ICANN Approves generic top-level domains | Hacker News Reader

ICANN Approves generic top-level domains

icann.org

200 pointsgert136 comments

Meh, I doubt it'll have any major effect.

The vast majority of TLDs that are introduced fail very quickly (.name, .biz, .pro, a range of ccTLDs whose country's citizens opt for the 'big 3' - .com, .net and .org - regardless, etc)

All that'll happen here - IMO - is that a few of the bigger names will waste $185k on their unique TLDs, realise that their users get confused by them (never underestimate the ability of big companies to misunderstand it's users..) and either switch back or stick ridgidly to using them with no positive effect.

So yep, I'm not too concerned about this. ICANN aren't a great organisation, and they seem to mainly bring out new TLDs to get more money.

</Rant>

So yep, I'm not too concerned about this. ICANN aren't a great organisation, and they seem to mainly bring out new TLDs to get more money.

ICANN is a non-profit whom CEO gets ~1M/year. Executives earn a bonus of more than 50% of their base salaries if specific goals are reached, such as adding a new gTLD.

Source: http://www.thedomains.com/2010/02/11/nonprofit-icann-release...

I really hope this top-level domain thing isn't going anywhere.

For comparison, the CEO of the American Red Cross has an annual salary of $500k.

I don't think this is a relevant metric at all for judging a non-profit's goals or performance. There is a certain point where the organization has become so big or important that any person whom the stakeholders can trust to run it will command a high salary.

So as a stakeholder in a non-profit, I'd want the non-profit to be as efficient with money as possible... yet I'd want a a CEO that burns through way more money than he needs? Please explain.

I'm not the grandparent poster, but I think he's saying that executives skilled enough to run the Red Cross are motivated primarily by their salaries, and so those salaries have to be competitive with what the same highly-skilled exec could make in the private sector.

I might argue that the situation is in part a matter of "like attracts like". I've you're going to get executive types to contribute and to encourage/force their workforces to contribute, you have to present them with an "executive" type interface. (If it walks like a duck and talks like a duck... and plays golf like a duck...)

My thoughts next turn to "hypocrisy". These are supposedly charity endeavors. But much of their activity -- fundraising activity, at least -- is just more status seeking and building and fraternizing.

There, for me, lies the heart of the matter. When in the minds and/or actions of the top level participants, charity work is anything but.

(I'm "on the outside" looking in, and perhaps I'm too cynical. I want to believe that people have more depth than this, and some individual acquaintances who travel in those circles sure seem to. But it makes me wonder.)

No wonder they are not more effective.

I'm not entirely comfortable with my comment, but I'll make it for the sake of the point about "like attracts like".

Efficient is not strictly synonymous with cheap.

It's not strictly synonymous with "expensive" either. If it correlates, it probably correlates in the direction of "cheap", if only marginally. And any way you slice it, "expensive" is synonymous with "expensive".

Let's come up for air here and ask the simple question: do we believe we have better domain policy because we pay some guy at ICANN a 7-figure annual comp package?

I don't believe that. Do you?

Well, define "better domain policy"...?

Due to its ARPANET origins, the Internet namespace was technically property of the United States. Obviously that situation had to change, or the other nearly 200 nations on the planet would have difficulty accepting the Internet as the true global network.

The ICANN is necessarily a compromise between the interests of powerful nations. It's not reasonable to assume that it could be run completely differently from all the other global intergovernmental organizations. Most politicians don't consider Internet policy to be all that special compared to trade, foreign aid, patents and everything else that needs to be similarly balanced and negotiated. Hence, the head of the ICANN is someone who's acceptable to governments foremost.

I'm not really inclined to figure out what the appropriate compensation for ICANN's CEO is. My comment was not about ICANN in particular.

Gail McGovern, the CEO of the American Red Cross, makes $995,000. Charity Navigator gives the Red Cross 3 of 4 stars.

By contrast, Ophelia Dahl of the 4-star rated Partners in Health makes $86,000.

