TLD Graveyard
dzdb.caida.org
dzdb.caida.org
http://domainincite.com/26198-breaking-failed-org-buyer-etho...
https://domainnamewire.com/2020/11/19/donuts-is-acquiring-af...
https://domainnamewire.com/2021/01/22/breaking-ethos-capital...
List of gTLDs run by Donuts: https://icannwiki.org/Donuts
Or you can just google "donuts abry acquisition dot org"
Or you can google "Ethos capital .org": https://www.google.com/search?client=firefox-b-d&q=ethos+cap...
Or you can google "Fadi chehade ethos capital": https://www.google.com/search?client=firefox-b-d&q=fadi+cheh...
In my personal opinion it's nothing but a bunch of rent-seekers, in the classical economics textbook sense of the word.
DNS is quite literally all about names, it’s what the N stands for.
Can you give an example where name ownership was unimportant to folks using DNS?
DNS used to be a text file that people would pass around. Now… this. Something has gone terribly, terribly wrong.
Even for TLDs I think the problems/issues tends to be more about operations of the TLD, because it’s generally a little more work than just a zone under a TLD.
But yes. Changing ownership of a TLD that’s operating under existing guidelines for all the domains that use it, is a bad thing. I’ll caveat that with, if it can no longer be operated by the group doing so it should change ownership, but the same protections for the zones inside it should be preserved in the sale.
Because one private for-profit company holds the sole rights to register things on it, and can charge what they want?
Why should ICANN be giving a license to private rent-seekers the ability to charge more than a fair profitable price for something that exists in a fully-automated software purchase/checkout/zone file entry work flow?
- .an was the Netherlands Antilles, now replaced by its constituents .cw, .sx and .bq
- .tp was East Timor, now renamed Timor Leste (.tl)
There are plenty more not listed here though, including .su (Soviet Union) and .yu (Yugoslavia).
Come 2091, people will celebrate the 100th anniversary of the fall of the Soviet Union, and .su will still be alive and kicking.
But I'm sure somebody somewhere is going to be celebrating.
America's hawkish (neo)conservatives continue to celebrate the fall of the Soviet Union. "Presidents Reagan and Bush Sr beat the USSR!" (Maybe that is an oversimplification of the truth, but it sounds good to their ears.) In 2091, there is likely to exist some intellectual successor/descendant to that school of thought, eager to mark that centenary.
What do the Russian Orthodox think? They went from an authoritarian regime which officially opposed the Russian Orthodox Church to a (still rather authoritarian, but arguably somewhat less so) regime which wants to be publicly seen currying favour with it. I'm sure some of them view that as an improvement. If that's what some of them think, why wouldn't they celebrate in 2091?
What about the countries which gained their independence through the downfall of the Soviet Union? I think many people in many of those countries will be celebrating the centenary of their independence.
For sure. https://imgur.com/gallery/L9VNpP4
heh
There are probably a lot of firms that serve many of these countries, and having a single unified presence at foo.su is probably more efficient than registering foo.ru, foo.ua, foo.kz, foo.lv, etc.
1. Sites that were created back then and just stayed (a very low number).
2. Sites of organisations operating in most of the ex-USSR or CIS countries.
3. People who are nostalgic or associate themselves more with the Soviet Union than with any of the states that exist in its territory today.
4. Just-for-fun types of people.
In Soviet Russia, top level authority expires you!
Was the first one funny but then people kept making more which aren't funny?
Why is it a joke to use sentence structure from another language?
Also, OPs post was actually clever, which can't be said of 99% of these things
As for the foreign language part, it's probably meant to lean into stereotypes about Russian English speakers to emphasize the "trashiness" of the joke. You could have an ethical discussion about that.
Because the template can be filled with a wide variety of specific nouns and verbs, it has become a bit of a snowclone[1] or a meme in its own right[2], like the "Xzibit Yo Dawg" meme[3].
[0] https://en.wikipedia.org/wiki/In_Soviet_Russia
[1] https://en.wikipedia.org/wiki/Snowclone
> and the inverted Soviet form something menacing or dysfunctional, satirizing life under a communist dictatorship
A proper use of the "In Soviet Russia" meme isn't just reversing the words to make it sound like "Russian"; it' supposed to also suddenly construct a sentence that implies something dystopian. (And ideally, somewhat humorously, usually contextually so.)
I think the right ones are generally only at "sensible chuckle" level — you're not going to die laughing usually. I don't think the one above really fits the template though.
