[RFC] Public Name System: Resolution and RDF

Unless @rob, you are suggesting we just call them “Names”?

Because again, I don’t think the aim here should be to find a name for the system, but first find a name for the thing that people will type into the address bar.

I know you, and a group of other computer scientists and engineers, will find this excruciating, but I’d argue that domain name is probably the best name we have, in that it is what the vast majority of people understand to be the thing that they type in the address bar.

Yes, we don’t that network domains in the engineering sense like that of the web, but most folk will have zero understanding of that, nor give two hoots about it. They just need to know what to call the thing they type in the address bar.

But I’d argue we do still have “domains” in the safe network, just that the meaning is slightly different, but for the end user functionally the same. I.e a domain of information over which I have control, and only I can publish too… so I give it a name. I used to have to pay GoDaddy $100 a year for doing that, and they were in control, and I used to worry about getting it stolen, but now I don’t, I’m in control of it. And I can make as many sub domains as I want.


I guess I missed that direction in the topic; anyway, I find ‘global’ just as weak as ‘public’ for describing the system. There are a lot of properties that are not relevant or unexpected. Most acronyms I’ve seen make it clear it’s about some kind of global system already.

Some single words I found that have some relevancy:

  • Alias
  • Handle
  • Label
  • Identifier
  • Name
  • Nickname
  • Record
  • Referral (?)
  • Scope
  • Tag

For now, I’d still be in favour for SAFE Domain Name System (SDNS) and referring to them like on the web (domain or domain name). No need to introduce a new term while having so many similarities to DNS from a user’s perspective. Users can even act as a registrar if they happen to capture popular domains. (Anyway, how the system should be used has been discussed many times before on the regular forum, I believe (especially squatting).)

Good lord I hope not


Because it is not “public” or “global”

It is a naming resolution for any class of name. Local, global, singular, public or private, or anonymous or publicised

Go back to what a general definition of what a domain is. And there is your answer of why domain is not really applicable. There is no domain be it quantum, magnetic, electrical, computer networking, female spider (biological), political, poetic, or anything.

It is a equally distributed naming system, not hierarchical or community which is what domains live in

Reading this the entity I’m leaning towards would be:

  • Safe Name or
  • Safe Domain or
  • Safe Domain Name which then will tend to be shortened to one or other of the above

I also think Safe ID could work (never heard of the TV series despite it sounding my kind of thing at that age - maybe it didn’t air in UK).

I’m not taken with ‘Global’. That’s an intent, but when people are adopting this I think it sounds less credible than ‘Safe’ or ‘Domain’ Name.

Then for the system we have things like NRS, SNRS, SDNS, SNS etc depending on the entity.

I accept Dug’s points that we should try to focus on a meaningful entity first and the term for the system second, and that ‘Public’ is not suitable.


(just dropping some thoughts)
It basically is a pointer within the network - so a safenet pointer - or a redirect to the data location…

How about

SND - safe names directory

Everyone would see in the first moment that it’s like DNS just for safe (and ‘safe name’ still works - just like with domain names ; )

(Ps: or how about ‘listed safe data’ :sweat_smile:)

Oh - ah =D my favorite for now:

TRN == Truly Readable Name (spoken: troon) )

I’d vote for something to too boring

Other option: Number as Name - NaN


naming of map - NOM

Holistic Names Services - HNS (because its names for everything safe)

No Nations Names - NNN

Eternal Names Service - ENS

Hidden Service Names - HSN

Decent safe names - DSN


There is no mechanism to auction names, so I’d expect a high degree of squatting.


Hmm - but one could implent it pretty easily…

Updating (versioning) of appendable data costs money…

I could store the link to a website in hash(name) - the second person wanting that name would use hash(name+1) and when calling the website the browser just calls the 10 first possibilities +takes the one with the highest version count on first visit or the last visited if the user has bookmarked the site in his account [actually then you obviously would resolve it with the bookmark]
(having a switch option to choose a different version if it’s wanted and showing a warning if the last revision happened ‘lately’ [whatever that means])

… If at some point still many names have been taken we can include hash(name +thenChosenOffset+x)

(not too well thought through - maybe just a mind fart I had… But in this second I like the idea :innocent:)

Ps: the offset number to the name could be factored into safe links and therefore always link to the same resource if wanted

Sounds like that would basically end up being a petname system

While I love the idea of a pet name system this would be globally available competing names chosen by the creator of the website and ranked by how much he is willing to pay for being the most prominent owner of it - not by the visitor - so it is no petname system :thinking: (except for the bookmarking part I guess)

I like NOM since it sounds like name in English, is name in French and isn’t far off in Italian, Portuguese, Dutch, German…
Or how about TAG? Means name or label, no abbreviation necessary.

1 Like

Ah, so to add a new version/reference you need to pay a higher amount than the last guy? Or it just records how much was paid by each person to find the highest ranked? Interesting idea. I guess that while similar to appendable data, it would need to be a new data type so that this bid data would need to be recorded. It could also have a special fee structure too to avoid spamming the list of bids.

Actually I was even thinking in simpler ways with standard appendable data.