PIH is a smaller organization than the Red Cross, but there are people who could argue that PIH is also more impactful. It's also the case that part of the reason the Red Cross is so much larger is that it's a key mission of the Red Cross to be larger. It's not unlikely that charities with highly-compensated CEOs share that trait. You pay a CEO a lot of money because their market value is high. What makes a manager-type CEO's market value high? Sales and financing skill.

It's absolutely reasonable to point out and question the salary of a nonprofit's CEO.

> By contrast, Ophelia Dahl of the 4-star rated Partners in Health makes $86,000.

That's still tens of thousands of dollars. Can you point to a CEO of a non-profit making less than $10k? Less than $1k?

http://kittyrescueandadoption.org/how_to_help.html

"The shelter depends entirely on the generosity of others to sustain our efforts. There are no paid personnel and all monies/donations are used to care for our cats. KRA is a non-profit 501 (C3) corporation and donations are tax-deductible."

$86k in many places (say NYC or SFBay) is (just below) a living wage. Must all do-gooders be penniless?

SFBay average is about 76K:

http://www.bayareacensus.ca.gov/bayarea.htm

Equating 86K a year to penniless is a statement very very far removed from reality.

My point's just: there are CEOs of highly effective well regarded charities who are not taking 7 figure salaries. It is not true that charities basically have to shell out 7 figures because that's what people who can manage charities cost.

I don't understand what point you're trying to make.

I'm sure there are many instances where the chief executive of a charity does not take a salary or personally contributes more to the charity than their salary.

It's reasonable and it tells you something about the organization's values, but it only goes so far. Overhead costs are a poor way to judge a charity's effectiveness. More:

http://blog.givewell.org/2009/12/01/the-worst-way-to-pick-a-...

It does not ring true to me that looking at overhead costs are the "worst" way to pick a charity. Is there more to the picture? Sure. Is the Red Cross as impactful as PIH dollar for dollar? I doubt it. And if you look at the numbers and engage critical thinking, instead of just setting an arbitrary "overhead" threshold, you can spot things like growth for growth's sake, or fundraising for fundraising's sake.

If profit is their motive, why didn't they do this 20 years ago? That's 20 years of lost profit.

Market timing. 20 years ago a top level domain wouldn't have been such a valuable asset.

ICANN is less than 13 years old. The Internet only went fully commercial 16 years ago.

~1M a year!? That poor CEO...

Agreed -- if anything, it will just make .com domains even more valuable. The average Internet user won't understand these new TLDs.

Really? I mean how valuable is a .com domain to an average user? As far as I can tell, most people don't type in URLs at all, they type the site name into Google or have a bookmark. They mostly don't even look at the url, they look at what the page says. If you want to send the link to someone, you copy the whole thing anyway. If you do a little word of mouth, you probably won't include any TLD at all, so Google is again your target destination. I don't have any knowledge on how Google weighs TLD, but I highly suspect that they don't care. It's about content, they might even favor specific TLDs like "mybusiness.usa" over "mybusiness.com"

Domains couldn't be more irrelevant right now, and adding more TLDs is only exacerbating the situation.

> I mean how valuable is a .com domain to an average user?

> If you do a little word of mouth, you probably won't include any TLD at all

From my experience, mentioning .com verbally after the domain name is what makes regular people realize that they can put that name in Google and reach the website.

Google most definitely does care about the TLD. It's well known that .edu pages are given more weight than others.

The average user types 'facebook' and through the magic of browser search gets to Facebook.com already. This just cuts out one step.

How do you decide that a TLD has "failed"? Do the domains still resolve? They are obviously not going to attain a critical mass of public awareness overnight.

If startups are willing to go with goofy abuses of ccTLDs like .ly and .us then they would likely accept or even prefer gTLDs, when there is a rich variety of them available.

The domain scarcity problem will only get worse with time. Unviable options will eventually become viable by sheer necessity.

Let me illustrate how they failed with the example of the .aero TLD. When I want to visit the website of British Airways, one of the largest airlines in the world, I typically Google for "ba". The first search result is britisharways.com, not britishairways.aero. Now you can repeat the same experiment with all the other major or minor airlines, and the top search result always will be a .com (or a ccTLD) domain.

