h a l f b a k e r yA dish best served not.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
If you know how DNS stuff works you can skip my shoddy overview/background info.
BACKGROUND INFO: Every computer has a IP (i.e. 65.48.13.10). When you go to a web address (http://www.halfbakery.com) your computer connects to a DNS (Domain Name Server). There are a set number of established DNSs
[servers]. Most people have to pay to get on this list. So when you buy a domain name (www.halfbakery.com) you then typically pay to get on a DNS server so that when people type in www.halfbakery.com they get an IP that is the actuall address of the webserver.
IDEA: My idea is to start a new DNS type or standard. This would solve some "Internet 3"'s (Post by JakePatterson) proposed solutions.
The service would allow people that have an IP (probably static only) to register a domain name for free. This would open up a fun "Can-O-worms". BUT the kicker is there can only be one domain name per IP. Also there could be NO selling of domain names. People could report someone selling a domain name and it would be reset...aka open to the public. That way there could be no grabbing/hording of domain names.
Another fun feature could be a 2D address. An example could be:
"half
bakery" OR
"OO
>
__"
the FINAL bonus would be that you wouldn't have to conform to ".net" or ".com" your address could simply be "halfbakery"
*note that for some of these features to work there might have to be some browser "tweaks" or redesign.
AlterNIC.org
http://www.alternic.org/
One of many groups who've tried to take over the DNS namespace [phoenix, Dec 03 2001, last modified Oct 04 2004]
...and what happens when you don't play nice
http://news.cnet.co...-200-323630,00.html
[phoenix, Dec 03 2001, last modified Oct 04 2004]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Destination URL.
E.g., https://www.coffee.com/
Description (displayed with the short name and URL.)
|
| |
This doesn't address the problem of namespace clashes. |
|
| |
I think you are saying the clashes between the old system and the new...if that is the case the new dns server would after searching itself (and if it found nothing...no dns entry) would search the old/standard DNS. |
|
| |
1) AlterNIC exists. Just point your computer toward it if you want to use it. |
|
| |
2) IP addresses are dynamically assigned (there is nothing stopping me from stealing the HB's IP address). |
|
| |
3) A DNS zone in which every machine is in the TLD would take forever to search (thus the hierarchical DNS system currently in use) |
|
| |
reguarding phoenix's comments: |
|
| |
[1] That is similar to what I was thinking except you are stuck with a "dot something" still. |
|
| |
[2] I would like to see you take HB's IP...they have a static IP. Just like my server...if you are outside the block I am in you can't get it. |
|
| |
[3] You could make it hierarchical still...you could break it down by first character etc. |
|
| |
phoenix: The names could be storred sorted, allowing a fast binary search. Insertion into the table would be more computationally expensive, but that doesn't need to happen as often. Removal would be much more computationally expensive, since on average half the list would have to be moved. However, there are many other data structures (sorted linked list, hash table, binary tree, etc.) that could be used to avoid the need for a heirarchical structure. |
|
| |
This idea as stated has no appeal whatever to me, as I rarely consider static IP addresses (or top level domains, for that matter) unless I'm linking to or tracing to an address. I also have no e-pire building aspirations, so on with regard to internet3. |
|
| |
I will, nonetheless, pass out pastries to all for the germinal idea of 2-D domain names. This is really cute, and though the rendering of these names would probably require a layer of code for which there is no current standard, it might extend the fun of symbol searching. On our intranet, by way of example, you can search: ... , ./ , and so on and get hits; so, why not extend that to domain names, image metatagging, rollover .gifs and the like. There'd be a lot more room to play out in the open. |
|
| |
Thanks, part of that could be in reguards to business cards/CDs/Diskettes. They could hold these images/text layouts. |
|
| |
[benevia]
1) The dot-something is due to the hierarchical nature of the system.
2) Lessee: PING www.halfbakery.com (for IP adx); right-click on 'My Network Places' (Win2k Adv Server, sorry) select 'Properties'; right-click my public network connection, select 'Properties'; double-click TCP/IP entry; change 'IP address' to 207.46.230.219. Done.
