RFC - Decentralised Naming System II - continuous auction (by Seneca)

[polpolrene]

@riddim

it could just register another url and publish this new url so its 100k followers use the new one instead Oo

Do you think that would be a handy situation?

[riddim]

that would be annoying but i don’t see a problem - it also annoys me that I’m not the ruler of the world … sometimes you have to find a way to get along with the circumstances …

[catbert]

###Trust levels in delegated system

• a user recieves a pre-configured tree of namespaces when registering with MadeSafe
• each naming service has got a pre-configured trust level in his settings
• each mapping provided by a naming service also includes a trust level
• SAFE browser computes resulting trust level and displays that as address bar color

###Trust levels info displayed by SAFE browser

• each naming service can supply additional info for each name mapping it provides:
  ** company name
  ** jurisdiction
  ** street address

• when a user hovers the address bar he can see trust chain info; example:
  ** official namespace is curated by SAFE foundation (registered..)
  confidence level in this information is 95%
  this namesapce has a record that
  ** sub-namespace ABC is run by company XYZ (registered..)
  confidence level in this information is 75%
  this namespace has a record that
  ** website def is run by company PQR (registred..)
  confidence level in this information is 50%

This may create a lot of drive for companies to register in the official namespace - to obtain a high trust level. The level of trust assigned by Foundation’s proxy company to a name mapping will depend on the documents presented. Similar to ICANN business.

• adavanced users would mount new namespaces and assign trust levels at will
• a user would be able to present a Q-code to the camera and assign a trust level for the resulting local bookmark (they will need their own place in the namespace)

[Seneca]

Good example and I agree, the fact that you need to have wealth to make sure you keep your desired domain name is a drawback in these cases. It’s a bit of a double edged sword.

@polpolrene

Do you think that would be a handy situation?

Well, his registration under that domain isn’t gone, it’s just not the top ranked anymore. It can still be reached through that domain name, but users need to take that extra step to check out the registrations list under that name rather than defaulting to the top one. Browsers can be conditioned to show that registration list automatically under particular circumstances. Bookmarks will keep the internal address as well.

[riddim]

hmmmm - and how do honest drug lords (or non-official non-profit-organizations in difficult countries) ensure their trust level is high …? maybe suddenly a cruel official emperor/security agency captures the site and tricks whistle-blowers into revealing their identities …

your are creating centralized directories with not anonymous information here - that isn’t exactly the aim of MS as I got it till now

[Artiscience]

A lot of interesting discussions here. Just want to add some ideas.

The main “threat” that some people see is that first movers register so many domains that for the rest only random letters and numbers remain. As far as I see this is actually rather a routing problem.

1. User may register a Domain Naming System by burning a relatively high, dynamically calculated ammount of Safecoin (i.e. the ammount of Safecoin gathered by sharing the network avarage ammount of GB over a certain period, which could be i.e. ~1000 $). The user who creates the Naming System doesn´t own it, but can decide how much the registry of a domain name costs. He*she won´t receive any of the fees users have to pay when a name is registered.

2. Users may generate a (random) native address that looks like duskgytldkxiuqc6.onion by burning a relatively low, dynamically calculated ammount of Safecoin (i.e. the ammount of Safecoin gathered by sharing the network avarage ammount of GB over a certain period, which could be ~ 1$).

3. They may then register this address to a DNS of their choice at the pricing of the correspondent DNS.

4. When starting a browser, users may chose between all existing DNS. The plugin can offer them with popularity stats, a search function and by default choses the most popular DNS.

The most likely consequence is that after some time there will be one very large Domain Name System and several smaller ones. This allows the necessary centralization to keep users in one system, but also enough flexibility to change systems. Imagine one system is flooded with registries and a large part of the network is unhappy with that, they can always collectively move to a different place.

[catbert]

@riddim

hmmmm - and how do honest drug lords (or non-official non-profit-organizations in difficult countries) ensure their trust level is high …? maybe suddenly a cruel official emperor/security agency captures the site and tricks whistle-blowers into revealing their identities …

Good question. Under the delegated naming system…

#####Case 1.

There is one surviving member of the website crew who has access to the private key.

My proposed system is 2-tier. On the lower tier public key hashes are mapped to routing info. This information is stored by the core SAFE network in a distributed hash table. This info has to be signed by the private key to be valid. Maybe we could allow any surviving member of the website team who still has access to the private key to publish a revocation message saying that this binary address is no longer mapped to any website? Just an idea. Not sure who would pay for keeping this info afloat.

