On 7 July 2019, Florian sent comments on the proposed Evergreen Patent Policy in Feedback on the proposed Evergreen Patent Policy and Approach to an Everblue Patent / IPR policy. The goal of these messages was to a) fix some problems inherent to the current drafting of the Evergreen Patent Policy and b) make some largely cosmetic changes (renaming terms to be more generic), and import a few missing sections from the currently-active W3C Patent Policy, to make the proposed Evergreen Patent Policy something that could be used for the W3C Process as a whole, as per the recommendation of the AB taskforce working on Evergreen and related topics.
To better characterize the suggested changes and to provide concrete proposals for incorporation into the proposal, we imported PSIG’s DRAFT Evergreen Patent Policy (as it was on July 24th) into a GitHub repository, so as to rigorously track proposed changes. We first, however, did some cleanup of the source code and markup to bring it in line with the latest W3C document-editing best practices.
The rest of this document explains the problems outlined in the two emails (along with a few new ones discovered along the way), and their proposed solutions. We hope that PSIG reviews these changes and updates their draft to reflect them in spirit, if not in letter. We also hope that it finds its way into a git repository so that the Advisory Board, PSIG, and the W3C Process CG can better coordinate as the draft evolves (and we are happy to hand over our git repository as a starting point for such an effort).
Some best practices we adopted as the basis for editing this document, which do not have any effect on the actual prose, but make it easier to maintain:
name attributes in favor of id, etc.
p.s. PSIG might want to consider using Bikeshed to generate its drafts, as it handles cross-references, section numbering, and table-of-contents generation automatically. We are happy to make the conversion if PSIG so desires.
Merged text in 2.10.1 and 2.10 to produce a combined definition for “Working Group Participant”. (Changeset, Result, Diff)1. Both "participant" and "evergreen participant" are defined, but only the later is used. I suggest merging. Also, suggested name: Working Group Participant.
2. Circular definitions in 5.1, 5.2, and 5.3: the words "Contribution Licensing Obligations" in 5.2 (resp "Review Draft Licensing Obligations" in 5.3) link to section 5.2 (resp 5.3), which does not contain a definition of these words. The same words in 5.1 also link to 5.2 (resp 5.3), but do look like a definition. These words in 5.1 should be marked up as the definition, and the link in 5.2 (resp 5.3) should be changed to point to that definition.
Fixed as described. (Changeset; no change to visible prose)
3. The 5.4 and 5.4.1 sections both ask the question of how often Review Drafts should be published, and provide an answer (6 to 24 months). However but the Evergreen Process document also does (https://w3c.github.io/w3process/evergreen/#evergreen-rec-snapshot) Redundant text is bad, as it can easily lead to contradictions. For example, the Process version says "24 months **if** there are significant changes" (I think it means substantive rather than significant), but the Evergreen Patent Policy has 24 months unconditionally. No strong opinion if that single place should be the Process of the Patent Policy.
Split 5.4 to have several subsections: one to generically define what a Patent Review draft is, one to define its exclusion period, one to define its publication cadence, and one to define its notification requirements. Folded in annotated thoughts from the PSIG draft into potential normative text. Improved practicality of the cadence requirements, added an issue that maybe this should be moved to the Process document which is easier to update. (Changeset, Result, Diff)
4. "Essential Contribution Claims" seems to have broken phrasing that undermines what it is trying to do. compare (emphasis mine): > "Essential Review Draft Claims" [...] means all claims [...] that would necessarily be infringed by **implementation of** the Review Draft. with > "Essential Contribution Claims" [...] means all claims [...] that would necessarily be infringed by the Contribution, or by combination of the Contribution with the Evergreen Specification [...] A contribution is a "proposal to make changes to a [...] Specification". It cannot infringe anything. Only **implementations of** the contribution can infringe on patents, but that word is missing from the definition of "Essential Contribution Claims", which by my reading, risks making the whole thing nonsensical. It should instead say: > "Essential Contribution Claims" [...] means all claims [...] that would necessarily be infringed by **implementations of** the Contribution, or by **implementations of** the combination of the Contribution with the Evergreen Specification [...] (this "bug" is shared with the WHATWG IPR Policy)
Fixed as described. (Changeset, Result, Diff)
Fixed by replacing:7. In "2.1. Contribution", "Contribution" is defined as a "proposal to make changes to a particular Evergreen Specification". "Evergreen Specification" is defined not to include "An editor's draft that has never received an indication of Working Group consensus". However, Editor's Drafts are the thing about which WG member make "proposals to make changes". Also, in the Evergreen Process, it is said that "Preliminary Drafts do not necessarily represent a consensus of the Working Group.". So, if a bunch of contributions are made to an new ED, which is then published as an PD (and maybe iterated further upon as ED/PD), which is then published as an Evergreen REC, it could be said that there has never been an indication of Working Group consensus, and therefore that there is no Contribution Licensing Obligations on any of the content (and it may be a long time before there is a Review Draft). That seems bad. I don't think EDs should be excluded, even with a qualifier as currently written. If the ED is being worked on by the working group, I think it is fair to incur licensing obligations when making a contribution, regardless of whether there is yet consensus on the whole document or not. And if the ED started its life outside the WG, the point at which it is brought to the WG is the contribution. Moreover, WGs that do internal incubation generally use ED only for a while, and use the first TR publication as a sign of maturity and of graduating from incubation. We should not put incubating WGs at a disadvantage to incubating CGs.
