its own policies.

    4) N months later, the next entity in the pool is asked the same question, and it has to come up with a new name or loose the turn.

    Makes sense?

  • What does the DNS do anyway? First, it provides names that are easier to remember than IP addresses. Second, it puts a level of abstraction above IP addresses. That way, IP addresses can change, be aggregated under a single name or several functions can be aggregated on a single interface and split up later.

    One thing it was not supposed to provide is the function of a directory. It is often suggested that domain names should be handed out according to a content-based hierarchy (where freshmeat would be known as or according to a location based hierarchy ( It would not work. The world as we see it is not a hierarchy where everything has one or only a few correct classifications. Sites like Slashdot would fit in all sorts of subtrees ranging from humor to bizzare and who gets to decide which classifications are correct?

    What is happening now is that the second function of DNS, the abstraction layer above IP addresses, is corrupted by the first function of DNS, the easy to remember names. I think that a stable abstraction layer is a really essential function. In my opinion DNS should concentrate on that and sacrifice the other function should that be necessary (and it looks like it is). This way competing or even conflicting namespaces would not be that big a problem. You could always fall back to the then somewhat uncomfortable but guaranteed-to-work DNS abstraction layer. This fallback does not exist today for several reasons, virtual webservers being only one of them.

    Conflicting name spaces will not be avoided. Money is to be made. The draft documents want to make us believe that this can be solved on a technical level: "Your internet domain name system root zone is violating internet standard RFCblabla because someone is already offering that TLD. Go away." Like that's going to work. It's important that these kinds of clashes will not harm the general interoperability of the net. Because of that, I suggest to stop handing out names on request and start handing out random strings instead (preferably only consonants or something like that to make sure it is really incomprehensible). Just make sure there is one set of really non-ambiguous addresses on top of IP addresses, available to anyone in large quantities, meaningless to humans.

  • These problems are entirely due to ICANN Mismanagement - their delay in bringing out new TLDs.

    Millions of gTLDs are being sat on by ICANN. They are the worlds greatest 'cybersquatters' (to use their definition of word).

    What sort of world would it be if we only had two hundred or so towns - each allowed only one High Street (Americans may call it Main Street)?

    The authorities have been giving you lies, spin and propaganda. They always known the solution to trademark problems -

    The .reg is instead of trademark symbol. This then acts as certificate of authentication and directory.

    This does not stop business using other domains for advertising e.g. current dot com.

    I have been communicating with US Government authorities - they have been unable to deny my assertions.

    Trademarks 'raison d'être'. To use Attorneys words, "The basic tenet of trademark law is to protect consumers and trademark owners from confusion in the marketplace"

    They know by using the DNS as a trademark system abridges free speech - violating First Amendment.

    But US Government don't give a doggies droppings about the law or your rights.

    Please visit my site to see full explaination - []
  • by sulli ( 195030 )
    That would be the most user-hostile thing I could think of. Pop-up windows for X10 cameras are bad enough - I can't bear to imagine having to think about the root server when just using the goddamn web!
  • The internet is a free software phenomenon. A bazaar. Collaborative development. That kind of thing is co-ordinated, not led. If you lead, nobody's going to follow. IBM noticed that in the late 80's...

    The internet is a lot of things... it's also a site of business opportunity, a growing economic center, a gathering point for both anti-capitalist and libertarian groups, and, of course, the focal point of Jon Katz's revolution.

    This kind of thing is not coordinated. That's the problem. You try to coordinate anything here, and it goes nowhere. Too many divided interests... and they're far from equal in voice.

    What's needed is leadership. Not what ICANN is doing ('Hey! We're here to lead, so follow us!), but actual leadership. That entails someone who has a vision of what the internet is, and what it should be... and someone who can convince the various groups that this is their vision too. Call it politics if you like (I wish... politics would be so much better with some decent leaders), but it's leadership.

