- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Wed, 5 Feb 2025 11:19:22 +0100
- To: ietf-http-wg@w3.org
On 05.02.2025 10:56, Michael Toomim wrote: > On 2/5/25 1:29 AM, Julian Reschke wrote: >> As I explained, a version *is* a resource. In your explanation, the URI >> for it is optional. But the fact that it *can* have a URI by definition >> means that it *is* resource. Please read the beginning of RFC 3986. > > Julian, you seem to be claiming that a version *is* a resource. But is > it not more accurate to say that a version *can be* a resource? No. You just quoted the text that shows that it *is* resource (but maybe not an HTTP resource). > Just because something *can* have a URI does not mean that the thing > *is* a resource. My name (Michael Toomim) can have a URI constructed for Yes, it does. > it, like urn:toomim:Michael%20Toomim, but this does not imply that my > name is a Resource. It is my name. Likewise, I could define multiple How is it not a resouce if you have a URI identifying it? > resources for my name: > > urn:toomim:Michael%20Toomim > urn:foobar:Michael%20Toomim > https://toomim.net/Michael%20Toomim > https://example.com/names/Michael%20Toomim That's not a problem. > Now we have a one-to-many mapping between my name and resources. You > cannot say that my name is a Resource. If it's a Resource itself, then > which Resource is it out of that list? A URI is not a resource. It's an *identifier* of a resource. There's no problem if there are multiple ones for the same resource. Consider http://example.com/?a=b&c=d Many servers will choose to consider it to identify the same resource as http://example.com/?c=d&a=b > Similarly, in the Bank example I gave earlier, there is a many-to-many > mapping between Versions (e.g. "transaction-1") and Resources: > > Resource: Version: Value: > https://alice.com/alice "start" $20 > https://bob.com/bob "start" $20 > https://alice.com/alice "transaction-1" $10 > https://bob.com/bob "transaction-1" $30 > > Now, we could also come up with many URI definitions for "transaction-1": > > https://alice.com/versions/transaction-1 > https://bob.com/versions/transaction-1 > > Each of those are Resources. But the version "transaction-1" itself is > not a Resource. If it is a Resource, then which of those Resources is it? You claim that these are different resources. They are not. In git (the distributed system I know), a commit is a resource, as is the version of a resource at the time the commit was made. These are different resources with different URIs. > Again, it seems like you are thinking with an Option 1 mindset while > reading Option 2. Please try to understand the difference between mental > model of Option 1 and Option 2. > ... I believe I do. This discussion really is leading nowhere. Please read RFC 3986 (only the first sections), and consider how HTTP servers for Git (be it Github or Gitbox) identify these different concepts. Best regards, Julian PS: I will now try to stop responding until we have other participants in the discussion.
Received on Wednesday, 5 February 2025 10:19:28 UTC