- American guy goes to the dentist and says "do it fast!"
- German guy goes to the dentist and says "do it well!"
- Norwegian guy goes to the dentist and says "does it hurt?"
It is a thing in some cultures. Norwegians poke fun at the Swedes (and ourselves), Englishmen make jokes about the Irish and probably the other way around too.
I hope we can agree that it is OK also in the future. I don't want the world to be too dull.
This is true not only for jokes but for pretty much every good idea. More often these are just a reference to something that is (was) funny, like movie reference or quote.
I don't know why you think this references language structure (see next point).
OP may have been clever (or not), but in my reply I sought to assert the canonical form where you take 'A <verb> B' and write 'In Soviet Russia B <verb> A'.
Namebase is a marketplace atop of the Handshake protocol. Puma Browser supports it natively, as well as NextDNS. If you really want your own TLD (or even to have your own name as a TLD), Namebase is the way to go (for now).
Maybe I should've added this information to my initial comment, I seem to have angered some folks.
Every single one of them has the exact same problem: the purpose of a domain name system is to agree upon naming. Alternate roots break that model, and any attempt to use them seriously immediately runs into massive interoperability problems (e.g. cannot send/receive email, cannot obtain SSL certificates, etc).
However, I can speak to your last points.
Email sending between addresses utlizing Handshake domains has been demonstrated. I am unable to find any tweets but here's a video explaining how to do it[0]. Concerning security certificates, this is possibly via DANE rather than a centralized authority like LetsEncrypt. There is a tutorial[1] for that as well.
Regarding name collisions, Handshake's purpose isn't to replace legacy DNS, but to build atop it. It is for this reason that the Alexa top sites aren't available for bidding on Namebase. Again, I haven't looked at the amateur projects you've listed but it's possibly Handshake's focus is a key distinction.
--
Which fails to solve the problem of being able to interoperate with normal people on normal email servers.
> Concerning security certificates, this is possibly via DANE
DANE is dead. Nothing supports it, and there are no meaningful plans for that to change. Given that DNSSEC is still often dependent on RSA-1024 for security, it's not clear that DANE would even provide an adequate level of security.
> It is for this reason that the Alexa top sites aren't available for bidding on Namebase.
This is a bizarre assertion, given that Handshake is selling access to what are effectively TLDs, not domain names. A more believable explanation is simply that Handshake was attempting to avoid selling trademarked names associated with companies that would object to this.
"believable explanation" is subjective. I don't work for Namebase and am not involved in Handshake development. I just like the project. (:
We're still in the early days so your complaints may be met with solutions.
.cs was the first removed TLD as far as I was able to find.
I’d personally love to see a proposal to drop puny-code and adopt utf8.
One issue that I see is that the UIs where this is rendered, render the utf8, ie the look alike code points. So if it’s to protect users, then we’d need to present the puny-code and not the rendered native language text. But, why would do this? It would defeat a primary reason for DNS, to bridge humans to networks. So I just don’t buy it.
So the look alike issue is still there. So if it’s not the users, then who are we protecting from the look-alike characters? It’s definitely not a benefit for those of us who’ve written the code to parse and store DNS names (at least from my personal perspective).
The example of umlaut is really good one though, as the umlaut can be an extra code point, or part of the actual character. I can see why we'd want to be explicit that both are not allowed in DNS...
It still feels like we're solving this issue at the wrong layer, i.e. the protocol layer, as opposed to the application layer. I'm 100% in agreement on needing solutions here, but I think your example illustrates that it's a higher-order issue for applications, and not the protocol.
Yes. 手表 (shóubiǎo) means wristwatch, with 手 (shǒu) meaning hand and 表 (biǎo) referring to a time piece.
The verb watch as in "watching a movie" would be 看 （kàn）as in 看电影 (kàn diànyǐng).
At this point, why not just accept any string of characters up to a certain lenght as a TLD? We've already lost the advantages of having fewer TLDs, no one can remember what is and isn't a TLD, so we have to assume that anything could be a TLD.
I don't realy agree with the proliferation of so many TLDs, but the damage is already done and the logical step forward is to let anytjing be a TLD.
https://webmasters.stackexchange.com/questions/24460/how-can...
Yes. This has been the correct way to handle it for at least a couple of decades and perhaps arguably more. RFC 2606 reserves a handful of TLDs (such as .example and .invalid) and it would be sane to treat these specially (e.g. this-is.invalid just isn't a valid DNS name) but otherwise the most you could say for sure about any TLD is that you don't know whether it exists.
