Revise language around dependent claims

[msporny]

Some people are getting confused by the language in Section 9.6 around atomic claims:

https://w3c.github.io/vc-data-model/#bundling-dependent-claims

The confusion seems to be around us suggesting that atomic claims are a best practice, which they are no longer since we have standardized selective disclosure and unlinkable disclosure schemes.

[msporny]

A PR should be raised that notes that selective/unlinkable disclosure schemes should be used and dependent claims should not be allowed to be disclosed independently of one another.

[iherman]

This was discussed during the vcwg meeting on 11 February 2026.

View the transcript

w3c/vc-data-model#1617

Brent: Manu raised this in Oct. haven't yet discussed it.

Manu: way back in another universe thought the best thing to do would be issuing one credential claim. A driver's license would be split into 35 different credentials, for example. Now we have selective disclosure so atomic credentials may not be the best approach.

Manu: if you want to selectively disclose use it not individual credentials

<manu> I'd be fine w/ removing the section, too.

Ivan: is that a normative section?

Brent: no it's not

Dmitri: hasn't come in a while - is a VC implementation Guide a responsibility of this group.

Dmitri: minimize the amount of claims, kitchen sink credentials done with selective disclosure - belongs in implementation Guide

Dmitri: Have had the conversation about composite verifiable credentials. I.e., credentials which contain credentials.

Dmitri: CLRv2 from 1Edtech - expresses school transcript as a VC. It is compound as a credential for each class is then wrapped by an outer credential

Dmitri: technically this is correct vis-a-vis VCDM but awkward.

Dmitri: should instead of representing compound credentials in compound form use id digest hash to link to them

Dmitri: do we want to repurpose dependent claims to address this concept?>

<dmitriz> agreed, we don't want to make removi8ng the section dependent

<dmitriz> let's remove

<manu> +1 to that ^

Brent: if anyone is volunteering to become an editor of the implementation guide (not revised since 2019)? Bundling dependent claims fits well into the revised implementation guide. But doesn't want to close this issue dependent on revising the implementation guide.

<dmitriz> so, +1 from me to removing this current section & closing this issue

<dlongley> +1 to remove the section and let the future take whatever direction wrt. linked VCs

<dmitriz> I can do that.

Brent: will open an issue in the implementation guide repo addressing this issue, link to it and remove the section PR.

Brent: Dmitri has volunteered

Brent: would raise the issue in the implementation guide or Dmitri would you like to?

<brent> w3c/vc-imp-guide

Dmitri: willing to take this on. Close this one after PR re: implementation Guide

---

[msporny]

PR #1623 has been raised to address this issue. This issue will be closed once PR #1623 has been merged.