    IBM may have created it, but that by itself doesn't grant them the mantle of leadership. ICANN may have held that ability for awhile, but they seem to have forgotten their mandate. Now that they're focused on maintaining their position, they're no longer leading.

    What's needed is someone, some organization, some non-profit, that can say "We've got a solution to these problems, and here it is... it's fair to the individual, it's fair to business, it's open-minded, and here's how we can implement it. Oh, and it's easy to learn."

    Personally, I don't see why we're sticking with this arbitrary system of names and numbers. Why not map the whole thing to a virtual space. We've all read SnowCrash... same deal. If the VR is too much to implement, do it 2D and without the avatars; the key idea is the physical allocation of 'geographical' space. Map it to real estate and then the rules map as well.... clears up lots of problems.

  • Damn where are my mod points when I need them! This is the most lucid comment on gTLD's that I've read yet!

    How about writing up an IETF draft for a competing standard that real humans will use. No ".com" ".org" ".net", etc.... just nicknames.

    The only problem is if companies with patents on this crap get in the way, like RealNames?
  • Although your philosophy may seem sound the practability is insane. Domain name registration simply can't be done in a totally anarchistic manner because there has to be somebody to avoid conflicts e.g. when two people claim to have registered the same name first. As soon as you have a body who can intervene in the registration you have a controlling body. It may have a different abreviated name, but it will essentially be the same thing. In short we need a naming authority. Just not like the present one.

  • by ackthpt ( 218170 ) on Wednesday May 30, 2001 @10:05AM (#188655) Homepage Journal
    How dare anyone challenge the monopoly? Don't they realize that by bothering ICANN, they're only suppressing innovation?

    Lie of the 70's = The check is in the mail
    Lie of the 80's = Trickle down economics
    Lie of the 90's = I have not had sex with that woman/man/computer/etc.
    Lie of the 00's = Monopoly promotes innovation

    All your .sig are belong to us!

  • Well, essentially your first idea is another iteration of trying to construct a Dewey decimal system for content; lots of people have been working on taxonomy stuff for a while, but I don't think you'll see much success - look how fragmented Yahoo, for example. Any working taxonomy is at the very least going to have to allow a concept like "aliases" or keywords in the tree, for things which map to multiple locations -- just like a pizza house appearing under "take away food" and "restaurants" (yummy).

    As you point out, RealNames was a sadly proprietary first attempt at a directory-cum-search engine, but if you want to build a non-commercial version of anything like this, you need to look at SOAP, UDDI and WSDL. The corporate interests are forging ahead on this one.

    For example, check out this web services toolkit [] on alphaWorks.

    There's been talk about a UDDI registry run "at first" by IBM and its partners... sadly I don't have a link better than this rather lame news item []. Anyone care to oblige? Perhaps thats an 'upcoming' interesting registry that we ought to be keeping an eye on...?

  • Okay, but you still have to do some resolution somewhere...


    So how does my machine know what name servers to use when I type opennic:? Or icann:? What if opennic changes their DNS servers? What if I want to start my own alternative DNS? What if I name it alternic too?

    You're adding another level of indirection, but you still have to have a master authoritative list somewhere or you wouldn't know which DNS servers to use. You can't make DNS fully dynamic -- that would defeat the purpose of having it in the first place.

  • by AdamInParadise ( 257888 ) on Wednesday May 30, 2001 @09:34AM (#188669) Homepage
    Check out theses nice ICANN cartoons: []
  • NSI gets anywhere from $6-$35 per year for each one. That's $240,000,000 to $1,400,000,000 dollars anually, and they don't even maintain the servers.

    Untrue, VeriSign (owner of NSI) DOES run the servers for .com, .org and .net. They also run half of the roots and in particular the A-Root which is the master.

    Running the DNS toplevel servers is a non-trivial task. It is not simply a question of downloading the latest version of Bind and sticking it onto an Intel box running Linux.

    The central problem with the domain name system is that people are using it as a yellow pages directory when what it is designed to do is much closer to the service that maps 1-800 numbers to the actual telephone lines.