There's a lot of bad legacy code out there which will do things like reject email addresses with a TLD longer than three characters, don't do that even if you think .horse was a bad idea.
> At this point, why not just accept any string of characters up to a certain lenght as a TLD?
With the exceptions in RFC 2606 and the practical constraint that some TLDs are "poisoned" by decades of misuse (e.g. .corp) this is the situation today.
However the next question is - who shall operate the registry for each of these TLDs? So that's what the ICANN process is for. It's not a problem for .some-ludicrous-vanity-domain-for-a-corporation as on the one hand, nobody else wanted those names (often even the corporation itself never uses them, it's just burning money) - but it gets trickier for .horse where understandably the people who own a horse's name in .horse don't want to hear about how the company operating it went bankrupt and now they need to pay somebody else or their site stops working.
I would call it poisoned by decades of "misuse". From my POV, squatters rights should apply and ICANN has no business delegating widely used TLDs, even if they are not formally registered (in which case they would be expected to pay $$$$-$$$$$$ / y).
TLDs like .onion, .corp, .lan, and other widely used TLDs (say .bit, .dn42, ...) should not be delegated by ICANN, unless it is to the undisputable current users of the TLD.
There is no way a hobby project could pay $$$$$$ for registration, and if it does become successful, while not using the root nameserver services, I don't see why it should pay for the privilege.
It makes no sense to do this for say .corp because in fact the existing "squatter" users just arbitrarily assign whatever names they want, so unlike facebookcorewwwi.onion there is no agreed global meaning of headquarters.corp or hplaserjet.corp - the namespace is just poisoned.
No, it is not globally resolvable, but it is in the same namespace (no way to add a different namespace while retaining compatibility), is widely used and I don't see why a dozen or so widely used TLDs need to be auctioned off to Donuts co. or whoever. It's not like we have a shortage of words or even operational gTLDs these days.
I'm not suggesting any TLD with 3 users should be off-limits (though I don't agree with gTLDs altogether) just the really widely established ones. If a project/group/whatever successfully squats a TLD and gets widely established, then it should be protected as well. If nothing else, then by squatting itself.
We do not have another namespace for alternative resolution strategies, so ICANN should not simply rent away all names.
Last time I heard, ICANN was corrupt. As far as I am concerned, they have no more right to assign .corp or whatever, than I have a right to reassign and use it for other purposes on my systems. I am not paying 185000$ for gTLD, nor renting a long domain to use for internal purposes, or with my friends.
Who's the "they" in this sentence? ICANN delegating the TLD formally is the mechanism by which you pay money to delegate a TLD. Until said TLD is delegated, it's not a real TLD, and does not resolve on the Internet. You can't use it in a meaningful sense until it is a real TLD, and once it is a real TLD, global collision rules apply. The fake use cases and the real use cases are fundamentally incompatible.
"they" are TLDs
I can use tor (.onion), namecoin (.bit), dn42 test net (.dn42) or others. They work and people use them, so not fake, just not registered with ICANN. Most of them work fine along with ICANN DNS. They require alternative forms of DNS resolution, so they don't work with ICANN DNS, but that is actually desirable for many of them.
I can also have .lan addresses on my router, .corp addresses on corporate vpn, or whatever in my hosts file and they work. Hosts file works even better than global DNS, so the domains are definitely real, not fake. Just not global, even when globally used.
Which is demonstrably not the case:
(If you can't guess what that leads to, you might benefit from two pieces of context. The European broadcast region has an organisation for its Public Service Broadcasters, Eurovision, which holds an annual Song Contest that is known for quirky up-beat winning songs. There was a Channel 4 comedy show about the lives of Catholic priests living on a small Irish island and in one episode they enter the song contest...)
While some extensions are just absurd, I think there is a good branding potential with choosing some of these TLDs (like for crisp.chat, frame.work or magic.link).
I’ve heard of decentralized protocols like Handshake attempt to decentralize the entire sector of TLD’s but they just seem be riddled with early profit seekers buying up all the names, so it seems unlikely the large browsers will incorporate them. Can anyone speak on the TLD space and it’s future? What was the incentive of ICANN starting the gTLD’s and what are the assumptions behind browsers accepting TLD’s?
Yes, and any system will be riddled with early profit seekers. With Handshake, there's a lot less profit to seek:
1. The current TLDs are pre-reserved on Handshake for the current owners.
2. The number of Handshake TLDs is ~infinite. If it were to catch on, I'd expect a race-to-the-bottom for second level domain prices.
