The official domain name system is broken.
It's just a service anyone can run.
It's time that someone set up a better one with the original rules.
Amazon's recent moves to gain ownership of certain new generic top-level domain names have caused anger among writers and publishers. The company wants to take control of addresses including ".book," ".author" and ".read". The move has upset the Authors Guild and the Association of American Publishers (AAP), both of whom …
It's just a service anyone can run.
It's time that someone set up a better one with the original rules.
The domain name system was designed on the principle that domain names would be trustworthy identifiers of the people or organisations using them. The domain name system operates on the principle that domain names are stakes claimed in the limited number of symbol combinations that might be sufficiently recognisable to have a monetary value. No one is going to play by the old rules if there is more money to be made by playing to new ones. See under "banking" for more information.
The official domain name system was set up by a bunch of theorists with limited imagination and little real world connect for how their idea would actually be used. I mean, look at it: COMpanies, EDUcational establisments, ORGanizations. 19th century monolithic thinking meets 21st century infrastructure.
For all the talk, those "visionaries" had no real feel for exactly how empowering the World Wide Web would be (and a delightfully naive idea of what would be the most profitable for-pay leveraging of the technology from the get-go).
And when the marketers had gotten in on it, well, there went the initial intent. "Hey Mr Man-on-the-Street. What YOU need is your OWN DOTCOM!" (paraphrasing actual TV commercial from 1997-ish there). Classic case of an unenforceable standard.
The Genie is out of the bottle now, though. Can't go back. But it would be a refreshing change if the powers that be would get a fsking clue when it comes to not handing such useful domain names as "author" and "blog" to massive commercial for-pay enterprises.
Let's hope they market them responsibly and affordably.
Luckily it doesn't matter, as nobody will care more deeply than to think, in passing "oh, fancy address."
And to think Apple was the only one "monopolizing" the book market.
For Gods sake NO.
We would be better off without the expansion rather than let those buggers take over.
Between google searches, social media sites, barcodes with imbedded links, etc how many punters actually bother to remember a sites name much less which tld's it exists on? Is this a case of much ado about nothing or am I just too lazy? Of course there's the venerable bookmark and rss feeds. I'd be interested in learning the statistics of the preferred methods of finding sites... are tld's relevant?
"... are tld's relevant?"
No, not really once you've found the site & bookmarked it etc. but you've still got to get there in the first place.
The point (so I thought) of the new TLDs was to allow people/organisations/companies etc to get hold of simple/recognisable domain names now that .com etc are pretty much used up. Giving a whole TLD to Amazon or Google completely defeats the point of it (unless I'm missing something here!?).
But your average user just types amazon+ENTER don't they?
Well... isn't that a bit counterproductive? The more TLD's there are the possible combinations of FQDN's there are... ultimately making for an alphabet soup that the average end user will be annoyed to deal with. I understand the land grab mentality but don't major sites register on other tld's simply because they don't want anyone else infringing on their precious trademarks? How confusing would it be if it became common to have multiple distinct (popular) websites spread across multiple tld's?
I have to wonder if these new tld's make more sense from a need to reduce the amount of subdomains or to segregate user content to a separate domains to avoid security whitelist issues. I.e. you trust x.com but get infected by malware on blog.x.com because you whitelisted x.com and all subdomains.
I think there are a few things going on at once here. The system allows some big companies/organisations to have a big old global domain all to themselves. There's not really much problem (other than user confusion) if Google are allowed to own .google. Or Microsoft, Nike, Pepsi and the like. I guess some gentlemen from Columbia may wish to argue that .coke should unambiguously belong to Coca Cola... But here, you've got problems with Amazon and Apple (or Windows), where more than one user could have a legitimate reason to want the domain.
There's then the idea of regional domains, like .london, .scotland or .amazon. Although even that can create problems with things like .catalan.
And finally you've got the generic ones. I guess if some neutral party want to run .hotel, say along similar lines to .xxx where sites are checked up on a bit more, that might make some kind of sense. But what if Travelodge want to run .hotel, or Amazon to own .book? Then it just becomes a big old landgrab. I seem to recall that Google went for 20-odd gTLDs.
Don't know how you'd apportion .islam, but I don't understand why people are opposing the Vatican's registration. Surely the Pope is .catholic...
It's a right old mess. We'll have to retrain lots of ordinary internet users - or all this stuff's going to be ignored or a security nightmare. Or both. I suppose sending emails to bill.gates@microsoft might make some sense, as well as being where to go to download your software and look for info. Many users don't look at domains anyway though, so I don't see it as much of a gain for anyone.
If a registrar and the whole industry are willing to make an effort, there could be a security gain from controlled TLDs like .bank. As long as the existing big players don't just use it to lock out competition. Could kill phishing for bank details stone dead. Otherwise the generic ones just look like an opportunity to fleece companies for yet more defensive registrations.
I can maybe see the regional ones going down well in some places though. I wonder how many people would want a bob@london email address, or jock@scotland etc.?
Who decides who gets a .bank?
Just the Americans - so you would allow Lehman Brothers but no banks in the middle east
The swiss, Jersey, the Caymen islands?
Or do you let every country authorize .bank ? So liberia or somalia would happily hand out "h-s-b-c.bank" for $50
Who decides? In theory ICANN will appoint a responsible registrar, who have a plan to run a controlled domain space, where sites will be regulated (you need to be a properly nationally regulated bank in order to get a .bank address, plus there'll be daily (or hourly) scanning of the sites to make sure that there's no malware on them and that all the certificates are tickety-boo. That's my memory of the proposal. If it works, then browsers and email programs could be set to recognise the .bank suffix - and phishing for bank details would suddenly become incredibly difficult.