This how you can tell .aero has failed; it's main target users are simply not using it.

If you already have your .com then you don't need new TLDs. But .aero may come in handy the next time someone wants to start a new airline (which is admittedly pretty rare, so .aero in particular was probably a dumb idea).

I think the succesful gTLDs will be the ones that appear commonly in trademarks like photo, pizza, sports, mart, lube, etc. It may become much more common for the domain to be the primary trademark, even for businesses that aren't particularly internet related. Fredspizza.com would be a confusing name for a restaurant but Freds.pizza less so.

> But .aero may come in handy the next time someone wants to start a new airline

The problem is that ICANN has already set it up so that if you have the trademark on "flying.com" that nobody can get "flying.aero" (well they can, but you can take it off them easily). So why would you even bother with the .aero domain unless you own the .com too?

This is simply a money grab by ICANN.

You make .aero part of your trademark e.g. crazyclown.aero. As long as crazyclown.com is not another airline, you should be fine.

You're very unlikely to get away with that. Try starting facebook.aero and see how long it lasts.

If ICANN goes by trademark law, then you should be able to register something generic like friendly.aero without infringing on friendly.shoestore. If they don't then they are being stupid. If there are going to be hundreds of TLDs, you can't automatically protect every trademark in every TLD. Trademarks are not supposed to work like that.

They might not be supposed to work like that, but they do under WIPO.

You can see results here: http://www.wipo.int/amc/en/domains/decisionsx/index-gtld.htm...

Yeah, but who's actually going to do that?

Apparently, just about nobody.

I don't understand the situation well enough to say if it is a "money grab" or not, but it seems like another good reason to abolish intellectual property.

Doesn't sound like a failure at all; you navigate using Google. I would assume this is typically where their traffic comes from. With that being the case it might as well not have a TLD at all; it's all but invisible to the user. The domain could very well have been .aero and you still would have navigated to it successfully. Isn't that all that matters?

Ok here's another example: easyJet, one of the largest budget airlines in Europe, paints easyJet.com on the sides of their airplanes, not easyJet.aero, as you can see on this photo:

http://www.easyjet.com/common/img/photo030.jpg

What I'm saying, airlines simply don't use the .aero TLD.

I think you're missing the significance of this change. It's not clearly stated in the announcement (in fact, nothing is) but in section 1.3 of the Applicant Guidebook[1] it shows that for the first time it will be possible to have IDN gTLDs, that is, gTLDs in any standardised script. Latin1 is gone. This is big.

Sometimes the anglocentrism (and USA-centrism) on this site worries me.

[1] http://www.icann.org/en/topics/new-gtlds/rfp-clean-30may11-e...

I think you've missed a couple of IDN gTLDs, some launched more than a year ago, i.e. http://www.thewhir.com/web-hosting-news/111710_470000_Domain... .

They are country-code top-level domains. Currently IDNs are limited to country representations only.

The current non-IDNs are also limited.

Which makes the issue of IDNs unrelated to this discussion.

The anglocentrism on this site is nearly entirely because the overwhelming majority of developments in the computer industry, from the Mark I to the iPad, have been created by Anglophones.

Let's keep this tangent short but, so what? It doesn't mean we need to be blinkered about what the gradual relegation of English as tech lingua franca might mean for non-anglophones. Do we really need to be in a minority to be interested in remote perspectives?

Personally I want the Internet to bring the world together, and I see IDNs as a step toward maintaining the existing segregation between nations and languages.

And to do that you force everyone to use english?

English is a great lingua france but there are parts of the web that are only targeted at speakers of one language. Why should they have to type strange stuff into the address bar to get there?

Ausserdem habe ich so meine Zweifel ob du mich ploetzlich verstehst nur weil ich ISO 8859-1 verwende. Man kann die fehlenden Zeichen auch umschreiben.

Google and Mozilla have both stated intentions to get rid of the address bar; the number of users that will even see IDNs is a shrinking market. Rather than "typing strange stuff into the address bar", in a year or two most people will be "typing native stuff into a search box" -- if they aren't already.

This move by ICANN is already irrelevant.