This member can also go to all registration services he knows asking them to remove the name mapping or drive down the trust level.

####Case 2.

Nobody with access to the private key is alive. General public notices that things have gone wrong and notify naming services. Those naming services which have the ability to do that (curated) remove the mapping.

Please note that under the delegated system at least the curated naming services can remove the mapping in Case 2. Under the althernative proposals (first-come-first-served, auction) there may be nobody able to remove the mapping at all

@riddim

your are creating centralized directories with not anonymous information here - that isn’t exactly the aim of MS as I got it till now

Hmm… Don’t think this is exactly right:

• curated services can include anonymous mappings - no copmany names on them of course
• a curated service can include an anonymous service as a sub-domain
• under the proposal each user is free to mount additional or alternative naming services and these new services can be as anonymous as they wish
• the preconfigured set of naming services each new user receives at registration time can well include one or several totally ones - depending on what the Foundation thinks
• if we don’t want Foundation to decide smth this important, then each user can be asked to choose several from say 10 or 20 most popular naming services at registration time - and anonymous services will likely be part of the offer

So what the proposal is trying to do is to create is freedom and choice not to remove anonymity.

[catbert]

@Artiscience

4. When starting a browser, users may chose between all existing DNS. The plugin can offer them with popularity stats, a search function and by default choses the most popular DNS.

Fabulous! By why wouldn’t we allow in addition to what you proposed curated collections into the mix? And why don’t we allow the user to mount one naming service as .a.safe and another as .b.safe so that he can use both at the same time?

@Artiscience

The most likely consequence is that after some time there will be one very large Domain Name System and several smaller ones. This allows the necessary centralization to keep users in one system, but also enough flexibility to change systems. Imagine one system is flooded with registries and a large part of the network is unhappy with that, they can always collectively move to a different place.

This is exactly what I want.

[happybeing]

@Seneca

There’s no perfect solution here. I think squatting should be prevented above all, it would ruin SAFE’s DNS on day one.

I don’t think squatting is that bad. None of us like the idea of it, but at least anyone can afford to register a name and know it is there as long as they choose, for a minimal fee. Your scheme limits this based on wealth, which seems against SAFE to me: Secure Access For Everyone.

Also, I can start to see some rather unwanted behaviour arising out of the ability of anyone with enough money being able to disrupt an existing website.

I think this is important for everyone, not just big money earners. If I’ve put effort into a site that doesn’t generate money, but gets lots of traffic, I’m constantly going to worry if I’ve paid enough to keep ownership. Even if the site makes money and I can afford to pay a bit more, I could lose valuable or important traffic (and my users lose my service) if I don’t pay “enough” with no way to know what “enough” is.

To me this seems worse than the impact of domain squatting, which despite having created many websites hardly affects me. It means highly funded ventures who want a particular name have to pay a few thousand dollars for it, and those with less have to be more creative.

I don’t agree that “[domain] squatting should be prevented above all”. You have stated this but not made the case for it.

The inclusion of warning popups makes it worse IMO. Note: if losing your domain was not terrible, there would be no need to pay big bucks to keep it. But popups don’t solve this IMO.

Anything that delays a user getting to the content they seek loses traffic. This is a fact, and in business it matters, but also for the usability of the entire system. A pop-up which requires the user to make a decision, which to me looks like a pretty hard one for most people, will not only lose traffic for the website concerned, I think it would also risk turning people away from SAFE Network altogether.

To me either of these features (the risk of losing my domain and the need for popups where ownership has changed) is worse than the impact of domain squatting. 99.9% of users won’t notice the difference between a web with, and a web without, domain squatting, so we need to keep them in mind, and ensure everything we do improves things for Everyone. Or as many as possible!

I think we can make a system that is clearly better, and in no way worse, than the current one, and so can only encourage adoption.

On the current web, yes, most of the obvious names are gone and only affordable by business. But you can still catch them when they expire… I got my full name for about $100 - I learned about it because a broker offered to get it for me. I ignored his offer an bid myself

And even now, anyone with an imagination can still get a cool name like theWebalyst.com I even nabbed online-counselling.co.uk (or very similar only four years ago - I let it go).

However, with SAFE, choosing good unused domains will be even easier because there are no Top Level Domains to restrict us. This makes squatting much less restrictive, and if we limit the number of registrations per account, and include a charge etc, we can improve on the situation compared to the old web by making it harder and more costly to squat.

I think we need to be very careful before messing with something that works and which people already understand. We should certainly not insert a warning pop-up between a user request for information and supplying it, unless there really is no alternative. Warning popups are what we should aim to eliminate above all IMO!

