Clarified subtitle of Data Model #780
Conversation
|
Thinking out loud, feel free to ignore: The only remaining concern I have is with the word "provenance" -- wondering how many people are familiar with the word, but I have no better word other than "source". Trying to aim for the spirit of this phrase (best said with a US Southern accent): "Know where your data is coming from, ya'll." |
|
I think provenance is understood well enough. It's more than source/origin, and I think that's important to communicate, which would make the subtitle much longer with any other wording. We could make it link to a dictionary definition, the PROV DBpedia entity, the PROV Wikipedia page, or even the PROV-O Primer. |
|
The issue was discussed in a meeting on 2021-08-11
View the transcript3.6. Clarified subtitle of Data Model (pr vc-data-model#780)See github pull request #780. Ted Thibodeau Jr.: fairly self-explanatory, good discussion.
Wayne Chang: once the 14-day period is up we can merge |
|
I don't like using the word "provenance" in the subtitle; that has a specific meaning about deep sourcing that I don't think VCs fulfill. We're not describing where data comes from, or proving where it comes from, except in the most proximate sense -- it comes from the issuer. But where the issuer got it (what "provenance" suggests) is out of scope. I suggest we reword to "with verifiable source." Tagging @SmithSamuelM, who may also have an opinion. To be clear, that's a -1 from me on merging this. |
|
I hear you, @dhh1128. Upon further consideration, I think you're right; "verifiable source" is better. Now editing the PR.... |
index.html
Outdated
| @@ -19,7 +19,7 @@ | |||
| shortName: "vc-data-model", | |||
|
|
|||
| // subtitle for the spec | |||
| subtitle: "Expressing verifiable information on the Web", | |||
| subtitle: "Expressing information with verifiable provenance", | |||
There was a problem hiding this comment.
Suggesting this on behalf of @dhh1128. Can this achieve consensus?
| subtitle: "Expressing information with verifiable provenance", | |
| subtitle: "Expressing information with verifiable source", |
There was a problem hiding this comment.
Ooof, this evokes a "What does that mean?" response from me... that's why I suggested it as a worse alternative above. I don't see how it is any better than provenance...
index.html
Outdated
| @@ -19,7 +19,7 @@ | |||
| shortName: "vc-data-model", | |||
|
|
|||
| // subtitle for the spec | |||
| subtitle: "Expressing verifiable information on the Web", | |||
| subtitle: "Expressing information with verifiable provenance", | |||
There was a problem hiding this comment.
Another option:
| subtitle: "Expressing information with verifiable provenance", | |
| subtitle: "Expressing information with verifiable authorship", |
There was a problem hiding this comment.
This one feels better... except that we don't talk about "authors" in the document anywhere.
There was a problem hiding this comment.
It's more than source and/or authorship ...there is an an aspect that deals with the content of the claims as well.
A VC is expressing an trustable attestation that a set of claims has not been modified relative to a point in time (issuance of the VC).
Perhaps, "Expressing a set of immutable attributes relative to a point in time".
|
I think I'm OK with either "verifiable source" or "verifiable authorship". I wish there were an easy way to run a quick poll on Github! |
|
I'm aligned with Ted now. If Manu doesn't like "source" and only tolerates "authorship", and if Michael doesn't like authoriship, here are a few other options just to stimulate thinking (but to be clear, I'm still good with either of Ted's alternatives):
Etc. |
|
+1 for
but I believe the title of the specification needs to be changed given that it includes so many disconnected concepts and lacks any serious cohesion. See #791 |
|
In an attempt at compromise wrt. "source"... what about:
I'm back-tracking on "on the Web" because we really are using Web technologies here... and the Web does flow into our daily lives, even in the physical space, so it's probably fine. I'm suggesting "Confirming" as a synonym for "verifying"... it's typically a good idea to explain the title of the specification using different words in the subtitle... I'd be fine w/ going back to "Verifying", but suggest we find a good synonym. ... and this includes the word "source" in a way that I think is understandable to the general public (which is one good measure of a good subtitle). So, @dhh1128, @TallTed, what about?