I recall talking with someone who worked with a trade school, this was probably about 15 years ago, and they were very proud of their .edu domain. It was seen as a bdge that they were a federally-sanctioned college as opposed to just a Learning Annex with delusions of grandeur.
At the time, they could only get the one domain schoolname.edu and couldn't just buy "enrollatschoolname.edu" or "schoolnamegoldenhamsters.edu" for a marketing push. Not sure if that's changed.
I created auctions for the remainders.
Granted, unless your devices are using a Handshake resolver (or NextDNS[3]), you won't be able to actually use these domains but browser adoption seems (tentatively) likely. Puma Browser[4] (mobile app) supports Handshake so who knows...these dead TLDs may see life again.
--
[0]: https://namebase.io
[2]: https://nextdns.io
http://web.archive.org/web/20210226094451/https://dzdb.caida...
At any rate, it's a way for brand-owning companies, or squatters hoping to shake down those companies, to help support ICANN financially. TLDs are expensive to register.
What's mostly going on here is simply that companies bought something without a plan for using it and then dropped it after a period of time as the expenses continued adding up and they continued doing nothing with the TLD.
Any given potential name can be the subject of multiple (registered, not to mention unregistered) trademarks in different combination of field of business and territorial jurisdiction, so even if you have high trust that ICANN won't issue something that is a trademark somewhere in the world to anyone other than the trademark owner (which you probably shouldn't be, given that it's pretty much impossible), you have very little reason to be sure that (1) your brand isn't also someone else's trademark in some jurisdiction and field of business, and (2) that person won't get a TLD for it. Especially in the early round of newTLDs when it wasn't clear the degree to which the prediction that having your own TLD would be expectation as much as having a second-level domain in the most appropriate TLD had been previously would turn out to be true, preemptively acquiring the TLD even without a clear plan to use it made a lot of sense if the cost of doing so is a small fraction of the value of the brand.
I was doing registry work at the time the new gtld program was launching, and remember thinking how silly it all was but specifically some of the tlds being sought. Sure enough, I recognize a few of them listed.
So you'd go to `iphone.apple`, `zero.cocacola` or `order.macdonald` etc..
I guess that didn't workout yet.
I didn't know that companies could simply register TLDs.
- .chrysler
- .srt
- .uconnect
- .mopar
- .dodge
- .fiat
- .jeep
- .ram
So it does in fact look like that's exactly what was/is happening.
This is actually my project that I started years ago to track changes in root zones files. Let me know if you have any questions.
Keen to use a .land for something but would suck if it disappeared.
.land specifically seems to be owned by Donuts Inc, so that will stay [0]
Tl;dr they've thought about how bad for the Internet it would be if TLDs with many users on them could just fail and disappear, so they've taken steps to prevent that from ever happening. There's a reason almost everything on this list is a brand TLD. Your biggest risk as a consumer isn't that your TLD will disappear, it's that the price will be jacked up on it (though this has been happening on .com and .org too).
Thank you. That provides significant comfort. Want to use it for email so possibility of them pulling the plug would be not awesome
Glad that does directly apply as a possibility
Price increase - well I guess if it’s not absolutely insane increases I’ve got at least a year to transition
When will .exe, .jpg or .png ever also be in TLD?
It's just a programmers solution to a programmers problem.
The only reason .com even matters is because it got ingrained in people's minds early on.
You could drop all TLDs tomorrow and nobody would care.
Huh? That would cause all domain names to stop working. People would certainly care! Can you expand on what you mean by "drop all TLDs" and how all of the existing Web content and links therein would continue to work if this were done?
The high cost of the custom TLDs really keep them from being useful for internal systems though.
Internal TLDs are "free" in the sense that they are not registed with IANA.
Luckily I had to edit my hosts file for a new client and eventually realised the problem was because of a project many years ago.
What I’m thinking is having a seperate domain, which many companies already have, for corporate services. We have a .dk for our actual service, products and email, but we also use a .net domain for things like AD, Azure integration and a ton of auxilary services.
Having .company rather than .company.net is just a “nice to have” and certainly not worth the $185,000 registration fee. Had it been cheaper I could see many chooce to have their own TLD just to have a more clear border between “infrastructure” and “product”. It gives the IT department more control of the domains they wish to allocate without stealing the from the product and marketing teams. I worked for a company where we already used a series of subdomains, for backend stuff, but down the road those same subdomains became more and more relevant to free up for the public parts of our website, and it was a pain to move around.