However, in practise, it may be that the potential registrar are just after a big fat pile of money for little effort. Who can tell? They've managed to convince someone they're serious enough to get $200-$300k together to put this application in - and they'll get shot down in flames by the banks and governments if they don't look to be on the level. But this process can go horribly wrong in several ways. However I don't see ICANN over-ruling the governments on this one, or there will be serious trouble, and they've already lost a lot of credibility with this whole screwed-up process.
The point (so I thought) of the new TLDs was to allow people/organisations/companies etc to get hold of simple/recognisable domain names
The point of the new TLDs is to increase the demand for domain names. They serve no other useful purpose.
gTLDs are largely useless in the first place. In the rare event that entities from different categories under the existing system want the same domain name - say a hypothetical School of International Business Management wants "ibm.edu" - one is likely to get the preferred term, and the other likely to choose a different domain name, simply because users generally don't pay attention to TLDs. That's not to say there aren't collisions in DNS today that are only distinguished by TLD; but those are clearly not optimal cases, and generally the entities holding such domain names have other, better-distinguished ones.
ccTLDs make a little more sense, but only a little, since non-technical users still typically ignore them. Often non-US entities of significant size will go ahead and purchase gTLD-based domain names anyway, so that browser "search common domains" functionality will find their sites. The main effect of ccTLDs seems to have been to give countries like Tuvalu a chance to sell some domain names.
The TLDs do make it possible to partition the naming space for assignment and search purposes, but there are plenty of other ways to achieve that.
 For example: neu.edu belongs to Northeastern University (redirects to northeastern.edu); neu.com to some sort of German dating site, apparently (redirects to neu.de); neu.org apparently to the International Automotive Components Group, bizarrely. Looks like no one has made a play for neu.net yet.
I asked if I could be registrar for .clusterfuck, but it turns out ICANN already own it...
I thought they all made their bids last year, so what's this new turn of events?
So it's fine for a company to own a generic .com (DIY.com) but not a generic TLD? TLDs are profilerating to such an extent it seems little difference to me,
I guess it's getting publicity now, as the first lot of new gTLDs are starting to go through the approval process.
If you remember ICANN instituted 'digital archery' instead of picking numbers out of a hat, so it was the company that could get their ping closest to the correct time on one of ICANN's servers. They got to go first. Anyway, that turned out to be a rubbish idea, and got iCanned - so who knows how they picked the order? But they split last year's requests into groups, so they could keep the squabbling
going for longer to a manageable level.
I'm not sure why we ever needed domain names. Telephones don't have a similar concept to domain names but do have a similar concept to IP addresses. But no one (these days) has to remember telephone numbers - we use built-in address books. Why not do the same for websites and save ourselves the expense of buying domain names from squatters or thinking up none-dictionary domain names.
Hmm, what happens if the ip address changes then? How do you resolve conflicts when multiple sites are hosted on the same machine? You'd need some level of indirection there, maybe if the ip address was just an identifier or a name that you could used some sort of dynamic name service to resolve the real address.
>I'm not sure why we ever needed domain names.
>Telephones don't have a similar concept to domain names
I'd say that's incorrect. Call a large company, listen to the options menu press 1. 1 is likely a 'tele-dns*'
lookup to a ring-group, which looks up individual extensions to ring. The PBX doesn't share that information with the public telephone company at large.
*tele-dns is some crap I've made up that represents the data lookup the system does. In large systems this can get quite complex.
This would be a piss poor option.
A single server may host anywhere from 50 to 500 domain names, yet only have a single IP address.
Depending on the setup, you CAN actually go to the domains via the IP address on some server, but this generally requires using something similiar to ###.###.###.###/~username/
Do you think anyone wants to have to give out their account username to everyone?
Or, how about the fact that IPv4 address space is dwindling/nearly non-existant and most services are not yet IPv6 compliant, so doing this would require 50-500 IPs per server?
Currently, on most hosting companies systems "shared hosting" servers (generally any plan under $15 a month) , unless a user is paying for a separate dedicated IP they are using a servers "shared IP" which is shared by all other users on the same server.
"Telephones don't have a similar concept to domain names"
Oh yes they do.
44 = uk, 33=france, 1=north america, etc
44-114 = uk.sheffield, 33-1 = france.paris, 1-212 = usa.new york, etc.
...large ($1000) deposits to obtain a domain name. When companies sign up for them in large quantities, it might be a deterrent. Of course, requiring the deposit BEFORE the name is issued and making it NON refundable would also help.
I make this note as "accountonline.com" is registered to Citi Bank, which makes little sense to me!
In case you wonder why the downvotes - large companies spend more on toilet paper than your proposed deposit. It will /not/ deter anyone who already is in quasi-monopolistic position. It, however, will deter smaller companies, associations, user groups etc . , thus creating ideal conditions for the corporations to claim every domain name conceivable.
"The potential for abuse seems limitless," argued Authors Guild president Scott Turrow."
..... as all El Reg readers have been saying since this was first announced.
How will adverts (in print or vision) for lesser known companies list their website address? Seeing "mydomain.com" . ".co.uk" etc. makes it obvious.
You'd end up with them saying stuff like "website address: mydomain" or "http://mydomain/"
"mydomain.com" is cleaner, and instantly recognised