SAFE Network can have both a better name allocation system, and keep a familiar model that we know works, and has no quirks that might discourage adoption.

[Artiscience]

@catbert

Fabulous! By why wouldn’t we allow in addition to what you proposed curated collections into the mix? And why don’t we allow the user to mount one naming service as .a.safe and another as .b.safe so that he can use both at the same time?

Well, it´s not about “allowing”, is it? I mean, even if its not included in the plugin, no matter who can create a tool to mount addresses in a specific way and share them. The reason why I proposed it in the above way is that I believe there needs to be a threshold at system level because otherwise the “collections” turn into mere top level domains (maybe I got you wrong, in this case please elaborate). The proposal that I made would still allow a large corporation to create a DNS and then register all addresses selling them to users, but a) they´d need to pay for each and every address, so even if you´re very rich, you would soon run out of money and b) the money would be senselessly wasted, since as soon as people understand you are ripping them off, they will simply switch to a different service.

[Seneca]

@happybeing

But you can still catch them when they expire

But they don’t expire on SAFE, at least I haven’t seen a serious proposal that has that feature.

@happybeing

I don’t agree that “[domain] squatting should be prevented above all”. You have stated this but not made the case for it.

I think that squatting will be way worse on SAFE than it was/is on the current internet, because everyone now knows how profitable it can be plus they domain names don’t expire.

@happybeing

To me either of these features (the risk of losing my domain and the need for popups where ownership has changed) is worse than the impact of domain squatting. 99.9% of users won’t notice the difference between a web with, and a web without, domain squatting

This I dispute, I think squatting will be pervasive and very annoying to the average user. Practically all well-known domains on the current internet will lead to crap sites on SAFE.

@happybeing

Warning popups are what we should aim to eliminate above all IMO!

I agree that “popups” are bad, but I think that domain name take overs would be quite rare, especially high profile ones.

@happybeing

SAFE Network can have both a better name allocation system, and keep a familiar model that we know works, and has no quirks that might discourage adoption.

I have yet to see a proposal that delivers all of this. It seems we have different expectations of how things would turn out under the different models.

[catbert]

@Artiscience

no matter who can create a tool to mount addresses in a specific way and share them.

What I like about your proposal is that it allows multiple naming services. If we start with a system with only one naming service that assumption will feed into all software and will be impossible to change later on.

So I consider it is rather important that the initial implementation of SAFE plugin allows the user to choose from several naming services. This would ensure all software is written and tested in an environment where names can be resolved differently by different users.

And true I would love to see curated services supported in this fashion as early as possible.

@Artiscience

The reason why I proposed it in the above way is that I believe there needs to be a threshold at system level because otherwise the “collections” turn into mere top level domains

You are right - the collections do turn into domains! But…

• each user will have a choice of which services to mount as top-level domains and under what names
  is my .com resolved by naming service A or naming service B?
  do I want it it be .c instead of .com?
  well maybe in practice these will be .com.safe and .c.safe

• this is a hierarchical system - collections can turn into top-level domains or into sub-domains:
  imagine I mounted naming service 000ffaaaaeooo000.onion as a.safe
  imagine 000ffaaaaeooo000.onion has a record saying that
  another naming service 134ffaaaafffooaaa.onion is named b
  imagine I type name.b.a.safe
  then name is resolved by 134ffaaaafffooaaa.onion
  the computed trust level for name.a.b.safe then depends on trust level I have in
  000ffaaaaeooo000.onion, the trust level 000ffaaaaeooo000.onion has in
  134ffaaaafffooaaa.onion and trust level 134ffaaaafffooaaa.onion has in name

@Artiscience

The proposal that I made would still allow a large corporation to create a DNS and then register all addresses and offer them to their users, but a) they´d need to pay for each and every address, so even if you´re very rich, you would soon run out of money and b) the money would be senselessly wasted, since as soon as people understand you are ripping them off, they will simply switch to a different service.

I welcome any system to the mix - under the condition this is not the only system offered! And this vein I don’t see a reason why a big corporation shouldn’t be allowed to run its own registry. I believe it should be up to people if they want to mount these corporate services as top-level domains. I hope people will not want to do it. I think the Foundation or its proxy company will be very well positioned at the start to become the maintainer of the most popular curated collection. Then it will be up to people running it not to loose the race.

[tfa]

Things are even funnier than you think.

In @Seneca proposal the safecoins you have paid for a domain name have been recycled. Unless I am wrong, this means you don’t get them back if a higher bidder overtakes your domain name.

