On Sat, Nov 9, 2024 at 7:52 PM Vinnie Falco via Boost
On Wed, Oct 30, 2024 at 1:04 AM Thomas Fowlery via Boost < boost@lists.boost.org> wrote:
I worry about the integrity of the Boost review process. The review manager and all the accepting reviewers were/are affiliated with the C++ Alliance.
Perhaps you can help me understand how Alliance-affiliated participants may negatively impact the integrity of the review process?
Hi all, About the conflict of interest, maybe the initial remark stems from the fact that there are different expectations about what is a conflict of interest and what role (Review Manager or Reviewer) they affect. Different expectations from a text come from different interpretations of the text, which boils down to the text not being explicit enough. From my perspective, it means: *let's fix the guidelines*. We can do so by giving them a clear *Conflict of Interest* section, for both the *Review Manager *and the *Reviewer* roles. I am personally used to see this kind of clause of non-conflict of interest: 1. I will disclose any personal, professional, or financial relationships with authors, contributors, or stakeholders that could influence my judgment or create a perception of bias, 2. I will not participate in the review process for submissions in which I have contributed as an author or co-author. 3. I will not participate in the review process for submissions in which there exists any form of financial interest or contractual obligation that may compromise my objectivity. 4. I will not participate in the review process for submissions in which the authors are from the same institution, organization, or research group I am currently affiliated with. In this state, it can not be applied to Boost. From what I have seen on the ML there seems to be a confusion between clauses used by the outside world, the clauses formalized by Boost, the way people have done things, potential conflict of interests and perceived conflicts of interests. Let's try to disentangle this. Point 1 is explicitly mentioned in the Review Manager section: In order to avoid any conflicts of interest a potential review manager is expected to disclose to the Boost Community if they have any relationship to the author of the library or the library itself. -> IMO it is reasonable to require it for both Review Manager *and* Reviewers. Point 2 figures in the guidelines (in a diluted form it seems). -> IMO it is reasonable to explicitly require it for both Review Manager *and* Reviewers Point 3 does not figure in the guidelines. -> IMO it could maybe carefully/reasonably be included after careful phrasing, for example to discard the case where an employee of an organization is the Review Manager of their boss's proposal, or N employees weigh in the Review process of their boss's proposal, Point 4 does not figure in the guidelines. -> IMO it could be reasonable for a Review Manager, but can not reasonably figure in the guidelines for a Reviewer, as it would unfairly disadvantage organizations that are deeply involved with Boost. I mean, we all think about the C++ Alliance here, but if Boost loses the reviewing expertise of all its members because one of them happens to write a library that may hinder Boost end quality in the end. That being said, it must be weighted by the total number of external Reviewers. With all of these confusions made explicit, I believe they could be solved by adding something around the following in the Guidelines, and, to ease their onboarding, as checkboxes during the Review Manager and Reviewer submission process. *Conflict of Interest Guidelines:* *Clause for both roles: **All participants in the review process, including Review Managers and Reviewers, must disclose any personal, professional, or financial relationships with authors, contributors, or stakeholders that could influence their judgment or create a perception of bias.* *Clause for Review Manager:* *A Review Manager must not manage the review process for submissions in which (i) **They have contributed as an author or co-author (ii) **They have a financial interest or contractual obligation tied to the submission, (iii) **They are in a hierarchical relationship with one of the authors (e.g., an employee reviewing their manager’s submission).* *Clause for Reviewer:* *A Reviewer must not participate in reviewing submissions in which (i) **They have contributed as an author or co-author, (ii) **They have a direct financial interest or contractual obligation tied to the submission. * *Reviewers affiliated with the same organization as the author may participate, provided they disclose this relationship and their input is balanced by independent reviewers and reported at the discretion of the Review Manager.*This could ensure that organizations deeply involved with Boost, such as the C++ Alliance, can still contribute their expertise while preventing reviews dominated by internal bias. I hope this was useful for future rework of the guidelines, Best wishes and rainbow kitties, Arno