By Jason Levitt
January 21, 1997

The Internet Zone Takes You From Earls (URLs) To Urns (URNs) To Answer The Question: "How Can I Get A Web Address That Will Last Forever?"

"H"-"Tee"-"Tee"-"Pee", "Colon", "Slash"-"Slash", "W","W", "W", "Dot"...

How many times have you entered a Web address into your browser and gotten back "File Not Found"?

Web addresses are ephemeral, to be sure, and they are constantly getting shuffled around and renamed. What everyone wants, of course, is a Web address that will last forever, not something that can be outdated by simply being renamed or moved to another machine.

If you're a Web administrator at a large Web site, you're intimately familiar with this problem. Even if you've done your best to keep your Web site's links up-to-date, there still may be old information in the various search engines (Yahoo!, Lycos, Alta Vista, etc.) that frustrates visitors who click on outdated links to your site. To keep visitors from staring at "File Not Found" errors, many sites will redire ct browsers to a special page that lets them know they've tried to visit an outdated or nonexistent link. Take for example, the page returned from IBM's Web site when you try to go to the nonexistent page http://www.ibm.com/junk.html.

Sites that aren't quite as friendly as IBM's will simply let their Web server software return the standard error message. Some sites will also return "File Not Found" instead of "404 Not Found." In any case, you get the picture. Web addresses are easily outdated and difficult to maintain. It'd be nice if there was a more permanent way to identify a Web page.

(In Web parlance, addresses are called Uniform Resource Locators or URLs, pronounced "earls" For a more detailed description of URLs, see NCSA's "Beginner's Guide To URLs, A Solution.")

So ho w can we keep URLs, those ungainly strings of punctuation that now grace everything from billboards to milk cartons, from getting outdated? The short answer is, we can't. However, several Internet standards groups are working on defining a new identifier, called a Uniform Resource Name (URN, pronounced "urn"), that can serve as a permanent pointer to some resource on the Internet, such as a Web page.

URNs look something like URLs, but unlike URLs, they identify resources on the Internet in a location independent manner. But most importantly, URNs last forever (or as long as you want them to last, at least).

Here's how they work: When you enter a URL into a browser, the browser software finds the Internet Protocol (IP) address of the target Web site and fetches the page designated by the URL. For example, when I entered http://www.ibm.com/junk.html into my browser, the browser software resolved www.ibm.com into an IP address and then directly contacted the Web server on the machine www.ibm.com and aske d for the page named junk.html.

URNs, however, are resolved differently and therefore have special properties. Let's assume the folks at IBM have created an URN for the page junk.html.. You might refer to it like this; urn:weburns.internic.net:myjunk. Here, "weburns.internic.net" is a machine that handles URN to URL mappings, and 'myjunk" is the name we happened to give our Web page.

If browser software of the future is designed to understand URNs, then you'd enter that URN into your browser just like you do with URLs. However, the browser would handle it differently. First, the browser would contact the machine weburns.internic.net and ask it to provide the URL for the page that we've called "myjunk." The machine would then send back the URL (http://www.ibm.com/junk.html), or possibly send back the page itself.

Here, the machine weburns.internic.net is acting as a naming authority for the resource labeled "myjunk." The actual URL of the page could change, but the user would always find it becau se the naming authority, running on the machine weburns.internic.net, would always return the latest URL. Of course, it's up to the Web site administrator or whoever wants to maintain the URN to make sure that the appropriate naming authority is updated when the URL is changed.

	URN (Uniform Resource Name)	URL (Uniform Resource Locator)
Example	urn:nameauthority.ibm.com:myjunk	http://www.ibm.com/junk.html

Syntax:
The syntax of URNs will look something like this. urn: <NID> : <NSS> The string "urn:" serves the same purpose as "http:" in an URL--it identifies what protocol should be used to resolve the entire string. NID stands for Namespace Identifier, and NSS stands for Namespace Specifi c String. Although the example above uses a Web page, you can use URNs to identify any resource. An URN for a book might look like this: urn:isbn:X-13453829. For the current specification of URN syntax, see the Internet Engineering Task Force's draft on URN syntax.

The Future
It may be another six months before a formal RFC (Request For Comments) standard on URNs is published, and possibly six months more before URN support appears in browsers, but the possibility of persistent Web addresses, expressed as URNs, will make it worth the wait. Some sort of naming authority, possibly added to our existing Domain Name System (see references below), will also need to be in place if URNs are to be globally accessible.

More URN Info

WC3 Paper On Web Addressing Techniques a good introduction to URLs, URNs, and related stuff

URLs (URNs are mentioned) are defined in RFC 1738

This paper and this paper talk about using our existing Domain Name System (DNS) for URNs.

An interesting progress report (but outdated) from D-Lib Magazine

URN syntax definition

An NCSA archive of URN info is here

The URN mailing list for ongoing development discussions: Send E-mail to majordomo@bunyip.com with "subscribe urn-ietf" in the body of the message.

URN mailing-list archive at ftp://ftp.bunyip.com/pub/mail-lists/urn-ietf.archive

The home page for URN work is located at http://www.bunyip.com/research/urn-ietf

Return to AuthorITies or view AuthorITies archives

Comments?

http://www.informationweek.com

---