> Ausserdem habe ich so meine Zweifel ob du mich ploetzlich verstehst nur weil ich ISO 8859-1 verwende. Man kann die fehlenden Zeichen auch umschreiben.

"I also have my doubts whether you understand me suddenly just because I use ISO 8859-1. One can also rewrite the missing characters."

(Provide your own translation to the dominant language or rely on software to do it for you. That was Google's translation service.)

I think you missed the message.

The majority is actually not as overwhelming as it appears at first sight if you think of people like Linus Torvalds, Guido van Rossum or Bjarne Stroustrup.

BUT, I think most IT folks who are not natively anglophone have wholeheartedly embraced the english language as the lingua franca of the industry and of science in general. The practical advantages of having such a lingua franca are simply overwhelming, and I really think that any step towards localization is a step backwards for everybody.

Very good point. The development of Linux I think of as an English-speaking effort, but that's because Linus conducted all the discussion about it in English (because, as you say, English is the lingua franca of the technology world).

So I guess the actual truth is that the vast majority of technology developments have come from people who speak English (some Ruby stuff excepted), not necessarily native English speakers.

It isn't anglocentrism. It is that English is the language of science and diplomacy. If this whole crazy thing had started in the 18th or 19th centuries it would have been in French or German. If you want to communicate to the world in scientific literature, you publish in English. I think that this spilled over into technology for obvious reasons. You can argue owe the merits of this, but having a common lingua Franca (if you will) was important to bootstrap the Internet. Now that it is more commonplace, it is important to support more languages. But if we had started out supporting everything, it might not have gained traction. Or it would have balkanized and been multiple networks separated by language. (which is what we might get anyway)

People can register unicode domain names (not tld) for years now. Do you know any major sites with a non-latin primary domain name?

I don't get around all that much, but the biggest one I've seen was Daring Fireball's web shortener at ✪df.ws . It was later pulled due to widespread software problems with Unicode glyph handling, with a big offender being AT&T's name servers: http://daringfireball.net/2010/09/starstruck .

If it's not clearly stated, then perhaps people didn't miss it because of their bias but because... it wasn't clearly stated.

Whether a raft of new generic top-level domains succeeds or fails depends on whether there is a mind shift in the average user from ".com" to ".anything". IMHO, the new entrants like .biz haven't succeeded because there hasn't been that shift yet.

Now, if Google starts using .google for its products, and other key services you use every day stop using .com as their primary URLs, I think that mind shift could happen and you could get a very different result.

just names - one day it will be without www or top level domains just the pure company name or service. Like tinyUrl and other rerouts this will become possible and the background of the domain structure will only be a registration and admin thing.

That works just fine today. Type the name into your address bar and it will either go right there, or google it for you.

Technology has made this move by ICANN obsolete. However the big problem is if you're a large company you HAVE to protect your trademark, so you'll have to register these other domains anyway - meaning it's just a pure money grab from ICANN.

I wonder who's going to register "localhost" or "mail"? ;-)

EDIT: There's a list of reserved TLDs that won't be allowed to be registered. It includes localhost (but not mail)!

AFRINIC IANA-SERVERS NRO ALAC ICANN RFC-EDITOR APNIC IESG RIPE ARIN IETF ROOT-SERVERS ASO INTERNIC RSSAC CCNSO INVALID SSAC EXAMPLE* IRTF TEST* GAC ISTF TLD GNSO LACNIC WHOIS GTLD-SERVERS LOCAL WWW IAB LOCALHOST IANA NIC

(example and www have asterisks next to them because "in addition to the above strings, ICANN will reserve translations of the terms “test” and “example” in multiple languages. The remainder of the strings are reserved only in the form included above.")

localhost would be useless as it's in the hosts file by default. *.localdomain might be interesting. It's around, but I don't think it's the default (I don't know anyone that even uses it though).

*.local is used by all Apple devices for Zeroconf (Bonjour.)

Imagine your Mac's hostname is uranium, uranium.local will automatically resolve to the current IP of the laptop.

