Rethinking the registration/renewal model for ENS

💬 General Discussion
44

The conversation about “what is a squatter” is holding us (and ENS) back.
The competitors are not having this conversation, and
for ENS to progress, we must all resolve.

We need to come to an agreement:

  1. The “reoccurring fee” is a great feature. The “reoccurring fee” uses strong game theory, to keep names in circulation, (over the many decades), from lost wallet keys, etc.
  2. A “Squatter” is NOT the same as a “Flipper”; although a user could be both.
  3. Just because someone buys a name, (to later resell), they are NOT a squatter.
  4. A “Squatter” is a person who unlawfully occupies an ENS name (like “not your company’s name”, ie: google.eth, winklevoss.eth, mycrypto.eth).
  5. A “Flipper” is a person who buys ENS names (usually many), with the intention to immediately sell for them for “more than the spot cost” of the NFT-asset.
  6. A “Squatter” is bad for the system (because the action may be unlawful, and is unethical), but we can not stop squatters (since ENS is ~100% decentralized, immutable, and censorship resistant).
  7. In a properly designed system, a “Flipper” is NOT inherently bad for the system, but the behavior SHOULD be discouraged (see agreement #1).
  8. A “Flipper” helps with price discover, the help drive ENS income for promotion & development, and they own this right, as decentralized users & by paying ENS fees.
  9. ENS exists to provide the most decentralized naming service in the world. This means we want ENS to be available, easy to use, and useful, for any end user to register & manage ENS names. The goals of “users who buy ENS-names to sell” are NOT directly opposed to this.
  10. The base costs of ENS should “never go up” (ONLY DOWN). (We can reserve the right to review, after ETH "Crashes, all the way DOWN to $10,000; (and a loaf of bread costs ~$10/20)).

Fight me? Fight us.

We are one ENS.
All for one, and one for all.
All users are friends, or all users are enemies.
Strong Together–One Message; One Naming Service.

2 Likes
45

I like the concept of this proposal; it’s an improvement.
To say it again: I see the value in the “reoccurring fees”.
Just plz don’t increase the “one year” fees for users, imho.

I think Makoto has a good pricing table example,
…which keeps the “one year” fees consistent.

What are the next steps, or what specific questions do we need to hash out?

1 Like
46

I do use some of my domains actively. I trust no third parties so I rather use ens than checking every characters of an address twice. One pointing to a multisig wallet, one redirected to deposits at ftx one to binance etc. A few others I reserved for an ipfs mirror of my websites and some which I want to giveaway for free registration of subdomains. I’d hate to see other users fleeing this contract though.

47

The point of the temporary premium is to stop expiring names being snapped up by whoever pays the highest gas price and has a bot set up to grab them as soon as the name becomes available. Instead, it ensures anyone can get the name when it reaches a price they consider fair - or if nobody wants it, it remains unregistered and the price returns to normal.

Makoto has enhanced our bot so that it notifies users who have ENS names in their Twitter names when their names are expiring, and we see a significant number of those expiring. I don’t think it’s an unreasonable position to think that people accidentally allow their names to expire, and that this is a significant issue for the usability and safety of ENS names.

That is because most of our competitors are optimising for maximum revenue - which means encouraging squatting, as squatters spend a lot of money speculating on domains. ENS’s first priority is to be useful as a naming service, not to sell as many names as possible.

This isn’t about labelling users, it’s about behaviour we want to encourage and discourage. Names being bought and resold at premiums makes ENS less useful, so that’s behaviour we want to discourage.

I don’t think it’s really useful to distinguish the two here. Both impact usability and availability of names negatively, and both operate the same way with the same goals.

While it’s technically true that buying and reselling valuable names “helps with price discovery”, this “service” helps only the squatter themselves; everyone else is impacted negatively by this behaviour.

Again, can we stop focusing on particular potential pricing schemes for one potential proposal, and instead focus on the overall goals, and if there are adjustments that could be made to achieve them?

