Hacker News Reader

Comment by CaliforniaKarl | Hacker News Reader

UR_:

I: Identifier, is an identifier. "Oh, that's a [a thing, or a location]."

N: Name, tells what the thing is. "Oh, that's the Sears Tower."

L: Locator, tells how to get to something. "Oh, that's at 233 S. Wacker Drive Chicago, Illinois 60606"

That was the second example to come into my head (the first was the Bank of America building in San Francisco), and I think it speaks to things that both* examples have changed their names! The Sears Tower is now the "Willia Tower", and the "Bank of America Center" is now "555 California Street".

And even place names don't stay the same! In 2019, Stanford University renamed "Serra Mall" to "Jane Stanford Way": addresses changed; building names did not.

Just goes to show why it's so important to implement redirects, and (in HTTP at least) why the "301 Moved Permanently" response code exists.

View on HN

> That was the second example to come into my head (the first was the Bank of America building in San Francisco), and I think it speaks to things that both* examples have changed their names! The Sears Tower is now the "Willia Tower", and the "Bank of America Center" is now "555 California Street".

I was under the impression that a URN is an identifier that is not tied to a specific location, in fact it could apply to a resource that exists at many different ones. Is the "Sears Tower" example (or anything that could reasonably be associated with a single location) really the best in this case?

Here is a real URN which is also a good example of what you’d use a URN for - urn:isbn:9780133488838 - that’s one of the volumes of Knuth’s TAOCP

Thank you. Most examples of URNs that I've found use ISBNs to convey their use. I just wish that there was more information out there about use cases. This post's article does make a good reference to one that I found interesting though.

The RFC for a URN scheme for federated content: https://www.rfc-editor.org/rfc/rfc4198

I ended up using this for XML namespaces.

Also somewhat relatedly, XMPP tends to use URNs for the XML namespaces af the various stanzas, which in turn define the various parts of the IM message or what have you.

It’s hard to imagine URNs without a DNS based namespace to back up the universality part of it.

  urn:catalogue.isbn.org:1234

You could use a URL for information about the book as the URN for the book itself? Or is that “bad” because in 25 years we might not use either https or DNS?

  https://isbn.org/catalogue/1234

If the URN is a kind of type prefix for an otherwise meaningless number, are we supposed to use these inline / Hungarian notation types everywhere the number is used?:

  https://amazon.com/books/urn:isbn:1234/purchase

This is how XML namespaces work. They're URLs, but they identify resources rather than locating them. Often, conventionally, some information about the identified thing also happens to exist at the URL.

Reply on news.ycombinator.com