Re: draft-toomim-httpbis-versions HTTP mapping (and WebDAV Versioning)

On 02.11.2024 03:09, Michael Toomim wrote:
> Julian in response to the following (bolding is mine):
>
> On 11/1/24 4:05 PM, Julian Reschke wrote:
>> It could be used to discover a version, but *that version should also
>> (in general) have a URL*.
>
> I disagree that every version should have its own URL in general. I
> argue that making URLs for versions is a separate concern that should be
> handled in a separate spec.
>
>> if the version has a different URI then the version-controlled
>> resource, is it a resource on its own?
>
>
> In my world, it does not have a different URI than the versioned
> resource. Are you ok with separating these concerns into separate
> specifications, and considering this to be a spec for versioned
> resources where only the resource itself has a URI? Additional URIs can
> be added separately.

I believe that adding them later would add another protocol layer that
isn't really needed.

FWIW:

>            Version: "dkn7ov2vwg"

If you change that to

              Version: "urn:ietf:....:dkn7ov2vwg"

it already would be a URI.

> I also suggest using the term "versioned" rather than
> "version-controlled", because the spec also supports distributed
> systems, which allow peers to create versions without central control.
> (Control is done with validation, which is also a separate concern from
> versioning.) The term "version control" arose in centralized versioning
> systems that used "controls" like checkins, checkouts, and locks to
> prevent conflicts. We are just doing versioning here.

Yes, that makes sense.

Best regards, Julian

Received on Saturday, 2 November 2024 08:06:20 UTC