For example, do people generally agree with my argument that substantially discounting longer registrations will serve to make them more attractive to users, without reducing availability by making squatting more affordable?

48

I follow the bot. I do think it is an unreasonable position to think that people accidentally allow their names to expire or in any case I do not believe it is a significant amount of all expired domains nor a reason to change the current fee structure. Do you have any reason for your suspicion besides this hunch as most domains I’ve seen that expired come from accounts that have renewed other domains.

No. The registration fee is a very small portion of the total price to register a domain at the moment (~10%). At 105 wei rn registering a new domain for a year costs Ξ0.003 and gas fees are estimated at Ξ0.032. So I do not believe adjusting the registration fee will attract more users. I guess if you believe that gas fees will stay at prices they are at now for longer than the average domain registration length something could be said for that, but if it stays like this for longer than a year everything is useless imo.

One year free domain for new users prob better solution to attract more users

49

Yes–As long as “one year, reoccurring fees” do-not increase.

50

Even though I am the one who initially proposed the discount (in the team discussion), I am getting less and less supportive of this idea for the following reasons.

  • As @Dreko mentioned, simply giving a discount does not have much impact considering the current gas price is a lot higher than the annual fee. Even if we do “3 years for half price” rather than a gradual discount, it’s still from $15 to $7.5 which is not a lot.
  • That’s why your initial example increased the first year to make the discount look more effective but that would equally discourage users who are new to the ecosystem.

Thoughts on the squatter. It has been puzzling to me that the current high gas price hasn’t decreased the number of ENS registration (our daily average is between $100~$200). From the current price point of $5/y it is irrational that people spend $100 on these names unless they are buying ENS names in the expectation that their name price goes higher in value in the long term. I would assume that we don’t have non-squatters in the purist form @nick.eth defines and have to start thinking about living with it. This doesn’t mean we should encourage squatting but we should at least relax our attitude that squatters add no values. As they have skin in the game, they will be more passionate about promoting the value of ENS whereas casual users (non-squatters) are less likely to join our community and governance discussion simply because they have no reason to care.

The single most effective way to make the system less liquid is actually to revoke the NFT ability from .eth names, but everybody should agree that that’s a bad idea.

1 Like
51

I agree with everything you are saying here;
with caveat to the fact that (even with or without the high gas fees),
we should not underestimate the economics how the long-term-discount does help some/many people.

52

You can’t buy a three letter .com for under 10k USD. So where do you get that 1/75 of cost from?

You are right that you in theory could register it for 10 USD, same price as any other .com domain. But that just shows the problem. All three, four and five letter .com domains are gone, almost all 6 letter as well. A large number is owned buy domainers. So “normal” users don’t have access to these domains.

1 Like
53

Hi,

I thought I might give a short input from the domain market.
I work for the ccTLD .se, but everything here is my own, not my employers.

Nick, you are right, domainers/squatters are considered a problem in the TLD space by many. I personally have never understood why we as an industry try so hard to hate our best customers.

One thing is sure, human readable names have a value, and shorter or more recognizable names have higher value. No policy will ever change that. Every market has people speculating, in some more open than others, but it exists everywhere. If there is money to be made, people will find a way.

So instead of focusing on squatters, focusing on providing more value to users will be a better investment of time.

From a price perspective, even 20 USD would not be a problem if people see a value in it. Think of yourself, how many of your .com domains wouldn’t you renew for 20 USD. You might find it unreasonable, but would the price or the “unreasonable” be the problem. 100USD or 50USD would make you think, but not 20.
As reference, .se has increased prices twice in the last years and there is no decline in registration whatsoever. But of course. .se has a very unique selling point as being the TLD for Sweden.

And that would be my input, ENS is not in a position where people see it as the default name service of the blockchain world. Until then, I would ignore the squatters and try to maximize user value.