    The Internet is NOT designed to have two different owners of a DNS zone.

    If what people really want is a yellow pages lookup system then what they should look at is CNRP - common name resolution protocol.

    Incidentaly, the drafts are personal contributions by a guy called Simon Higgs who I have not heard of before. He is not speaking for the IETF and the article should not make it appear that the IETF has endorsed the draft, it has not and furthermore given that it does not appear to be addressed to any of the IETF process tracks the IETF cannot endorse it.

  • We could check with Paul Vixie to be sure (keeping in mind he is, um, "a man of strong opinions" :)) but I don't think it would be too tough to scale BIND up to 17576 TLDs

    Paul has done some great stuff but the capabilities of BIND are not relevant to the debate. BIND does the hard job of supporting DNS servers that applications can point their query at and have the DNS server work out how to resolve it. BIND can also be used to set up a DNS resolver. However the tlds and the root servers are pure resolvers. They don't need to cache, they don't need to pass on queries. They are simply in memory databases that respond to what we scientists call a humungous number of queries a second.

    As a matter of fact there is no technical obstacle to having an unrestricted top level directory. In fact dotcom is pretty much that already - it has the vast bulk of the names and most of the lookups are in dotcom. VeriSign would be more than happy to run such a system for a modest fee per name.

    Politically the obstacles are enormous. ICANN will stay in control of the root provided that they do not do anything really stupid. Throwing Cuba out of the root on state dept. orders or throwing the country TLDs out for not paying the dues ICANN would like to levy would cause the root to fracture.

    Unfortunately arbitrary resolution procedures that are stacked in favor of domain name grabbers with tenuous trademark claims is not going to be sufficient to split the root.

    However it is fortunate that the root will not split just for the benefit of the likes of Idealab! startups thrown together with little thought to try to make some bucks but which really add no value.

  • That may be, but it's completely irrelevant. The internet isn't designed to use IPv6 yet, it doesn't follow that IPv6 will destabilize the internet.

    The Internet is designed to support IPv6 and IPv6 was expressly designed to provide for a gradual deployment that does not destabilize the Internet.

    DNS on the other hand was designed with the assumption of a unique owner per name and all existing applications act on that assumption. Fracturing the root would if successful cause all the names in the fracture to be worthless or less than worthless.

    Consider the value of a name vs a dotcom name. If you buy a dotcom name for $35 you can build a brand image arround it. There are plenty of pronouncable sylable strings left. If you buy a name at least 95% of the Internet will not be able to resolve it. Anyone fancy building a brand arround such a name?

    The only hope for the irregular domains is to eventually get them sanctioned by ICANN and incorporated into the root. The only value in a name is the hope that that might occur.

    Should someone really get exclusive ownership of a desirable TLD just because they took a large amount of money for names before the tld was granted?

    Domain names are in effect trademarks in their own right, the value of is that someone can send mail to and connect to the right mail server. Domain names are the routing unit of the Internet.

  • by Zeinfeld ( 263942 ) on Wednesday May 30, 2001 @01:53PM (#188673) Homepage
    One point nobody seems to have picked up yet, the documents in question are no more than Internet Drafts, they are the personal opinion of the author. The obligatory preface states explicity that they do not represent IETF policy, are not guaranteed to become standards and will be consigned to the bit bucket in six months time.

    The documents do not therefore represent a fight between the IETF and ICANN, nor do they represent the position that the IETF would take. They are simply one person's personal view.

    The threads reflect the common misconception that DNS is a yellow pages directory. It is not, it is a name service that maps names that are intended to have a meaning fixed over the long term to Internet Protocol addresses that for various reasons are subject to change over relatively short periods of time.

    The IETF has developed a yellow pages type protocol - CNRP. With CNRP you can type in 'sex' and the client will search as many catalogue servers as you like for pr0n. Queries can also be made more specific, tailoring your search for strip clubs to your geographic locality, fetishes etc.

