The current domain name system (DNS) is difficult to use and creates artificial scarcity. I propose replacing it with an intuitive, decentralized system that is based on the way that humans naturally use names.
This may be complemented by the Distributed DNS project at Sourceforge: http://distributeddns.sourceforge.net/index.php?page=news
This intuitive system would require that each machine (PC, server) maintain its own name registry. The user would simply type in the name of the machine/website that he wants, and it would be contacted based on the information on that registry.
The user could define a number of registries to reference (for example: local, ISP, Google). The user could modify the local registry by a simple tool in any network software (such as the web-browser). If multiple websites are identified by the same name, the software would either open connections simultaneously (in tabs, for example) or provide the user with more information and ask for clarification (similar to results from a search-engine).
A user could seed his own registry with information from his ISP, a search engine, or a friend.
This would ultimately get rid of the scarcity of domain names, which drives companies to pay millions of dollars for good names, drives political fights in the organizations that manage the names, and drives late-comers to have unreasonably long names. This system could also make mis-spelled "web-addresses" less annoying.
I agree
No opinion
I disagree
Add your comment
It would take getting used to at first, but it certainly is a more elegant solution. Hopefully it won't go the way of Dvorak.
I forgot two points:
First, a person's local registry could expand by collecting information from web-links as that person browses the web. Names could be taken from the text that points to the webpage or from the page's title. This would be an easy way to eliminate the "machine language" that is placed in URLs, such as the directory information and the file extension.
Second, the main problem that I see with this system is the difficulty in updating the information in the registry, for example if a webpage is moved to a new server. The current centralized DNS just requires that the webmaster submit the new information to the DNS. A decentralized system would either need <ul><li>a centralized component to keep track of changed IP addresses, <li>a way for the server to announce the changed address, or<li>a way for the clients to track down the information after the address has changed. </ul>
To communicate these changes, the server could first notify any machine that connects to it that it will soon be moved to a different address. If a client can't find a particular machine, it may be able to get the new address from another machine that it has reason to believe may have this information...such as the Google servers or perhaps another web-server that was linked to the previous server.
I came across this article on New Scientist describing the political battles over the DNS system:http://www.newscientist.com/article.ns?id=dn7757
Further ideas:
It is important that this system can be implemented on an individual basis. Even if the rest of the world uses the traditional DNS system, individual users can adopt this system, so it can grow slowly instead of needing to be implemented all at once. Of course, the more people who use it, the more efficient it will be.
This system relies on manual confirmation that information has been retrieved from the proper server. This probably wouldn't work for email, so there'd have to be some sort of automatic confirmation that the email was sent to the proper server. A public key encryption system may be used to make sure that messages don't end up being sent to the wrong recipient.