According to their Applicant Guidebook:

String reviews (concerning the applied-for gTLD string). String reviews include a determination that the applied-for gTLD string is not likely to cause security or stability problems in the DNS, including problems caused by similarity to existing TLDs or reserved names.

So my phishing TLD .corn is out?

It wasn't until you called it a phishing TLD. ;-)

I always thought .con would make a good phishing TLD; it could be reserved for only bona fide criminals.

.mail would be excellent, really. All we need is $185,000, a stomach for paperwork and an established company looking to expand into the registry business.

Shame they aren't allowing www - then we could have had com.foo.www to remind us of JANET addressing.

or java package names

How possible would it be for an individual to register one of these, assuming they had the $185k figure I've seen floating around in articles related to this? I doubt it would be a useful or smart investment, but I just wonder how theoretically possible it would be for someone (or a small company they setup, if ICANN requires that) to buy one for personal use (e.g., for myself, me.grantpaul or something similar). Edit: this appears to be impossible, according to the guidebook: "Established corporations, organizations, or institutions in good standing may apply for a new gTLD. Applications from individuals or sole proprietorships will not be considered."

Or, alternatively — could a (funded) startup buy one? E.g. could AirBnB buy .bnb and then have sanfrancisco.bnb direct to the relevant San Francisco AirBnB page?

I've done my best to skim over the (huge) Applicant Guidebook and as far as I can tell, you don't need any reason per se to be granted a gTLD.

You have to be a corporation or organization with good legal standing and financial stability, and you have to slog through an epic and expensive beurocratic process. Your application can be rejected on various specific grounds, some technical and some political.

But I don't see anything that explains the purpose of gTLDs. So essentially, if there is no particular reason for you not to have one, you can have one.

Can I get .movie and sit on it until the movie industry pays me millions for it?

It does at least say that you need to run a registry. So you can't just squat a gTLD. But I don't know, in general, how the prices and policies of the registrar are evaluated and enforced. It might be in there somewhere but it didn't jump out while scanning.

Registry fees alone will cost you a minimum of $100k a year even if you sell no domains. You also need to submit business plans and "proof" of economic viability; without those your application will be rejected. One factor that most people aren't aware of is that ICANN will require you to have enough money in reserve to keep the registry alive until a new buyer can be found in case things go bad; this can be as much as $4 per domain sold in cash as reserves. If you predict you'll sell 4 million domains you'd better account for the fact you'll need to keep $16M in the bank.

IIRC if multiple companies apply for the same TLD it goes up for auction.

What I've gathered is that they require licensees to run a registry for the given TLD. If you are not operating a registry or are administering it poorly, they may revoke your license.

Maybe I'm just a hippy at heart, but that $185k really rubs me the wrong way. It's a step back to the bad old days of broadcasting licenses that only big companies can afford to buy — and given the stagnation of old media that makes me a bit glum.

The main reason it is $185k is because the process has dragged out so long; ICANN _must_ make back any money it spends on the process. The last round was $45k and in the future when next round comes around it will probably be less than $185k, but most likely more than $45k. All the background checks and due diligence costs money.

What background checks and due diligence? What background are you supposed to have to own a TLD, anyway?

If you start a new TLD, ICANN wants you to at least stay in business for the foreseeable future so that your customers don't get screwed. You also need to be able to survive the inevitable natural and artificial disasters that befall infrastructure.

So $185k buys what essentially amounts to the right to "broker" a gTLD suffix. This is building into the system a prohibitively high barrier to entry, which almost guarantees that the interests of the conglomerates will once again trump the small, innovative and more efficient.

More details on why the $185K -- http://www.bbc.co.uk/news/technology-13835997

My fear is that brokers won't be the ones to buy it: For example a big media company could get ".movies" to control a domain space and use it as an anti-competitive tool against indie film makers who don't even get in the front door. Look if Disney wants ".disney" I'm cool with that, but I don't like the idea of Disney owning ".animation".

We might as well just go to an AOL Keywords system.

For a disturbingly large percentage of the population, Google is that system.

Why is that disturbing?

I like to live in a fantasy land where everyone understands the basics of how the Internet works. The reality that many people don't have the faintest idea how to use a computer beyond the very, very basics -- and probably never will -- is unfortunate, IMO.

