- From: Bohdan Andriyiv <bohdan.andriyiv@validbook.org>
- Date: Tue, 19 Feb 2019 11:54:56 +0200
- To: "public-credentials" <public-credentials@w3.org>
- Cc: "Markus Sabadello" <markus@danubetech.com>
-1 - I think, DIDs should allow no-registry (no-ledger) methods . This would be _true_ decentralization. You've created DID, shared it with your friends, they have it and keep local provenance history for it - that's enough for you and your friends, no registries(ledgers) involved. This will still be DID and you kind of "registered" DID with your friends, so you would still need "register" and "registry" words around (e.g. "local DID registry", "global DID registry"). - "Ledger" is a rare, hard word. Another hurdle for understanding, that require additional explanation for uninitiated in blockchain subject. Let's simplify not complicate. For example, moving from Decentralized Identity to Decentralized Identifier was very good. Simple change, that brought a lot of clarity (now you do not have to discuss "What is identity?"). Let's continue in the same direction. No need to add another word that needs explanation. Depending on the context and audience one can interchangeably use registry/ledger/"decentralized registry", but in spec "registry" IMO is the most correct (general enough and simple enough) term. Bohdan ---- On Tue, 19 Feb 2019 06:54:22 +0200 Markus Sabadello <markus@danubetech.com> wrote ---- > Also -1 I think, "ledger" won't work for IPFS-based DIDs, or public-keys-wrapped-as-DIDs, or potentially many other DID methods; it doesn't feel broad enough. > Personally I also don't like the separation between "creating" and "writing" / "registering" a DID. This separation is something I have seen in Sovrin docs and discussions several times, but I think it's better to have a mental model where you have only one step for "creating" the DID. E.g. in Sovrin, you may create a wallet and a key pair as a preparatory step, but you haven't actually "created" your DID until you also write it to the ledger, because only then it becomes possible to "resolve" it. > Markus > > On 2/19/19 12:39 AM, =Drummond Reed wrote: > > On Mon, Feb 18, 2019 at 3:30 PM Kim Hamilton Duffy <kim@learningmachine.com> wrote: > > So I strongly believe that the sooner we fix this naming issue, the sooner we stop sending the wrong message to potential adopters about how DIDs actually work. > > I definitely agree sooner is better...if people are down for this exercise right now, I'm not stopping anyone > Cool. All in favor of moving from "DID registry" to "DID ledger", please +1. > If you strongly feel you have a better alternative, please advance that. > On Mon, Feb 18, 2019 at 3:26 PM =Drummond Reed <drummond.reed@evernym.com> wrote: > Kim, while I agree that it would be good to avoid a naming exercise right now, in fact a term was recently suggested to me that IMHO would be infinitely better than "DID registry". It is simply "DID ledger". > Note that the term "DID ledger" does not say "distributed ledger" or "blockchain" or anything that would imply that DID technology could only be written to one of those types of systems. In fact, "DID ledger" doesn't even mean that the ledger is decentralized. > What "DID ledger" DOES capture however is the idea that the DID controller writes the DID to the ledger. In all cases with DIDs, that's what happens (whether the DID is actually initially created entirely independent of the ledger, as with Sovrin DIDs, or it is created via the write transaction to the ledger, as with BTCR DIDs). > And that of course is exactly the OPPOSITE of what happens with "registries". The essence of the problem with the word "registry" is that it is always the registry that controls the rights to the identifier, not the registrant. > So I strongly believe that the sooner we fix this naming issue, the sooner we stop sending the wrong message to potential adopters about how DIDs actually work. > =D > On Mon, Feb 18, 2019 at 2:58 PM Kim Hamilton Duffy <kim@learningmachine.com> wrote: > I'm not sure we'll get a better candidate in the near future, but ditto on the problems caused by the use of the term "DID registry". > In fact, after my presentation at W3C Strong Authentication and Identity Workshop, I decided not to use that term unless I have ample time to qualify/caveat what it means. > At minimum, if we just mark it (perhaps create an issue) to revisit, that would probably be fine. Not sure we're in the mood for a naming exercise at the moment. > But also +1 to the improvements in this use case document. Great job Joe! > On Sat, Feb 16, 2019 at 8:37 PM =Drummond Reed <drummond.reed@evernym.com> wrote: > On Thu, Feb 14, 2019 at 8:01 AM Joe Andrieu <joe@legreq.com> wrote: > Folks, > > Based on the feedback from the call Tuesday, I have updated the DID Use Cases document. > > https://w3c-ccg.github.io/did-use-cases/ > > Please take a look and provide feedback. Please use the mailing list for general discussion and Github issues for specific places where the spec text could use improvement. Pull requests appreciated if you have suggestions for improvements. > > Joe, this is a big improvement. Thanks for doing this. I have some wording suggestions but unfortunately will probably not have time until RWOT to submit them, and they are minor anyway. > One terminology question, however: this is the first doc I've seen using the term "DID registry". While I get why that term seems attractive—it's the best analogy to the existing world of registries (especially DNS registries), I have avoided it all this time because the process of writing a DID to what the spec used to call a "target system" is SO different than conventional registries which ALWAYS involve centralization. This is true for every single target system I'm aware of. That's the whole point of decentralized systems—they don't involve the same power dynamics as centralized registries. > So I'm just wondering if the term "DID registries" has become established or if we can use a better term that reflects the unique nature of DIDs. > The key difference in this iteration is the addition of an extended discussion of what you can do with a DID and the 13 DID actions I've distilled from our conversations over the last couple of years. Hopefully this addition helps both with the big picture and gives concrete functionality. > > Note that not all DID Actions are supported by all methods and not all will be specified in the DID spec. However, these actions have informed the design of DIDs and hence represent the aspirations of the eventual system based on DIDs. > Agreed. I like the section on DID Actions very much, though I do have a few suggestions to clarify some of them. I'll see if I can get that to you before RWOT. > -- > Kim Hamilton Duffy CTO & Principal Architect Learning Machine Co-chair W3C Credentials Community Group kim@learningmachine.com > > -- > Kim Hamilton Duffy CTO & Principal Architect Learning Machine Co-chair W3C Credentials Community Group kim@learningmachine.com > >
Received on Tuesday, 19 February 2019 09:55:28 UTC