[TELECOM Digest Editor's Note: You are quite welcome, I am sure. I am always glad to be of assistance. PAT]
'Directory assistance' provides a 'name to number' mapping function, nothing more. Registrars provide a 'name to number' mapping function, nothing more ...
The only real difference between the two is that the 'numbers' are in different address-spaces.
The PSTN does not rely on 'directory assistance' for the basic functionality. calls *must* be placed to a 'number'. If you have a 'name', you must *first* translate it into a number, before you can attempt to make contact.
The Internet does not rely on 'naming services' for the basic functionality. Packets *must* be sent to a 'numeric address'. If you have a 'name, you must *first* translate it into a number, before you can attempt to make contact.
It is 'convenient' to remember and use names instead of numbers, and to use 'directory assistance' to map those names into telephone numbers.
It is 'convenient' to remember and use names instead of numbers, and to use the registrars databases to map those names into "internet" numbers.
The 'essentially identical' nature of the operations should be obvious.
(A) sometimes it is the _end-user_customer_ who tells 'directory assistance' what information should be there -- name *and* number, maybe including address. And directly _pays_ the operator of the directory-assistance service to carry that information.
(B) When you register a name, you have to provide the *address* of the machine(s) that will answer questions about things 'under' that name. If you fail to provide the addresses for those machines, then *nothing* works.
Rhetorical question #1: How does the PSTN know how to route your call, if directory assistance doesn't know about your number?
Rhetorical question #2: How do you think the Internet functioned _before_ there were 'root servers' (and DNS)?
It is really *easy*. The 'root servers' are *NOT*INVOLVED*AT*ALL* in getting packets to a _numeric_address_. Each and every router on the entire Internet has a set of 'forwarding rules' in it, that describes, for _every_possible_ address on the internet, where the 'next step' en route to that destination is. That is _all_ the router needs to know; where to send it 'next'. and that next step does the same thing. "And so on, and so on." Eventually, by recursive application of that 'send it to the next step along the way", it arrives at it's destination.
To make things "easy" on (a) users, (b) application software, and (c) software developers, the standard 'name to address' look-up functionality has always worked on the basis of 'given a name as input, go find the address for it; given an address as input, simply return _that_ address." Note that that latter functionality does*not* require any consultation with the 'root servers', or anything else (even a 'hosts' file) for that matter.
And you get a circuit, and some IP address numbers. *PERIOD*. That is _all_ you get when you buy basic service.
You do *not* have to do that. You _may_ ask them to handle "all that stuff" for you, but you're engaging in a purchase of 'additional', _optional_ services from the ISP. when you do that. Depending on the provider, they may offer to do it 'at no additional cost', or they may charge for it. Price on the 'basic service' is better from those who _do_ charge extra for that optional service.
Of course, you can be known to the world at large as 'spam.com'*without* any intervention by that ISP. *WITHOUT*, in fact, the ISP even being aware that you are using that name. You can either contract with "somebody else" (other than that ISP) to handle the DNS-related stuff, or you can 'do it yourself'.
My ISP, for example, has no idea what domain-names I am using for what machines, at which of the addresses they supplied me. I can change the host names, and domain names, any time I choose. Without their knowledge or consent. I can add a new domain name, and deploy servers under that name, and the ISP has no knowledge, nor any awareness that I have done so.
True. The ISP *tells* you that you _will_ be addressed as '184.108.40.206'. You do not have any real choice as to what your 'number' is -- you must use whatever number(s) you get assigned.
EXACTLY the way that the telephone company tells you what your phone number will be.
You apparently "don't know what you don't know" about how the process actually works.
The ISP says "I will play middle-man with the directory service, if you want me to, or you can have somebody else do it, or you can deal with them directly yourself."
The CLEC says "I will play middleman with the directory service, if you want me to, or you can have somebody else do it, or you can deal with them directly yourself."
As you say, "quite a difference."
A registrar doesn't "assign" anything either, it simply reports info on what names are 'in use'.