They do have an understanding of how the internet works. You type what you want into google and you get it. Maybe its not how the internet worked before, but hey its nice to have to not keep downloading HOSTS.TXT.

These days it seems to be Facebook keywords

Alright, who has the cash for .yc TLD? Let's do news.yc, paul.yc for all our yummy essays, apps.yc for applications, and search.yc for the purrfect HN search.

Srsly ICANN will make more bank than it ever has now.

Isn't 2-letter for countries? I think other TLDs have to be 3 or more.

yc evaluates to yyc so you could use that

http://en.wikipedia.org/wiki/Fixed_point_combinator#Y_combin...

Actually, I got that wrong: yc -> cyc and you couldn't really use that!

Losing the latin-1 restriction is a mixed bag.

The main advantage is that it's respectful to folks of other writing systems.

The main disadvantage is that unicode characters that look visually identical, but which are logically distinctive, will be inevitably be used for more elaborate phishing attempts.

With a tld? These get individual approval, with pretty strict scrutiny. No ones going to let me register a new tld '.cоm' where the 'о' is cyrillic.

Give it time. The safeguards will be steadily eroded by money.

if you used regular-expression to validate websites (and/or emails), you better be prepared.

I'm going to miss the Chrome (IE9 too) combined search/address bar. No more will I be able to type in "haircut" or "bank" and have my browser know to kick off a search as there's no clear distinction from TLDs anymore.

From a purely marketing perspective, non-standard top-level domains (heretofore referred to NG AOL keywords) are a problem. If I see an ad for http://cameras.canon, I'm not sure I recognize that as a URL.

If your email address is joe@sony.electronics, I'm not sure people are going to understand it. Oddly, if it were just three character gTLDs, I think people would be fine by it.

Still, I want .local. That way I can break every Mac on the planet all at once.

I wonder what the process for getting a new one approved would look like. For example, would Google be able to say, hey, we've got a large online presence, can we have the "google" gTLD?

If not, we have problems like it being possible to continuously create new TLDs and domain-squat on google.whatever or such.

If so, well, that's interesting, I actually can't wait for Google's url to be http://google. Maybe one day even //google.

Good times, good times.

Somehow I suspect the process will involve forking a bunch of money over to ICANN.

You mean I can't get one for $7/year on GoDaddy?

Just wait. They'll sell the first few for hundreds of thousands (millions?) of dollars to Google, Pepsi, and Sony, and in 10 years they'll be on GoDaddy fo $20.

From http://www.icann.org/en/topics/new-gtlds/factsheet-new-gtld-...

> How much is the evaluation fee?

The evaluation fee is estimated at US$185,000. Applicants will be required to pay a US$5,000 deposit fee per requested application slot when registering. The US$5,000 will be credited against the evaluation fee

> Are there any additional costs I should be aware of in applying for a new gTLD?

Yes. Applicants may be required to pay additional fees in certain cases where specialized process steps are applicable, and should expect to account for their own business startup costs. See Section 1.5.2 of the Applicant Guidebook.

And the guidebook (http://www.icann.org/en/topics/new-gtlds/rfp-clean-30may11-e...) mentions a number of different fees.

I was trying to determine: what portion of that, if any, is refundable should an application be turned down?

If you withdraw after the 1st phase of the application (application submitted and all applicants are visibile for that TLD) then you are entitled to 70% refund, the second phase (once all eligible applicants have withstood the examination process and survived contention) 30%, when the bidding starts - nada.

First thought: will have to rewrite my email regexps...

ha yeah. i was amazed when i found http://www.ex-parrot.com/~pdw/Mail-RFC822-Address.html

(via http://stackoverflow.com/questions/201323/what-is-the-best-r...)

If you have to rewrite them to accommodate new TLDs, you did them wrong in the first place.

The question is will the consumer buy into any new TLD's?

Seems to me that this is in part just a mechanism for ICANN to setup new revenues for all concerned. And let's not forget with search engines trying to make you, the average consumer, forget about TLD's altogether it seems a battle is shaping between TLD's and the search engines.

