URLs are fragile. Yet we persist in using them in bibliographic references in published books and journals. There is a better way.
Why not use a URN instead? Any capable search engine should be able to resolve the URN to a URL. In many cases, just a UUID would work. Stick a UUID in your document, publish the UUID, and then let anyone who wants to find your document search for the UUID. A standard way to use URNs could eventually be established. If we used URNs, documents could be moved around but we could still find them.
Some things could go wrong. URNs could be copied in order to spam those seeking a popular document. Modified documents could be disseminated, with the modifications difficult to detect. I think these problems could be fixed, but a solution doesn't come immediately to mind, except that one could include a domain name as part of the URN. A domain name cannot be easily hijacked, which would mitigate the spam problem.
Using URNs could also open up the door for content to be disseminated in other ways, such as file sharing networks.
So, here's a little test. I'm putting this UUID into this post. I'm going to try to find this post later by entering the UUID into a Google search.
urn:uuid:59c5b81b-49b4-4df5-bfb1-edda2e86cd66
Posted by Doug Sauder at November 18, 2003 08:07 AM