- From: Michael Toomim <toomim@gmail.com>
- Date: Fri, 1 Nov 2024 15:09:05 -0700
- To: Julian Reschke <julian.reschke@gmx.de>, Josh Cohen <joshco@gmail.com>
- Cc: ietf-http-wg@w3.org
- Message-ID: <5d689a18-8c48-45c5-835b-046cdfc13f9a@gmail.com>
Julian brings up an important deep distinction here: On 10/31/24 11:35 PM, Julian Reschke wrote: > A version *is* an HTTP resource. It has content; it has metadata, it can > be retrieved with GET. > > It seems we need to work on terminology: version-controlled resources > (aka versioned resources, mutable), and their versions (snapshots, in > general immutable). All of these are HTTP resources, because you can > interact with them using HTTP. Let me phrase this as a question: What's the relationship between *versions* and *resources*? I believe we agree that a resource *has* versions, because a resource changes its state over time, and each state can be identified with (or as) a version. We also agree that a version *can* be served *as* a resource, as this occurs e.g. whenever I make a "index2.html" file. However, I disagree that a version *necessarily is* a resource. Julian argues that a version is an HTTP resource because you can interact with it using HTTP. However, I don't think that being able to interact with something over HTTP is enough to make it a resource. If my resource /foo has two representations (one in HTML, and one in text/plain), then I can interact with those representations over HTTP, but that doesn't make each representation a separate resource. Likewise, I think a resource can have multiple versions, and those versions do not necessarily need to be their own resources. This semantic distinction seems to be at the root of our different perspectives on the versions-02 draft. Thoughts?
Received on Friday, 1 November 2024 22:09:12 UTC