    With CNRP it is possible for multiple people to bind to the same index term. With DNS the entire engineering purpose is lost if that happens.

    The Internet Drafts contain a massive logical falacy. They assume that conflict between 'alternative' roots can be avoided. This is not the case. Most of the domains are also hosted by other irregular roots. In many cases the other alternate roots were up and running earlier than The idea that 'destabilization' can be avoided by a central actor presupposes that that actor exists and is respected to some degree.

    Use of the alternative roots is negligible to nil. Nobody uses an alternative root for hyperlinks in public web sites or for email. The only possible use for the alternative roots is as a poor substitute for CNRP - as a service lookup. Since DNS is designed to support the type of use made of it that hyperlinks and email do and is not designed as a yellow pages the only people to be incommoded should ICANN issue a TLD that collides with an irregular one are the operators of the root and the people who paid them money thinking they would buy names.

  • Do you have an email address? Do you give it out to other people and expect that they can send mail to that address and it will get to you? Thank centralized namespace authority.

    Have you ever tried to fix a problem in the DNS, when one of the "authorities" disagrees with another?
    Ever have the TTL on your domain set to 20 years?
    Ever have an orginization take your domain name by threat of legal action, and then impersonate you?

    Thank centralized namespace authority.

    Ever send email to the wrong person?
    Did the internet become destabolized when it happened?

  • by AnotherBlackHat ( 265897 ) on Wednesday May 30, 2001 @10:37AM (#188680) Homepage
    For all the hype surrounding domain names, they really aren't anything more or less than a nickname. A simple, easy to remember word or phrase (nickname) is associated with a hard to remember IP address.

    The idea that we need a central authority to dictate nicknames is ludicrous. The idea that if nicknames collide, the internet is "destabolized" is equally silly. If more than one agency want to run a nickname listing service, then fine. If that means that when I type in "" into a browser, I go to and when you type it in, you go to so what? It's my choice which listing service I use.

    There are over 40 million registered top level domains. NSI gets anywhere from $6-$35 per year for each one. That's $240,000,000 to $1,400,000,000 dollars anually, and they don't even maintain the servers. If this is a public trust, then I'd like to see a public audit of the books.

    Just because there are shades of gray, it doesn't mean we can't tell black from white.

  • by dachshund ( 300733 ) on Wednesday May 30, 2001 @09:51AM (#188681)
    Therefore, any new TLD which conflicts with a pre-existing TLD is destabilizing, no matter where it comes from.

    Well, that may be true... But it seems to imply that everyone wildly creating new TLDs on their creaky 486 Linux box is a stabilizing thing; therefore anybody who creates the same TLD somewhere else is destabilizing the net. This is one way of looking at things, but why should it imply that the first person to get there has some particular right to run that TLD?

    I do not believe that ICANN "owns" all TLDs, or should have rights to them own them in the future. I'm simply pointing out that anyone can stake out a TLD, whether they have the resources to maintain it or not. Calling this "stabilizing" is a bit misleading.

  • by bartle ( 447377 ) on Wednesday May 30, 2001 @11:13AM (#188686) Homepage

    I really don't see what all the hubbub is about. Currently you can buy a domain name of your choosing for $35/yr, I can't think of anything else that gets you anywhere close to that kind of cost/benefit ratio. The money goes to maintain DNS servers that pretty much never go down, that level of reliability is critical for the Internet to function. I've had enough trouble just dealing with ISP DNS servers, a bunch of competing TLD servers are going to cause all kinds of problems if there isn't a centralized controlling presence.

    As for the dearth of top level names, I still don't see why anyone would care. .com, .org, and .net stopped being meaningful a long time ago and there really aren't too many reasons not to buy a .com. If your chosen name has already been taken, switching to a different TLD is kind of a piecemeal solution. In terms of branding, corporate or personal, you want as simple a domain name as you can get.