{Please don't beat me [jutta]. I didn't click 'Apply'} |
|
| |
[JakePatterson] No other system allows for unlimited growth. |
|
| |
[reensure] How would you input the query to search for, say, the Coca-Cola logo? [benevia] implies you'll insert a CD-ROM and search for it (or scan a business card). In what way would this be better than typing cocacola.com? To search for domain names, type one in the address bar at the top of your browser. To seach for a particular meta-tag value, go to Google. |
|
| |
[phoenix]
On the contrary, a sorted linked list would have just as much scalability as the current method, it would require storrage space proportional to N where N is the number of domains, a proof that no method exists that would require significantly less space is trivial. (we are talking orders of magnitude here.) A domain quary would be 0(lgN), as would the insertion or deletion of a domain. Insertion or deletion of a record in an unordered array is O(1), which is better, that comes at the cost of doing the quary which becomes O(N) which is much much worse. Dividing the data into seperate heirarchies is simply a lousy hack that decreases the size of N by a constant factor for a particular quary. |
|
| |
phoenix: what kind of experience do you have with networking...my guess is not a whole lot. While you can set your computers settings to think that is your IP you will be unable to establish a connection to the internet. |
|
| |
In addition: please keep in mind that these are "Half baked" ideas...we are just kicking them around. Please try not to be a "nay-sayer". Thanks a lot :) |
|
| |
[benevia] Working with computers since 1979 (before your time, I'll wager).
MCSE, MCP+I, MCT, A+, Net+, I-Net+, CNA, CCNA
It's all numbers, bro. Just gotta know how to work 'em.
[JakePatterson] I realize that other storage methods exist (I do a fair amount of database and SQL work), but you'll find that:
a) They're not extensible and/or
b) They'll take too long to resolve queries and/or
c) The number of name servers required would be extensive. |
|
| |
Nonetheless, feel free to give it a go. I've done my best to show you the shortcomings. I'll leave it to you to prove me wrong. |
|
| |
¯phoenix: Ok, let me give this a shot and try not to graft on more to explain. \ … How would you input the query to search for, say, the Coca-Cola logo? … \ I wouldn't, necessarily. I'm sure there are image copy-protection efforts underway now, and it seems to be difficult to copy-protect anything these days, AEB what I've heard will happen to the next gen of Windows XP, only to be available online. |
|
| |
What I see, regarding 2-D domain registry, is the potential for a going concern like the Coke™ logo, which is essentially three stylized stripes red-white-red if you will, to be given a static 2-D domain name that could be registered as:
Coca
—
Cola |
|
| |
Instead of cocacola.com, cokeisit, therealthing, tastesgoodtasteshappy.cng, or coca-fn-cola. Would this only be used to address a registered hash-protected logo, or would it be used as a pointer to coke.com? Ya got me by enough years that I'll not guess, but keep eatin. |
|
| |
[reensure] I'm not opposed to the suggestion of a graphical DNS service (nor am I opposed to modification or expansion of the existing one despite appearent evidence to the contrary). I'm just curious as to how you would input the query (at least without having an example image). I suppose it would be possible to adapt face recognition technology to read a product lable via a PC camera (is that a separate idea?). Hold a can of Coke up to your web cam and say "Go get it Lycos". The task would be complicated by the fact that a Coke can doesn't resemble a magazine ad which doesn't resemble a power tool. |
|
| |
Another take on this concept was the :CRQ :CueCat (yeah, that's how they spelled it) that tried to take off a few years ago. The device was a wedge barcode scanner (free) which, when run across a product barcode would automatically open the appropriate company's web page in your browser. The idea was that barcodes are ubiquitous, unique and 'free'. I still have the one Wired Magazine gave me (never used it), but I bet there aren't many others out there who still have one and I believe the company is no longer in business. |
|
| |