Just my 2c,

/Ulrich

2 Likes
54

I think a rethink is due as the current registration/renewal process makes unintended loss of names too likely, and also makes it impossible to guarantee permanent registration of a name, which is important for censorship resistance. I shared my thoughts on the latter point on Reddit:

Following the discussion I had with Nick, my latest suggestion would be to offer two types of registration:

default temporary: the current registration process
default permanent: deposit based registration, equal to $100 worth of ETH, with an additonal fee of $5 worth of ETH.

By giving options, ENS can allow registrants to use the option that suits them better. For example, Tornado Cash could use the default permanent registration method to link eth-1.tornadocash.eth to the 0x47CE0C6eD5B0Ce3d3A51fdb1C52DC66a7c3c2936 contract address.

EDIT: another potentially promising channel to explore is switching to an exclusively deposit-based registration process, but making the minimum deposit $100 of ETH, plus a $5 in ETH registration fee, to discourage speculators from sniping and flipping.

Perhaps the ETH in the ENS contract could also be lent out via an ultra-low-risk lending protocol, with the revenue paid out to the ENS multisig.

55

I’m naturally against permanent registration procedures because too many wallet holders tend to lose access over the years. That being said a permanent option could mean a waste of names worse than what squatters already cause imho.

56

It also has several drawbacks, the most notable of which is the risk that people forget to renew their domains.

Truthfully I don’t think this is a huge risk. That’s how the current domain name system works, and I don’t see why a decentralized domain name system should be any different.

Traditional DNS registrars allow auto-renewals, which ultimately depends on having a credit card tied to the account. Instead what might be possible for ENS is to allow “fixed rate renewals”.

Lets say you purchase a domain for 2 years at the price of 0.1ETH/year. You could deposit 0.2ETH into a “renewals contract” which could be incentivized via cron-job style keeper bots, or some sort of automated system. This would allow users to effectively “set and forget” the renewals of their domains.

Another alternative would be to have the ability to also delegate renewals of domains to some sort of “renewal service” that would allow you, or the renewal service to handle renewals of your domain for you.

57

I just gave one - the bot is frequently tweeting 7 day and 24 hour reminders to Twitter users who have ENS names in their handles, but are about to let them expire. I think that’s a fairly strong indicator they didn’t intend them to expire.

That’s not what I’m saying. The goal is not to attract more users, but to incentivise users who wish to retain their names to register them for long periods, without causing squatters to also register for long periods.

It’s rational for someone to register a name if they think it’s worth $100 to them. That could be because they’re speculating on reselling them for more - but it could equally be because they’re expecting to use them and get at least $100 worth of value out of them.

Thank you for your feedback! I think this is a bit of a false-dichotomy, though. Trying to optimise the pricing for end users and making the system more useful for everyone are goals we can pursue in parallel.

Thank you for joining the discussion! Your post was a large part of what started this ball rolling.

One issue with offering multiple options is that users will pick whichever option suits them best - meaning you can only deter squatters as well as the friendliest-to-them option deters them.

A practical issue with the deposit-based scheme is the way the permanent registrar is structured; without replacing it entirely (a significant migration), we’d either have to require users to ‘ping’ their name regularly - which reinstates the possibility of losing it - or rely on the controller to prevent names being replaced - which is less trust-minimised than the current solution, as keyholders can replace the controller fairly easily.

We also have the issue that in the past, squatters have been happy to deposit funds, knowing they can get them back when they flip the domain. The situation today is somewhat different to how it was then, though, with many competing uses for capital onchain.

Depoist-based schemes help avoid people leaving a name unused indefinitely, too, but don’t guard against loss of keys, which can result in names becoming permanently unavailable to everyone.

The primary reason is that we have far fewer options to remind people their names are expiring, which is the main thing that triggers people to renew their DNS domains. Unless a user opts in to reminders at the time they register - and we know many do not - they are liable to forget to renew them when the time comes.

We already support this by allowing renewals and registrations of any duration - but that just puts off the day when the user needs to remember, unless they’re prepared to pre-register for their entire lifetime!