@janitor

You pay $20, I buy it for $21 and sell it back it to you for $22. Everyone’s happy?

You cannot sell it back for $22. If I want to get it back I only have to pay $2 to the network and you get nothing. In this poker game one of us will not be happy at the end.

@janitor

2. Alternatively, auction off an alias layer

@Seneca proposal is also an auction layer but a decentralized one and with a very peculiar rule: the loser loses the domain name and the money too.

Maybe @Seneca isn’t going so far but personally I like this system. And I would love to see huge battles between big corporations for a domain name because this system would double the amount of recycled safecoins (and in fact more than double when both competitors want the name desperately). This would be fantastic for the network.

[janitor]

@tfa

You cannot sell it back for $22. If I want to get it back I only have to pay $2 to the network and you get nothing. In this poker game one of us will not be happy at the end.

All right, so the money goes to the network. I just need to constantly outbid you in increments of $0.01 or whatever is the minimum that triggers the change.

You wouldn’t see huge battles between corporations, but between the small fish. Corporations would mostly settle in the real world outside of MaidSafe because they aren’t anonymous. Individuals would not.

[riddim]

@happybeing

I think this is important for everyone, not just big money earners. If I’ve put effort into a site that doesn’t generate money, but gets lots of traffic, I’m constantly going to worry if I’ve paid enough to keep ownership.

I don’t expect this being such a big deal … someone wants a site → he looks for free domain names cause he doesn’t want to rally with someone who is longer active and probably will not give up at first try

So he chooses an unoccupied name from the unlimited possibilities in the safe network … we aren’t trying to limit the TLD for maximizing profit … so if rid.dim is occupied i’d take riD.Dim or riddim.de riddim.home, riddim.at-work riddim.the-magician riddim.ruler-of-the-world, wise.riddim or vegan-cooking-with.riddim

[catbert]

@riddim

if rid.dim is occupied i’d take riD.Dim or riddim.de riddim.home, riddim.at-work riddim.the-magician riddim.ruler-of-the-world, wise.riddim or vegan-cooking-with.riddim

A lot of websites which

• used to be run by ordinary people or small companies
• used to contain useful info
• have gained a google rank
• haven’t had the lease on their names extended by owners

have now turned into link farms. Guess the “squatters” are trying to ride high google rank of now deceased websites. Something like this is a concern for SAFE - taking over the names of popular but not profitable websites for all kinds of gains.

[riddim]

@catbert

Something like this is a concern for SAFE - taking over the names of popular but not profitable websites for all kinds of gains.

if they are popular people would be willing to donate for it to persist … people are donating huge amounts for financing wikipedia-servers too (and they would safe this money when wikipedia moves inside the safe network)

and riding high google rank of now deceased websites is not in the interest of google - so google will try to prevent this kind of abuse and it can’t work too good (so prices obviously will be low) plus the risk inside safe is always that you loose all money and get nothing when you try this stuff here … i can’t imagine something like that turning out profitable

[smacz]

@Seneca

For every domain name, the network maintains a ranked list of such addresses.

It sounds like we’re going to be utilizing a large public decentralized ledger of some sort then, huh?

EDIT: Or can this take on the MaidSafe mantra of: “Each computer has a unique view of the network.”

[Traktion]

@janitor

I forgot to note: the client who’s visiting the site for the first time every won’t have any public key that is known.

You can have a chain of ownership, much like we discussed in the blockchain thread. Each new owner’s public key (or hash of it) would need to be added to the end of the chain. You would immediately see the previous owners, as well as the current owner.

This could be used to judge whether ownership had changed recently, especially if you could judge approximately the time of the event.

(side note - creating a hash of a public event could be used here. Maybe a hash of the last winning lottery numbers or some such. Easy to verify, but near impossible to guess. As we have no concept of network time, we would need something like this)

[Traktion]

@polpolrene

What about that little kid in India who created a brilliant video-site for SAFE. It registered safe:videotube and know has 100K users and makes some money. Next moment some big “Silk Road” site pays a lot of cash for that domain, and decides that before they link to the kids site, first they’ll post advertisement for their own service. So the kid is now screwed, it has advertisement on “his” website without anything he can do about it.

You put a grace/cool off period in, giving the kid a chance to upload a redirect to his vault address or his new name. He could also publish a message to the site prior to the switch over.

If the cool off period was a week, a month or some such, then there would be ample time to make adjustments. Even if the network could use an approximate time metric (based on some network metric that could be used - connections per second, safe coins farmed over a period, etc).