An editor's draft that has never received an indication of Working Group consensus is not an W3C Specificationby:
A document that Working Group does not have consensus to work on is not an W3C Specification; Working Groups must clearly identify documents that they have consensus to work on.We feel this is an improvement because:
(Changeset, Result: terminology updated throughout the document)
1.1. Call it something non evergreen specific (i.e. "W3C Patent Policy" or "W3C IPR Policy").
Renamed to “W3C Patent Policy”.
1.2. Change all references to "Evergreen Recommendations" / "Evergreen Specifications" into "W3C Technical Reports with a maturity other than Note" (this phrasing covers both REC track documents and Evergreen ones). This is a bit of a mouthful, so we could define a local term ("W3C Specification"?) to avoid repetition.Renamed to “W3C Specifications”, and updated the definition to reference work produced under charters referencing this policy. (Changeset, Result)
1.3. Change all references to "Evergreen Participants" into "Working Group Participants"
Renamed to “Working Group Participants”.
3. Change all references to "Last Call" in the Process into "Review Draft" (or possibly "IPR Review Draft", and use that term in the IPR Policy as well). This would cover: (1) CRs, (2) RECs+Call for Review of Proposed Corrections, (3) RECs + Call for Review of Proposed Additions. (Note to Don Deutsch: if you're not familiar with (2) and (3), that's part of the so-called Everblue Proposal https://www.w3.org/wiki/Everwhat_dashboard#EverBlue)
To make these less confusing to reference in the Process, which has multiple sources of review (such as wide review, horizontal review, AC review...), we renamed “Review Draft” to “Patent Review Draft” so that references to it in the Process can be clear about their intent.
The Evergreen process can say “Patent Review Draft” where it used to say “Review Draft”, and the REC Track can say “Patent Review Draft” where it used to say “Last Call Working Draft” (Changeset in the Process, Result, Diff).
These changes fill in gaps in the proposal with respect to the current W3C Patent Policy.
Copied over Note on Licensing Commitments for Invited Experts, which is referenced by the Process and seems like a useful clarification.
2. Restore the section about member submissions, which isn't part of the evergreen/whatwg IPR policy
Copied over Licensing Commitments in W3C Submissions and adapted it to use updated terminology such as “Essential Contribution Claim”. (Changeset, Result, Diff from 2017 Patent Policy)
9. The classic W3C Patent Policy has specific rules for FPWD (and subsequent WDs prior to "Last Call"). The WHATWG has no equivalent to FPWD (or to WDs), and therefore no specific clause in their IPR Policy, but they do have a requirement that a Review Draft be published within 6 months of the Workstream being formed. Evergreen has "Preliminary Drafts", which are equivalent to WDs (and of which there is a first one). Even if an Evergreen REC is published without Preliminary Drafts first, it may be 24 months from the First Everegreen REC before the first Review Draft, during which there's no clarity on the Patent situation. The Contribution Licensing Obligations provides some measure of comfort, but there could still be WG participants who did not contribute but have claims they will eventually exclude, but are keeping initially silent about. The concern is stronger if incubation of that specification happened outside of the Working Group. An Evergreen Patent policy should either: * restore the relevant clauses from the classic W3C Patent Policy into the Evergreen Patent Policy (pros: reusing the existing rules probably results in easier acceptance by the AC, especially given that we want to apply the improvements of the evergreen/whatwg policy to the REC track as well; cons: the rules are more complex) * declare the first publication of an Evergreen REC (or even possibly of the first Preliminary Draft?) to be a Review Draft (pros: simple, maximum coverage; cons: more limited exclusion opportunity than in the classic Patent Policy) * Require a Review Draft within 6 months an evergreen deliverable being included in a charter (pro: same as WHATWG; cons: more complicated than First-Draft=Review-Draft, 6 months without clarity)
We evaluated the handling of First Public Working Drafts, as well as joining and leaving a Working Group, in the current W3C Patent Policy and came up with the following changes as a reasonable incorporation of their intent:
8. The classic W3C Patent Policy has a 60 day exclusion period (most of the time, there are also 150 and 90 periods associated with FPWD). The proposed evergreen one changes that to 45 days (except when leaving a group, where it is 60). Is this difference intentional / necessary? I worry it will generate needless controversy.
We noticed that the PSIG draft added this as an open issue in their draft; we left it as an open issue. We recommend aligning with the classic W3C Patent Policy to avoid getting push back on the whole document merely due to a difference of duration.
Note: We did not address issue 6 of the Feedback on the proposed Evergreen Patent Policy mail here, but instead filed it against the process, as it felt orthogonal to this discussion.
If we want to have a Trademark Policy, it should probably be written from scratch, rather than copied from WHATWG. Since we do have sections on our web site covering this already, it is not clear that we need this section in the Patent Policy at all, and maybe it should be removed.