- From: Michael Toomim <toomim@gmail.com>
- Date: Wed, 5 Feb 2025 02:21:33 -0800
- To: Julian Reschke <julian.reschke@gmx.de>, ietf-http-wg@w3.org
On 2/5/25 2:07 AM, Julian Reschke wrote: > >> You can think of a Version as a "Version of the World", and within the >> World there are Resources. You do not know all the Resources that exist >> in the World. So you need a way to refer to a Version without regard for >> any particular Resource. Thus, your Version must be an identifier that >> exists separately from any particular Resource, and then you also need >> to be able to specify a Version of a Resource. > > Yes, a commit. In git it is sometimes called a commit, but also each item in the stash, and the working copy, all fit the definition of Version. Other Version Control Systems have even more features that fit the definition of a Version, with more terms, and you'll find ever more terms in CRDT and OT research, databases, log-structured filesystems, and transaction memory systems. If your problem is that I'm using the term Version to mean "a point in distributed time", and you have a better idea for what to call this concept, then please let me know! >> If a Version is an identifier that exists apart from any particular >> Resource, then what it is actually identifying is a point in distributed >> time, because a point in distributed time is precisely what uniquely >> identifies that Version of the World, and also that Version of any >> particular Resource. > > Is a git commit a "point in distributed time"? Yes. >> I don't think you understand the difference between Option 1 and Option >> 2, yet. It seems that you are critiquing Option 2 from an Option 1 >> mindset. > > I believe I do. Can you please restate what you see as the difference between Option 1 and 2? Then I can see where we have common ground in terminology, which will really help in moving this conversation forward! Thanks, Julian!
Received on Wednesday, 5 February 2025 10:21:39 UTC