Authorship standards for AGPs

Continuing the discussion from AGP discussion: Community Review Period, since we veered away from discussing the AGP itself to discussing the concept of authorship and contribution attribution.

It’s worth explaining here (and perhaps in AGP-1 itself, which would require a new Meta AGP) why an author is required to be listed for an AGP in the first place. Put simply, an author is required so that people know who to direct questions about the AGP to. It’s not necessarily for attribution; although people who help contribute to an AGP may want some kind of attribution, ghost authorship or a “contributors” section of an AGP are perfectly ok too. But there should be at least one person listed as an author who can answer questions and respond to/ incorporate the feedback of others. Usually this would be the person who opens the PR in the AGPs repo.

On the topic of collaboration process: there is no specific “process” for AGP collaboration, other than what is mentioned in AGP-1 which is to submit a PR and discuss feedback / objections in the PR thread, a related GitHub issue, or a discussion somewhere else like here on the forum. AGP authors can use whatever process they want to include other people in the authoring of a proposal.

2 Likes

I guess it’s more of having respect. If someone proposes a well thought out change to the governance, I would expect if someone is going to create an AGP some kind of collaboration is due / should be offered.

I think the author location is the best place to attribute people who contributed their intellect and ideas to the proposal. I disagree that “author” should be the person who submitted a PR or is available to answer questions (only one person can submit a PR, and how do you know they wouldn’t be able to help answer questions or at least collectively make the decision on who should take point.) I’ve seen multiple EIPs etc with multiple authors listed.

2 Likes

I did say that there should be at least one person listed as an author. Indeed, we have already several AGPs with multiple authors listed. I also said that if it is one person it is usually the person who opened the PR, again there has been at least one AGP that comes to mind where the author listed was not the same person who opened the PR.

To disambiguate “who should be given credit for writing the AGP” from “who should questions about a proposal be directed to” perhaps we could change the nomenclature from “author” to “champion” as the ZIP process does and add a section to the templates for “Contributions and Prior Work” that authors/champions could add attribution info to.

2 Likes

Technically I started Stage II and did not do the work to take it to Stage III, nor did you bump the thread and mention you were interested in taking it to Stage III. Instead, you decided to bypass your part of Stage II and go to Stage III instead, yet linked my Stage II “work”.

I don’t know, I think this use case may be a bit different. I don’t think it makes sense for someone to decide to take an idea to Stage III before informing the person who did Stage II.

I’m not angry or anything, and am not demanding a change to how you have attributed authorship. I would just propose these standards in the future.

I like the champion idea, but it should perhaps be documented like this:

author: Alice, Bob (champion)

But I also think the person who started Stage II should also be who decides “championship” / given the opportunity to be the champion.

It is a new edge case where someone started the discussion with the AGP idea but didn’t follow up on it, then someone else decided to take the initiative to actually turn the idea into a formal proposal. I don’t think we should restrict anyone from taking such initiative.

Speaking as an editor I’m not going to go look and see who wrote what first and when except in cases of actual plagiarism where there’s a dispute because someone tries to pass off someone else’s proposal as their own (this isn’t discussed in AGP-1, this is just how I would try to resolve this kind of dispute as an editor).

In the case of the Aragon Network Fiscal Year proposal it certainly wasn’t my intent to plagiarize and I did reference your thread in the first post on the discussion thread about this proposal. Looking at this as an editor I wouldn’t call this plagiarism but then I am biased so if you do feel uncomfortable about how this AGP is presented I would have to let another editor (@lkngtn) decide how to deal with it (and would prefer to move that discussion either back to the discussion thread for that AGP or start a new thread about it).

Hey John! It’s all good. Like I said I’m not angry and I’m not trying to change anything that happened.

I was trying to understand authorship standards for cases like this, edge case or not, as that is the thread topic.

I’m not going to push any further on this and I apologize if it seemed that what I was bringing up was hostile or off topic - I could have probably worded what I was trying to communicate better. I am glad that the AGP has been created - thanks for taking the initiative.

I’m just going to drop this here:


From the article:

How can we be good stewards of collaborative trust?
TL;DR: This essay makes a lot of suggestions, but the most useful/non-obvious/actionable are likely: (1) Be generous. (2) Use author contribution statements. (3) Put “author order not finalized” if it hasn’t been.

A lot of the best research in machine learning comes from collaborations. In fact, many of the most significant papers in the last few years (TensorFlow, AlphaGo, etc) come from collaborations of 20+ people. These collaborations are made possible by goodwill and trust between researchers.

This goodwill and trust is a precious shared resource, and it can be a fragile thing. When people work together, it’s easy to have conflict, especially around attribution and credit. If dealt with poorly, attribution issues can fester. I’m aware of several cases of collaborations dieing, or people leaving teams and organizations, where and the underlying issue was hurt feelings and lost trust around collaboration. This strikes me as rather sad.

We often talk about credit issues in kind of binary terms. But if the thing we care about is this shared trust, I think it’s not enough to just avoid doing anything wrong. We must also avoid any feeling or appearance of unfairness. In fact, we’d ideally actively cultivate the opposite, to behave in ways that add back to the pool of goodwill.

We should also be mindful that credit issues can easily be perceived as, and likely often are, linked to privilege or power gradients. This might be gender or race, but it can also be things like being a remote collaborator (ie. geographically removed from others), being an engineer or designer instead of a researcher, or being at a lower level professionally. A perception that junior collaborators or those from under-represented minorities are taken advantage not only harms the research community, but the larger cause of making sure all humans are treated fairly.

Core Principles

  • Always check in with any person who could plausibly be an author or feel like they should be, even if you disagree . Never have authorship or authorship order be decided behind closed doors or without giving people an opportunity to advocate for themselves.
  • Err on the side of sharing credit. Credit isn’t zero sum. It is often in everyone’s benefit to be generous with credit, because it creates an incentive for others to help in the future. It also makes sense to be risk-averse to the possibility of not crediting someone who deserves it, because the harm of not crediting someone who deserves it is often greater than the harm of crediting someone who doesn’t deserve it.
  • Acknowledge anyone you can remember talking to about your research significantly. It costs nothing and builds good will. You can still use stronger language to highlight people who helped you more.
  • Avoid diffusion of responsibility . For example, have someone clearly responsible for checking in with everyone on authorship.
  • Don’t reveal someone else’s unpublished work or merge it into your own without their consent.
  • Remember that you are likely overestimating your own contributions.
  • Act in ways that will make people want to work with you. Enthusiastic collaborators are one of the most precious thing you can have as a researcher.
  • There’s no substitute for emotional labor. Humans have feelings. No magic bullet will remove the need for us to invest energy understanding them and talking them through together.

The article goes on to explore common problems and common solutions around trust and collaboration. The author is one of the leading researchers in the machine learning space. It’s a 5min read and I highly recommend giving it a shot.

3 Likes

I couldn’t agree with this more

More importantly, collaboration isn’t a zero-sum game this is what makes being a member of this community so attractive, rewarding and enriching

1 Like

I agree with the “champion” or “chair” nomenclature here, as it is common in other community-driven processes, and in particular we may want to consider requiring at least one to be nominated / designated before moving a proposal to Stage III. There is an important distinction between someone who simply authors a pre-proposal (Stage II) and one who wants to take responsibility over the course of its life as an actual proposal draft (Stage III).

The current process has this responsibility implicitly assigned: we assume the author will take responsibility. We can still make this implicit assumption, but perhaps define standards to mark certain pre-proposals as “abandoned” and up for championing without explicit permission from previous authors / champions (e.g. was not actively championed in a previous ANV period).

1 Like