    There are also the international TLDs to worry about; it seems that these definately require a strong centralized authority to dish out. The Internet may provide the illusion of a united world, but things are still very much focused on individual countries and the international TLDs reflect that. Currently each country is given their own TLD which they can treat as national property, that system makes sense to me. Their sovereignty should not be affected by some random reseller.

    In my mind ICANN provides a much needed layer of stability and control over the Internet. For the Internet to work well, there needs to be some entity that provides such a stabilizing influence.

  • by darthtuttle ( 448989 ) <> on Wednesday May 30, 2001 @11:14AM (#188688) Homepage
    Postel's probably turning in his grave. Granted, there are a lot more computers and networks on the internet, but looks what ICANN and Verisign are, and what Postel was. I miss the old days.

    I think we have to understand that this is not a fight we are going to win. We have lost control of the internet. It is in the hands of the companies now. But now what?

    I think it's time to look at alternatives to "The Internet." It is well within our means to simply use the Internet as a transport method and develop our own networks and interconnect them if we choose. Ok, so some of our ISP's don't want us VPNing, but we can switch ISP's usually.

    The question is, is it time to walk away from the public, develop technology, and let the public come to us (again)?
    Thought Architect
  • by Spiffy Biff ( 451062 ) on Wednesday May 30, 2001 @10:16AM (#188689) Homepage

    The internet was founded on the idea of routing around damage. How do we route around the cancer that ICANN is becoming?

    You really should check out OpenNIC ( It's a (seemingly) democratic organization that recognizes the ICANN root, legitimate alternate roots, and its own namespace. What this means is that you can join immediately, your DNS won't be "broken," and you can have a say in how the DNS namespace will be organized. I've only recently signed up, so I don't know yet whether this is the solution, but at least I feel like I'm no longer part of the problem.

    P.S. The astute reader will note that I have registered "" in the ICANN root. I can only say that it was done some time ago, before I understood my options. I will not renew it.

  • I haven't read the RFC's myself, and I don't believe I need to. Two things should be apparent to anyone with reasonable intelligence: 1) The internet is the greatest tool for human communication since spoken language; and 2) Any central authority, even if it is just a naming authority, is incompatible with the basic requirements of free expression, and thus should be regarded as a (logical, practical, ideological) contradiction of that freedom. Someone once said something along the lines of "free market economy is the worst system - unless you look at the alternatives". I believe that applies here, as well. Regardless of the technical issues, any centralization of the internet is a blow to the fundamental concepts held by the people who make the internet work. The worst part is, ICANN's power is largely political in nature - and as with all things, any ideology that is predominant is a threat to the minorities, as long as that ideology is based on the idea of control (although I suspect ICANN and it's supporters are the minority, in this case).

  • I never recommended wild abandon in our internet naming system. I did not, in fact, offer any specific alternative. What I said was that a central authority with the arrogance of ICANN is wrong. I have no problem with a central authority, as long as it is chosen by the people it governs. The USA was founded on the idea of 'consent of the governed' as it's basic license for governing power - The internet deserves no less noble of an ideal.
  • Gah, I have been unclear, thus seeming to contradict myself. Let me clarify.

    Anarchy is incompatible with freedom, as it is simply a form of mob rule. (We won't go into the whole 'anarchy is a system based on the denial of the validity of systems' bit..). Democracy is not much better, it is simply organized mob rule - but at least with a democracy, one has the right and power to suggest alternatives and attempt to sway the system. Look at the US government - a central authority... but we, the people, have the power to affect it, overrule it, blow it off the map if we feel too threatened (at least in theory). Under an unaccountable organization such as ICANN, an individual has no reasonable channels for voicing of concerns, complaints, etc. Vast power is consolidated in the hands of a few who sit on a panel, and individuals have no direct representation on that panel. THAT is my specific complaint. I believe it's generally bad form to complain without offering at least a possible solution, but I have none at present, other than a vague idea of a compartmental naming authority with directly-elected members.

A consultant is a person who borrows your watch, tells you what time it is, pockets the watch, and sends you a bill for it.