W3C Council Report on the Formal Objection Against the RDF-star Working Group Charter

Council Report,

Copyright © 2025 World Wide Web Consortium. W3C® liability, trademark and permissive document license rules apply.

1. Introduction

A detailed exposition of this case may be found in the report prepared by the W3C Team, and the elaboration on the objection, by the objector.

The initially-proposed RDF-star charter included

For every recommendation updated by this Working Group, the pending editorial errata will also be addressed.
...
Other normative deliverables, or non-editorial errata to the deliverable above, might be considered by the Working Group during its lifetime.

2 comments in the AC poll suggested “extending the scope to all kinds of errata, including the substantive ones”, without formally objecting.

The charter was revised to incorporate that change, updating the wording to:

For every recommendation updated by this Working Group, the pending errata will also be addressed.
...
Other normative deliverables might be considered by the Working Group during its lifetime.

A W3C member objected to that change on the grounds that:

  1. The group does not have time to complete its primary objectives and also to address non-editorial errata.

  2. The change opens the door to additional features, which won’t comply with the W3C Process’s requirements for revising a Recommendation.

  3. The new charter encourages piecemeal instead of coherent feature additions.

There is also a discussion on SPARQL EXISTS (issue; email; discussion) that seems to be underlying this objection to the charter. However, when the Team Contact suggested formally objecting to just that Group Decision, the objector did not think that limited objection would be enough to get the group to finish its deliverables on the charter’s schedule.

The Team did not propose to return to the originally-proposed charter because the intent had never been to exclude non-editorial errata from the scope. Further, as Florian Rivoal argued, Working Groups are expected to address at least some of the long standing errata on each republication and may not simply ignore the non-editorial ones.

A Council was formed to rule on this objection. This Council Report documents the conclusions of this Council.

2. Decision

The Council resolved to accept the Team’s recommendation to overrule the objection.

The charter may proceed.

3. Rationale

We are overruling the objection because there seems to be broad support within the Working Group for addressing their substantive errata, and because it’s important for working groups to address substantive errata while maintaining their specifications.

The detailed objection makes three arguments that we’ll address individually:

3.1. Not enough time

The objector argues convincingly that the group cannot close all of its open issues before it is scheduled to publish new versions of its deliverables. However, as the Team report says, the scope allows the errata to be addressed after the new Recommendations are published (“The group will then maintain the set of resulting recommendations”).

The charter does restrict the group to avoid scheduling errata fixes in ways that would delay the deliverables past their due dates. However, even if the deliverables miss their due dates, as often happens, there may be some errata fixes that can be applied without further delaying the deliverables. We don’t need to change the charter to delay all errata fixes in order to get the deliverables as soon as possible.

3.2. New substantive features must go through formal review

The objector argues that section 6.3.11.4 of the W3C Process, about adding new features to Recommendations in-place, constrains the changes that the WG wants to make to their documents.

First, as discussed in the next section, the charter does not allow the Working Group to add features outside of RDF-star itself. This formal objection does not target RDF-star, so section 6.3.11.4 of the W3C Process doesn’t apply to this formal objection.

Second, this section of the Process says

To make changes which introduce a new feature to a Recommendation that does not allow new features, W3C must create a new technical report ...

The documents that the Working Group is changing are new technical reports. For example, the WG is not updating rdf11-concepts in-place. It has instead started work on rdf12-concepts, in accordance with this section of the Process. (However, the charter copies wording from the 2022 charter that implies a different plan, which would have violated the Process, and we recommend fixing that below.)

Third, focusing on the principle at play rather than the specific cited section of the Process, all of this WG’s changes will be formally reviewed as the new technical reports advance through the Recommendation track, without needing any changes to the charter.

3.3. Errata fixes should not introduce patchwork features

The objector argues that “addressing” an erratum by adding a targeted feature could produce a worse final language than could be achieved by considering the whole space of feature requests and fixing them with coherent language changes. This is correct. However, the W3C Process defines an erratum as

any error that can be resolved by one or more changes in classes 1-3 of section §6.2.3 Classes of Changes.

That is, the revised charter does not allow the Working Group to add features in order to fix errors. Errors that can only be fixed by the addition of new features will require the group to recharter, at which point a coherent set of features can be added.

4. Recommendations

We would like to remind the AC that it’s possible to formally object to any decision, not just to formal polls via the WBS system. The W3C Council process is able to make better decisions if formal objections are about more specific decisions.

Not everyone in the W3C community is aware that the Process gives a precise definition for “errata”, so we recommend that the charter be editorially changed to link the use in the scope section to its definition in the Process.

We also encourage the objector and the WG chairs to work together to find a mutually agreeable schedule to accomplish the group’s chartered work.

Finally, in reviewing how section 6.3.11.4 of the W3C Process applies to this charter, we noticed that the charter’s use of "Adopted Draft" implies that the WG intends to continue work on the RDF 1.1 Recommendations by renaming them to "RDF 1.2". Section 6.3.11.4 would prohibit doing this, since the RDF 1.1 Recommendations are not identified as allowing new features. This mistake was copied from the 2022 RDF-star WG charter. The WG did not actually intend to rename the RDF 1.1 series and has in fact started on new technical reports as the Process requires, so there is no problem in practice, just with the wording of the charter. This problem is not the focus of this formal objection, so it’s not a reason for us to uphold this objection, but it should still be fixed. We recommend that the charter template be updated with clear wording for charters to use when they intend to use the text of a published Recommendation as the initial content of a new technical report, and that all active and draft charters that use "Adopted Draft" for this purpose be updated with the new wording.

5. Miscellanea

Appendix A: Council Participation

The Council was formed on February 19, 2025 from the members of the TAG and the AB who had also been members on December 11, 2024 to be included in the call for dismissal reasons, plus the CEO. No-one renounced their seats on the Council; No-one was dismissed. Therefore, the actual membership of this council was:

Jeffrey Yasskin was appointed as chair.

Of those qualified to serve, the following participated in the final decision:

The decision was made by consensus.