Remove broken links

[OR13]

Relies on issuer defined terms, which is the simplest way to define terms, and does not require the implementer to download several separate JSON-LD context files to make use of the examples in the spec.

• https://www.w3.org/ns/credentials/examples/v2

^ this link is 404, and causes the examples in the spec to be invalid.

---

Preview | Diff

[msporny]

Objection to the PR, it would make just about every example in the spec use the "undefined terms" property, which would set a bad precedent.

Alternatives that could achieve consensus include:

1. Create a "Getting Started" section that outlines how to get started by using the base context, and then layer semantic meaning after the developer has stabilized the terms they want to use.
2. Use JSON-LD Contexts that are being used in pilot and production settings.
3. Upgrade the base context to include terms like "name" and "website", or other generally useful properties for a credential subject, which could then be used in examples.
4. Use JSON-LD Contexts from the VC Specs Directory.

The options above are being discussed in issue #1071 and a PR here #1064 (comment).

It is highly unlikely that there will be consensus to downgrade almost every example in the specification to, what is effectively, a bad practice in general.

[OR13]

Objection to the PR, it would make just about every example in the spec use the "undefined terms" property, which would set a bad precedent.

Actually this fixes that problem (which exists in the spec today), it defines the terms using the default vocab.

The current examples are malformed, and this PR is the simplest possible fix to them (since if defines their terms using the default context).

Alternatives that could achieve consensus include:

1. Create a "Getting Started" section that outlines how to get started by using the base context, and then layer semantic meaning after the developer has stabilized the terms they want to use.

This does not fix the problem that the examples in the spec are malformed currently.

Happy to address this in a separate PR.

2. Use JSON-LD Contexts that are being used in pilot and production settings.

Introduces further bias, and makes the examples in the spec more complicated.

3. Upgrade the base context to include terms like "name" and "website", or other generally useful properties for a credential subject, which could then be used in examples.

This seems like maybe a path forward, make the examples simpler, so you don't need context complexity in them.

You can also use the issuer defined terms to achieve this, that is exactly what the current PR does.

4. Use JSON-LD Contexts from the VC Specs Directory.

Introduces further bias, and makes the examples in the spec more complicated.

[selfissued]

I agree that all the hyperlinks in our specification should be valid.

[OR13]

@Sakurann @brentzundel I would like some call time to discuss this and the JSON Schema PR:

• update json schema to accurately reflect context defn #1064

[OR13]

I am fine with the working group adding links that have consensus back to the examples, we can keep having that conversation here:

#1071

It seems better to remove broken links first, and then add non broken links back.

The links that are removed are not required and the examples are non normative.

[TallTed]

It seems better to remove broken links first, and then add non broken links back.

It seems to me harder to discover where those broken links were once they're removed, so I'd prefer to keep each (easy to locate) broken link until it is corrected, unless/until there's a determination that that link is not going to be corrected, for whatever reason.

[Sakurann]

It seems to me harder to discover where those broken links were once they're removed,

why does the WG need to discover old broken links? the one who is interested in fixing it will know where they are

[OR13]

I'm closing this PR in favor of https://github.com/w3c/vc-data-model/pull/1111/files

[iherman]

The issue was discussed in a meeting on 2023-05-03

• no resolutions were taken

View the transcript

3.4. (pr vc-data-model#1066)

See github pull request vc-data-model#1066.

Brent Zundel: next up is PR 1066.
… request for changes from manu but not here.

Orie Steele: goals.
… technical recommendation contains malformed JSON-LD.
… PR intends to remove 404.
… manu and dave and potentially others have a model of what example they'd like to see examples in the core spec.
… examples that contain multiple contexts.
… but i don't like examples with multiple contexts.
… just having some conversation on different opinions on examples.

Dave Longley: just quick clarification.
… we should have examples of both.
… first examples should be issuer independent vocabs.

Orie Steele: I am also in favor of examples of both, as I stated.

Dave Longley: and then issuer dependent vocabs but we should also say that this could be a risk.

Orie Steele: Not sure I agree regarding the "market failures" statement.

Orie Steele: but i look forward to reviewing text that elaborates on that.

Dave Longley: we should fix context instead of deleting them from examples.

Brent Zundel: two paths possible.
… one PR gets merged and then issue gets created.

Orie Steele: There are already open issues that track handling examples.

Brent Zundel: agree that the links in the examples should work.
… the other option is to create another PR that fixes the broken links.

Dave Longley: .
… another PR somewhere.
… to define an examples v2 context.
… that has examples vocab in it.
… not looked into it recently.
… but if gets traction.
… then we can merge that.
… because fixes broken links.
… and then can iterate on context.

See github pull request vc-data-model#1110.

Dave Longley: to improve examples.

Orie Steele: ^ that is the PR he is refering too.

Brent Zundel: orie?

Orie Steele: haven't had a chance to review it.

[brentzundel]

I'm closing this PR in favor of https://github.com/w3c/vc-data-model/pull/1111/files

did you mean #1110 ?