There is a design difference in the architecture -- the name-to-address mapping service in the Internet realm requires that names be 'unique'; In database terms, you are only allowed one record with any particular 'key'. The telephone 'name-to-address mapping service' is not that restrictive. There can be several "John A. Smith" listings, with different numbers. How do you know _which_one_ is the one you're looking for? That _is_ the problem -- there's no way to tell. You have to get all the numbers, and call each one and ask "are you _the_ John. A. Smith that...?"
Totally ignoring the fact, as described in the botch-edited material above, that the spammer can do everything he needs, *without* relying on a domain- name _at_all_. Domain names are a 'convenience', nothing more. HTML makes it 'trivially doable' to 'conceal' the fact that one is *not* using a domain-name -- _and_ to give the appearance of using _somebody_else's_ domain name, but -actually- connecting to your own servers instead.
You continue to display your lack of understanding of how things actually work.
There isn't any need for any sort of 'under the table' dealing. You just order basic service from the ISP. Period. You can then: (a) handle domain-name stuff _yourself_, without *any* ISP involvement, (b) contract with the ISP to handle it for you, (c) contract with 'somebody else' to handle it for you, or (d) not bother with it _at_all_. ALL the ISP knows is whether or not you contracted with them to handle things. If you didn't, they have no way of telling whether (a), (c) or (d) applies. Nor do they care -- either way, _they_ aren't providing any related service, and that is the entire extent of their interest (more properly lack thereof :) in the matter.
Almost all big commercial accounts buy 'just connectivity' from the ISP, or more likely ISPs (plural) that they use. And handle all the 'other stuff', including interfacing with 'directory assistance' themselves. The people that provide Internet connectivity to General Mills don't have _any_idea_ as to what domain-names are being used. They don't care either -- *all* the data packets they see have a _numeric_ address in them; all they have to do is get things to the proper numeric address, and let the customer do whatever processing is appropriate.
ISPs _alone_ could completely clean up the mess. If they wanted to.
The problem is that, collectively, they *don't* want to. And there is nothing that we, the users who _do_ care, can do to make them change their mind about it -- as long as there are "sufficient numbers" of people who are willing to buy services from those 'uncaring' providers.
Since 'many' ISPs are demonstrably *not* interested in doing so, the idea of 'ISPs working in concert with registrars' is similarly nothing but a pipe dream. I wish it wasn't that way, but it *is*. "Reality sucks" applies.
Don't I wish!! Unfortunately that approach works *only* when 'almost all' of the players agree on, and _enforce_ the same set of rules. When the 'node' (or 'network') that carries 40% of _all_ the traffic in North America decides that they _will_ deal with 'John Q. Spammer', regardless of his history, it _really_ "doesn't matter much" what the 'rest of the world thinks' about it. They _are_ too big and 'too important' to be _effectively_ 'shunned'.
It's like the old joke: "what are the little brown bumps between elephant's toes?" Answer: "Slow natives."
And you apparently don't remember the great schism in Fidonet -- when two major nodes blackballed each other; and the 'rest of the world' had to choose sides. Resulting in two different 'fido nets' that didn't talk to each other.
Unfortunately, on the Internet, there is nobody in that '800 lb gorilla' position with the interest/gumption to do that black-balling. And when the pygmy tries it, he just ends up as another 'little brown bump'.[TELECOM Digest Editor's Note: Since you seem to have so many hassles with my editing, why don't _you_ start a Digest in which you could witness the Gospel to everyone? I do remember the Fidonet schism, and it was unfortunate, but it all eventually came back together did it not? And thats really what we need here on Internet, where a large number of the 'pygmies' as you call us, walk away and start doing our own thing, a sort of 'Internet2' approach. And when the 800 pound gorilla MCI comes around saying, "oh you must really be sorry about losing all our customers (who by and large, as Spamhaus indicates are spammers) from your circle of communications," my response would be "not really. Numbers do not mean everything; so now we have only 60 percent of the users we used to have ... so what .. we have the _quality_ users with us." Of the approximatly thousand items of mail this 'pygmy' recieved today, if I had not gotten 400 of them, and 395 of those 400 were spam anyway, somehow I think I would get over it. Yeah, Robert, you really have it made; start your own BONOMI Digest (as I offered to help you with when I sent back that 48 K-byte rebuttal you sent a couple weeks ago) and you will never again have to worry or fret over my 'ham-handed' editing. PAT]