An example - your name is drehb - so you want to have people to end up at your hompeage if they visit ‘safe://drehb’ - and you just reserve the name drehb therefore - since you are the first one that just costs you 3 puts or so …

then there is your evil twin with the same name - and he wants the same - so he reserves ‘drehb’ too; which is located at hash(drehb+1) - but since both of your version numbers of the appendables are the same at first the algorithm will probably first choose your website if someone wants to visit safe://drehb - now your evil twin wants to make sure that you don’t want to try to win this little war so he does 100 000 appends to the appendable so it looks like this now:

1: hisWebsite
2: hisWebsite

100001: hisWebsite
=> version number 100001 [after doing a put for each additional value]

with 1000 puts per dollar that did cost him 100 dollar and since your appendable (located at hash(drehb) would need to be altered 100002 times to have a higher rank again than his reserved name you have the option to either refer to your website as safe://drehb$0 or you just register dreh_b or something else that is not occupied yet.

ps: i know this sounds a bit like an insane idea and creates a lot of ‘unnecessary’ network operations … but it would result in a kind of continuous auction system where domain squatting would not be possible anymore

pps: and since mutables are limited to 1000 keys for now i’m not sure how the effect on the performance is if someone did a million appends and people always first check the version of the data + then request the last entry to end up at the wanted website…

1 Like

The big problem with this is that the safeweb is no longer secure. Mr Scammer just pays a little more for a popular site and shows the same pages but the details link to the scam and he/she profits off a popular safesite. It may take a week or more for the popular site to realise and to pay even more to get the name back.


That’s why this idea reminded me of petnames. There isn’t really any ‘truth’, you just need to pick the mapping that you like best.

Well - that’s why I mentioned that a warning should pop up when the value has been updated. (+the ability to have a link with fixed offset in it - links could automatically be resolved to those in the browser)

It’s not like the safe network would guarantee safety for its users in any form… :man_shrugging: if you don’t use explicit names there is room for mistakes - yes

But imho that price would be a pretty small one compared to large scale squatting at day 1 +once taken names are lost forever if the owner looses access (+large corporations that want credibility can stake large amounts on their name so it wouldn’t be cheap to pull off such a stunt)

or alternatively because of the squatting multiple name services competing and none of them becoming ‘the one’ => the network loosing attractivity because names are a mess or taken there …
[and it’s not like you could expect it to be expensive to just reserve every name appearing in a dictionary… just to sabotage the naming system or make money off it…]

Ps: since I just saw that this really is exactly how the current name system works (just without the offset+ranking if more than one possibility exists) this wouldn’t even be a large change to the name system used atm… And I may mention that just reserving a name costs 1 put - no more… So super cheap… Even in the current limited network one can knock out 1000 names without even having to ask for an additional invite… (PySafe will have the capability to do multiple calls in parallel - so squatting the words in a dictionary should be an affair of some seconds… But even if it’s minutes, hours or some days that doesn’t change anything… ) and since it’s just ‘random spread mutables’ nobody can identify them as registered public names :man_shrugging: in contrast to regular user data… And I guess we don’t want to throttle regular user activity…

Pps: btw - are current data maps for name resolution encrypted with their corresponding public names @bochaco / @joshuef ?

One of the fundamentals is for the non loss of web sites. Yes I know that could involve going back through the ADs to retrieve the site of a year ago.

But if you make it that safesites can be taken over by paying a little then lets say we have a education site (now not being updated) but was supplying a lot of great information on the shape of the earth. Rotation speeds how to detect rotation, gravity, and a wealth of information. They originally took a scam site away from the scammer who no longer cares

Now others decide to scam that site that was so useful for people wanting the ways to find out about their world. And we have flat earth scammers, flat earth believers and hollow earth and who knows what buying the site name (over and over again) and changing key parts of the site to make it appear all the science behind gravity, rotation etc is faked.

Other sites link to the site and the visitor is now presented with proof that the scientists are scamming the world and the various science around gravity, rotation, shape is fake.

So now the visitor saw a warning but what do they do because the original also said science was lying.

This becomes so messy

Now multiply this problem with 1000’s of important information sites that are in a war over their “perfect” name and every new name they try.

I doubt it will work very well when you consider scammer, anonymous, nasty people, etc, etc decide to wage a well funded war on 1000s or millions of safe sites.

It will make safe sites useless when these wars rage in safe space.

Its not about making SAFEsites safe since scammers can make any site anyhow, but it is about making it reliable when you do go to a site. So many people ignore warnings especially if the site looks correct. You want the site to remain in the hands of the original site builders. Yes there is a huge problem of squating but its better than scrambling the site names simply by paying a little more for the name and having wars on keeping the name

Well - it just claims to never loose data - there is absolutely no reason to keep track of were precisely my public name pointed when… (and btw that’s just as true for the currently suggested system) if you want reliable links you link to the xor name (+revision number - because that is not included in the public name either and in your logic is a huge mess too)

In that case we could go back to just using the hash of the name couldn’t we? It’s not like testing a method like this would destroy the network for all eternity (especially if applied in test networks…)

… Or even after 100 years just ‘restart’ the naming system as simple first come first serve with a different hashing algorithm :man_shrugging:

… I don’t see why we should start with the system that for sure will have problematic properties because we are afraid to test something new that might have unwanted properties/use cases :man_shrugging:

Isn’t all that basically allowing users to virtually steal ownership? virtually because when you fetch a website/webid/webapp/etc. (even wallets published behind a URL) you don’t know who owns it. Or, isn’t that allowing only richest to play a “game” and removing access to the rest to have/own a site, i.e. if I publish my little humble website with my favourite pictures (even with an ugly name) I never know when someone will steal it unless I keep playing that “game” forever?