58

The primary reason is that we have far fewer options to remind people their names are expiring, which is the main thing that triggers people to renew their DNS domains. Unless a user opts in to reminders at the time they register - and we know many do not - they are liable to forget to renew them when the time comes.

Thats a fair point, but to be honest I think that’s the users fault and not something ENS should worry about. ENS already provides the notification option. Users could set a date in their calendar, set it in their notes, etc… At some point the user cant be absolved of their responsibility in having to renew their domains.

We already support this by allowing renewals and registrations of any duration - but that just puts off the day when the user needs to remember, unless they’re prepared to pre-register for their entire lifetime!

Correct me if I’m wrong but currently there is no “auto renewal”, a user can elect to renew their domain at any time but they still need to go through the process, what I was thinking is an auto-renewal that will renew the domain for them when the expiration date is within X days. So the user never needs to renew themselves, they just need to top up a balance somewhere and have some sort of automated service ala keeper bots or something to trigger the renewals for them.

59

In your stated goals, stop placing your intents to “hurt” flippers/squatters, which is splitting your focus and minds attention, while weakening the wider ENS community.
Yes, it is good to deter flippers/squatters–We do not want all names taken by them (and held like NFT-art), but domain flippers are part of this community, too.
Please, just focus your attention on making the ENS system its very best for all users.
Please build for everyone, or the competitors will be able to burn ENS on the issues.

60

It’s certainly the user’s responsibility to remember to renew. But if we see a lot of users forgetting, that’s a fairly clear sign that the UX needs improvement. We really want to avoid anyone losing their name by accident if we can, even if it’s “their fault”.

We don’t have that - but prepaying for x years seems equivalent, and avoids extra transactions. If you have an account that needs topping up, then you have to remember to do that, too!

Our goal is not to “hurt” anyone; it’s to disincentivise behaviour that hurts the ecosystem. Every name registered by a squatter exerts a deadweight loss on other users.

1 Like
61

One issue with offering multiple options is that users will pick whichever option suits them best - meaning you can only deter squatters as well as the friendliest-to-them option deters them.

A practical issue with the deposit-based scheme is the way the permanent registrar is structured; without replacing it entirely (a significant migration), we’d either have to require users to ‘ping’ their name regularly - which reinstates the possibility of losing it - or rely on the controller to prevent names being replaced - which is less trust-minimised than the current solution, as keyholders can replace the controller fairly easily.

We also have the issue that in the past, squatters have been happy to deposit funds, knowing they can get them back when they flip the domain. The situation today is somewhat different to how it was then, though, with many competing uses for capital onchain.

Depoist-based schemes help avoid people leaving a name unused indefinitely, too, but don’t guard against loss of keys, which can result in names becoming permanently unavailable to everyone.

Thank you for the feedback @nick.eth.

Regarding the point about deposit-based registration making it possible that names become permanently unavailable to everyone, I hadn’t considered that risk…

I am just trying to find a way where ENS can help in the creation of an immutable key-value pair consisting of a name and the address of an immutable Ethereum contract. The textbook example would be the Tornado Cash ENS name linked to the address of the unalterable Tornado Cash Ethereum contract. Right now the ENS name could be changed, by both the current registrant, and a successor registrant should the current registrant fail to renew the name.

I wonder if it would be possible/desireable to provide a method for explicitly making a particular name-resolver mapping immutable? If it were deposit based, the registrant would lose access to that deposit forever, so from their perspective, the deposit would be burned. I imagine such an option would be seldom used, but useful for certain applications like the above Tornado Cash one.

In any case, this doesn’t address the practical implementation issues you brought up with respect to instituting deposit-based registration.

1 Like
62

This is definitely doable now - just transfer ownership to an inaccessible address. The same would work under a deposit model.

1 Like
63

Yes, but having a general deposit-based registration method would pose the problem you cited, of people losing keys, and thus names becoming unintentionally unavailable to every one for ever.

Restricting the use of deposit-based registration to names with immutable resolver values would ensure that the only names that ‘become unavailable to every one forever’ were intended to be that way.

1 Like