Is It Time For .tel? 292
Vitaly Friedman writes "ICANN, the body responsible for creating top-level domains, is considering a new one. Conceived as a way to easily manage contact information in an age where many people have numerous contact numbers, the proposed .tel TLD would allow individuals and companies to keep all of their contact information in an easily accessible location. Companies would get companyname.tel while individuals would be able to register firstnamelastname.tel." This idea has been kicked around for quite a while; one of the question is the whole name-space collision issue. For instance, there's me and then there's other me. Lemme tell how strange it is getting fan mail for country music stars.
eliminate top-level domains ? (Score:2, Interesting)
If I want a web site, why can't it be www.boxlight -- or www.boxlight.this.is.cool -- why does it have to end in
boxlight
Too little too late (Score:3, Interesting)
It's because they were so late to introduce a large variety that ".com" become synonymous with "web" and everybody wanted his site to be a ".com"
Should've they introduced domains like
An idiotic idea that shows domain names are broken (Score:3, Interesting)
Think about it. Do we still need domain names? People made up the "I'm feeling lucky" ifl: protocol as a joke, but isn't it true? Can't we find anything with Google anyway? Why should we have to remember a particular address with a complicated system of slashes and characters to get to a particular page? Right now, my URL is http://slashdot.org/comments.pl?sid=183301&thresh
Here's what I'm proposing:
Let's extend ifl: or something like into a real protocol. A trusted source, or better a network of user selectable sources, assigns keywords to URLs based on tagging by users via hyperlinks to the source and delicious-like tags. Normally, the URL bar shows nothing but the title the site has given itself (in our case, "Slashdot") and the particular page being viewed ("Reply to thread"), but on request, the URL bar can generate a user shareable set of keyword tags for the site with hash codes for pages to prevent collision (think about the addresses generated snipurl and the like; "ifl:Slashdot/4bacc23"). For the purposes of bookmarks, traditional URLs can be stored, but since these URLs won't be exposed to users, Ford Motor company can use a23rf2.ifl and Ford Modeling Agency can use j737bdh.ifl, and no one will care, since it won't be possible to hijack a keyword without the agreement of the majority of users. (No more Whitehouse.coms!) Domain names can stick around, so that people are free to assign multiple IPs to the same site, but the concept will become a background detail that users need to know nothing about. Until the technology is built into all browsers, URL-to-ifl translator sites can fill in the gap: "go to http://ifl.com/Slashdot/4bacc23 [ifl.com] or just ifl:Slashdot/4bacc23..." but since this won't be hard to integrate into browsers as a plug-in, I imagine it can be implemented quickly.
So, what do you guys think? Am I being naive about the possibility of the keyword space being kept pure without a registrar? Need I point out that the keyword space is *already* polluted, inspite of that barrier?
All contact info in one place - FOR TELEMARKETERS (Score:4, Interesting)
Yes, that sounds like a GREAT idea - I think I'll also put my social security number, my alarm codes, a Google maps link to my house, a picture of my house key, and my bank account numbers up there as well.
Look, if my company wants to set up a contact page they can set up a web page under their already existing domain name. If I want a contact page, I can set it up under my already existing personal web space. What does a new TLD add to this?
Now, *IF* they were talking about a new transport class (like http:// and ftp://) for encapsulating telephone numbers, such that a link to tel://8675309 would get me Jenny on the line, that *might* be useful.
But hell - I haven't even signed up for MYCALL@arrl.net to avoid being spammed by any asshole who scrapes my callsign (and I already have this one jackass who has done exactly that - he scraped my callsign and now he keeps adding me to stupid services like plaxo and the like, even though I've told this tool quite sharply that I don't want him bothering me.)
Re:Intended purposes (Score:5, Interesting)
Yes, but that doesn't stop plenty of people in the UK, like me, (ab)using the global
And why not? I'm no more a "com"pany or an "org"anisation than I am a "net"work provider. I'm not a "biz"ness, and I'm not dedicated to providing "info"rmation, and the domain is not my real "name". But nor do I want a country-specific domain -- my site is of very limited interest to the vast majority of people, but the tiny community it interests is spread right across the globe. My site isn't aimed particularly at people in the UK, so why should it have a misleading
What it comes down to is, there is no point whatsoever in trying to force an artificial hierarchy onto something like the internet, which is an interconnected network, not a neat and nicely categorised tree. It doesn't work. It's pointless and confusing. Let's just give it up already, okay?
Re:eliminate top-level domains ? (Score:2, Interesting)
Well... really what would have to change is (non-recursive-querying) resolver code, and since that is distributed to practically every Internet host, that likely would take time.
However, the server-side and administrative-side changes would stay largely the same, and there is no need to abandon hierarchical delegation of parts of the global distributed dabase.
There are two obvious approaches.
The first possible approach is to group arbitrary strings into three-or-four character groups and look up those substrings hierarchically. For example, "unformatteddomainname" would result in lookups for "name" "main" "eddo" etc., essentially as if it were originally written "u.nfor.matt.eddo.main.name." under the current system, with authoritative nameservers for the root and each subdomain.
An alternative hierarchicalization of arbitrary domain names -- implemented and demonstrated in practice -- is detailed in RFC 2317 [faqs.org] which allows for non-octet-boundary IN-ADDR.ARPA. DNS names, since classless inter-domain routing has a different hierarchy from the legacy DNS.
Re:Intended purposes (Score:3, Interesting)
Now that I think about it, the divisions into "com", "net", "org" help avoid collisions, but do so in a most useless manner. Say for example, that McDonalds is a purveyor of fine foods, and registers mcdonalds.com. Now, Old McDonald had a farm but now sells farm equipment and would like to register mcdonalds.com. Trademark law allows McDonald's Fast Food and McDonald's Farm Equipment to coexist, but it is unfair to allow one to register the domain but not the other. One soultion would be to deny mcdonalds.com from everyone, requiring mcdonalds-fast-food and mcdonalds-farm-equipment, but in a brand driven society this would probably fail for brands spanning a wide range of products.
I propose completely arbitrary TLDs to avoid collisions as an admission of the futility of attempting to classify domains with TLDs while keeping some collision avoidance. The new TLDs might be colors for example. mcdonalds.red and mcdonalds.blue would point to different sites, and McDonald's Farm Equipment would not be allowed to register in more than one TLD. The TLDs would need to be great enough in number and generic enough so that one TLD would not become more desireable than any other. That may be impossible to solve... Colors may not work, alphabets have an order (is mcdonalds.a superior to mcdonalds.b?), numbers have both an order and are not easy to remember, and any words might combine fortunately or unfortunately with the domain a la del.icio.us.
But, if an arbitrary TLD scheme could be come up with, typing "mcdonalds" in a browser would find all registered mcdonaldses and display them possibly along with a short description (from a meta html tag scraped from the website? or a new DNS entry?). If there was only one mcdonalds, then the browser would simply go there. This might be an unfair advantage to unique trademarks that are then able to keep out any business with the same name...
And for those who need domain to IP address lookups without going through a disambiguation process, they'd simply need to remember the correct TLD (red, blue, etc.) much as is done now.