|
|
There should be no requirement/limitation to use VCs "on the Web". UPDATE: ...or leave VCs to be focused "on the Web" in favor of having a more general concept (e.g. Structured Credentials) that captures the scope of credentials used both on the Web as well as beyond the Web (e.g. #788 (comment)) I favor the latter. |
|
"Expressing verifiably authored data" suggests that the statements within the VC are verifiably authored (issued), not that the collection is... |
|
The issue was discussed in a meeting on 2021-11-10
View the transcript3.1. Clarified subtitle of Data Model (pr vc-data-model#780)See github pull request vc-data-model#780. Manu Sporny: I prefer we close this PR and try again later.
Ted Thibodeau Jr.: it has been a lot of bike shedding because it is nuanced and getting this right is difficult.
|
|
This PR is intended to address issue #105 While there has been a fair bit of discussion on this PR itself about the topic, our go to solution is to label the issue rather than the PR, so I'll be removing the Errata and Editorial labels from this PR and making sure the issue is linked. Given that issue has been closed already and this PR is open I believe the intent is for discussion to continue here until we're able to merge it. Any issues with this @brentzundel @wyc or @msporny ? |
kdenhartog
removed
editorial
|
I agree that this PR does not need an errata label. |
|
"Expressing data in verifiably assembled containers"? |
|
"linking data on the web to identifiable entities using standardized containers"? |
|
Some other ideas:
|
This one seems the most "explain like I'm 5" of them all - I'd be good with this one |
+1 |
My remaining concern with this subtitle is that it sounds more generic than we intended with the VC data model. A VC is a credential, not just any arbitrary digital document. VCs are about expressing the attributes of entities / describing subjects / making claims about subjects. This is more narrow than "expressing data". This is an important distinction that I don't think we want to lose -- and it seems to me that failing to highlight it in the past has caused confusion. |
|
|
We need to either close this issue, or rank choice vote these options and put it to the CCG... we're going around in circles. |
|
Giving a short list of options to the CCG for voting on, is one way to resolve this, providing we can agree on the short list of say half a dozen options. |
|
Another option: "Verifiably authentic containers for claims" |
Perhaps more generic than you intended, but I absolutely meant it to be more generic than you appear to be describing here. Further, to my eyes, "claims" doesn't have a clear link to the meaning I think most here would be intending, and "expressing claims" is just as vague. Once again, the reason we changed from "claims" to "credentials" was that the former suggested to people that the statements (which we intended to be the "claims" under discussion) contained in the were verifiable, not that the utterance (the container, the "credential", the collection of "claims") was verifiably sourced. As @msporny says, we are going around in the same circles, again. We might be better off with no subtitle, than with continuing to cycle through many if not all of the suggestions made to date and likely to follow. |
+1 to that. I actually think it's the best proposal -- remove the subtitle. We're trying to communicate too much with too few words -- let people read the introduction / whatever they need to. |
Co-authored-by: Manu Sporny <[email protected]>
|
It's all about the outer part of the credential being verifiable and only this part (the credential envelope). The VC "concept" says nothing about the verifiability the inner part of the credential ...i.e. the parts inside the credentialSubject (aka the data part) ...other than it (the envelope and the blob that is inside) hasn't been tampered with once signed. Watch https://youtu.be/kM30pd3w8qE and then https://youtu.be/9RLYS7Xvabc |
|
I can approve going with no subtitle |
|
The issue was discussed in a meeting on 2021-12-15
View the transcript2.1. Clarified subtitle of Data Model (pr vc-data-model#780)See github pull request vc-data-model#780. Manu Sporny: We've got one we should be able to resolve very quickly. Brent Zundel: Any objections?. Manu Sporny: That was 6 months in the making :). |
|
Editorial, multiple reviews, changes requested and made, no objections, merging. |
Preview | Diff