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

I’m not massively against that, but I do wonder whether giving no ability to choose a free name, for fear of some people getting the ‘best’ ones, is preferable. Why spoil it for everyone, just because some may grab the obvious ones?

I would much rather have a slightly long phrase than a completely random string of digits as my base address. SillyDancingBrownGerbils is highly unlikely to be chosen, but it may be memorable. At5hy08jhqifDJI82aDFppLiO is utterly forgetful. It reminds me of xkcd: Password Strength

In short, with no ability to pick a name, then you absolutely need a way to resolve to it. Being able to pick a name - no matter how whacky - potentially avoids this. If you want a simpler name, say Gerbils, then it could be paid for.

IMO, all approaches have value. It is good to have memorable free native addresses. It is good to overlay shorter paid for names. It is good to have great, distributed search engines. Each brings something to the party.

or we could use a date blockchain to register a name per day. example: safe:www.google could be registered on 05.09.2015, and also on 06.09.2015 and so on. this a name can be reused daily, so the name itself will have no value. and when i tell a person to access a site, i will tel him to access safe:www.google from 05.09.2015
if a domain name is already register on a specific date, the user is told that he could try to register the domain name another day. also, a delay between registering names is welcome, this way automatic attempts could be prevented. or any other measures to make automatic process less successful

I could be wrong, but I do not think the SAFE Network has the ability to tell time.

then they have to build an internal clock, which i think it could be useful for other things

I think @dirvine has a good reason for not having a clock.

2 Likes

well i think it doesn’t have to be a clock like outsideworld because of the time-differences between user geolocations. but it could be a brand new network clock that starts the moment the network goes live. and all the users of the network will have the same internal time. but maybe technology does not allow to do this

You get one of two outcomes:

  • the richest guy wins (outcome that you supposedly want to avoid)
  • no one wants the domain because there is no clear winner and it’d be silly to invest in it knot knowing you’ll be able to keep it without the constant spending to outspend all the other contestants

What spending? Domain names would be free.

I just need to spend money on ads that would make people go to my address.

(A funny issue with your idea is that the network can’t know which of the three sites the user actually wants to visit. I see you have a “solution” which is to ask the user, but of course how could they know which of the three hard-to-remember keys belongs to the site they are trying to reach?
So it would make sense to send them directly to my complicated address (in order to not was my ad budget), but then the alias wouldn’t even be triggered and you couldn’t really tell that my site is “popular”)

This would be true no matter what system is used.

There would be no hard-to-remember keys. The user would only see the number of views. The user would choose which site to go to, or let the browser choose for them based on number of views.

Example:

Input = SAFE:www.site

Output =

SAFE:www.site # of views 1,000,000
SAFE:www.site # of views 1,000
SAFE:www.site # of views 100

Each domain name could have meta data added to it.

Example:

SAFE:WWW.site
Rank = #1
Number of views = 1,000,00
Name: John Smith
Company: Smith and Company
Adress: 107 Smith Lane
Smith, NY

No thank you, to this^

1 Like

All meta data would be optional to input.

This is gameable as hell, just like there are companies from which you can buy Facebook “likes”, we’ll have companies that sell fake views to get you a domain. This problem is essentially the same as the consensus problem that Bitcoin solved with Proof of Work and SAFE will solve with Proof of Resource. There has to be a cost involved, or else the system will be easy to game. This is also why there is criticism on Proof of Stake, there is no real cost there if you look closely.

2 Likes

That is literally Zooko’s Triangle.

  1. “Native addresses” = Securely Unique
  2. “Shorter…names” = Memorable
  3. “Distributed Search Engines” = Global

Check out how implementing Petname Systems addresses those three.

P.S. “Memorable free native addresses” doesn’t work without suffering real harm from squatters. See Namecoin’s blunders.There have been systems that try to combine one or more of those elements when designing a system. Spoiler alert - they weren’t able to accomplish them.

3 Likes

The cost could be that each user’s view would only be counted once. So if you wanted to pay people to view your site it would get expensive quick.

Example
Joe views SAFE:www.site 1,000 times, Joe’s site views would only count as 1 view to the sites ranking.

1 Like

Joe can make virtually infinite accounts. Tracking views per user is also both a great strain on the network and a privacy risk.

3 Likes

This is another major issue that needs to be addressed.

I believe the dev team is already working on this issue.