Are we applying for .hacker or .yc? it'd be cool to be able to identify all yc companies instantly by domain.

Doesn't linking to icann's front page mean that nobody can submit it again without being sent to this item id?

The dupe checker only checks the cache, and stories expire fast enough.

So much for neat and clever url hacks. Now everyone can do something like del.icio.us or script.aculo.us. :(

Everyone with $185k anyway.

which for many big companies is a drop in the bucket for marketing... :(

Won't be as much fun if you control all the parts of the domain name, since there's no need to be "clever" with a ccTLD.

"The Board approved a plan to dramatically increase the number of Internet domain name endings [..] from the current 22, which includes such familiar domains as .com, .org and .net"

Where do those 22 come from? doesn't each contry have its own TLD?

Countries have ccTLDs which are "different" from TLDs. The 22 are generic TLDs: http://en.wikipedia.org/wiki/List_of_Internet_top-level_doma...

These 22 are only gTLDs, which are not the only type of TLDs. There are also ccTLDs.

So for a few hundred thousand dollars I can actually buy clownpenis.fart ? Of course i'd have to hit the lotto first and even then saturday night live might sue me... maybe i should have gone into IP Law?

This would be a good thing if new TLDs represented the types of organizations, e.g. .news, .reviews, .cars, .health...

I'm afraid the real TLDs will be either more random and meaningless or trademarked.

Anyone think it's a coincidence that ICANN does this just after plans are announced by major browsers to eliminate the URL address bar? Are they rushing to sell-off?

Also, what about the possibility of disputes? What if both Facebook and twitter want to purchase '.social'? Who will resolve the dispute and on what grounds?

The meeting in question was for final approval; ICANN approved plans for opening up gTLD's back in 2008.

http://www.theregister.co.uk/2008/06/26/icann_approves_custo...

This didn't suddenly happen overnight; everyone was informed three years in advance. Just to illustrate how long ago that was in internet years: the first Windows public beta of Google Chrome wasn't even released until after ICANN's initial announcement.

The necessity of this, the price for which it is offered and all those other matters aside...

Iirc if you run your own DNS-server/bind-instance (for DNS lookups, not just responding to internet DNS requests), you have a options file which contains references to forwarders for all the TLDs.

How will managing new TLDs from a server/bind-perspective play out? Will there be a new, hidden ".global" TLD you can forward, or "default" forwarder you can configure to handle this?

Basically I'm not sure how this will play out with my bind-setup and if it will require any action from me as a server-admin, apart from "apt-get update". Anyone here have any inside info?

For at least one DNS server I've setup (I don't know much about it, to be upfront) all requests for which that server did not have an authoritative response were simply forwarded to an upstream ISP server. Maybe this is an atypical setup, or I'm describing a different problem to the one you're envisaging.

"Make it your upstream provider's problem" isn't a very helpful piece of advice if you're running that provider, I suppose!

Your DNS server will simply ask the root servers. Those are listed in a file (root.hint on BIND) and will tell you which server is authoritative for the TLD in question.

I'm not aware of any setups explicitly storing the TLD servers themselves (of course, they'll be in cache.)

...Brought to you by ICANN(R), the organization that managed to make misspellnigs a viable business.

Cool, they have just broken most of email regexps.

Funny, but email validations should already just check that the address contains an @ symbol and leave it at that.

They should also check that neither part is empty. But other than that you are right.

uh, nope. Email validation should check there's a valid MX record for the domain in question, and leave it at that.

Or a valid address record. Mailers fall back to 'A' records in the absence of a 'MX' record.

yes, and we could continue into checking something is actually listening for mail there, etc. But you get my point :)

No. There are 307 top-level domains right now and ICANN has added several dozen top-level domains in the last year and your regexp is probably already broken. You can be sure anyone using regular expressions has not been matching them correctly for some time. If you are using regexps, you're doing it wrong.

Legacy is a drag on any system. It has to be supported for those who followed best practices but not for those who followed worst practices. Don't hold progress back for them.

Reply on news.ycombinator.com