The Post-Meritocracy Comes to Linux

This is a developing and highly-controversial story, so I’ve got to give y’all a warning that this quick post is based on provisional research. There’s probably a lot more to the story than I know, and I’m going to try and follow it and update folks. But it definitely seems important enough to warrant this early note.

About a week ago I noticed a kind of weird story about Linus Torvalds (creator and leader of the Linux kernel project)[ref]An operating system kernel is “the core of a computer’s operating system, with complete control over everything in the system.” Wikipedia[/ref] apologizing for his infamously abrasive personality and promising to get better. What I didn’t notice at the time was that, simultaneous with his temporary departure, he also implemented a “Code of Conduct” for the Linux kernel project.[ref]You can see the community’s reaction on the GitHub repo.[/ref] Then, people’s heads started exploding. Why? Because many people believe that the Code of Conduct is basically a SJW[ref]Social Justice Warrior[/ref] take-over of Linux.

The best article I’ve found so far to explain that rationale is this one, which points out that the Code of Conduct was written by  Coraline Ada Ehmke, an LGBT activist who has a history of using her Contributor Covenant to go after people she thinks are bigots.

For example, she tried to have a core contributor to the Opal project expelled because–in an unrelated Tweet–he stated that “(trans people) not accepting reality is the problem here.” Although Ehmke wasn’t involved in the project or the conversation, she campaigned to have Elia removed. She failed in that regard, but got Opal to adopt her Contributor Covenant, and she then tried to modify Opal’s version of the Covenant in a way that would give her new ammunition to take on Elia.

The obvious threat is that activists like Ehmke will use the Linux kernel Code of Conduct to go after Linux developers the same way Ehmke went after Elia, and there are early signs that that is exactly what will happen. A “diversity and inclusion consultant” named Sage Sharp called out Linux developer Ted Tso in a Tweet:

Sharp subsequently backed off and claimed that “I don’t want Ted Tso reviewing Code of Conduct cases involving sexual assault or sexism” and that they “did not make a statement asking for him to be removed from the board.” Yeah, sure. Because it’s totally believable that Sharp is content with having a rape apologist on the board. They kicked Brandon Eich out of his own company for much less than this, and the only reason they’re not doing a full-court press on Tso (yet) is that the Linux community is both hostile and well-armed, up to and including a potential nuclear response. 

The nuclear deterrent was described in this (slightly sensational) article. Basically, some anonymous person posted to the Linux Kernel Mailing List with a suggestion that anybody kicked off the project for Code of Conduct violations could revoke permission to use their contributions to the Linux kernel. The legal case here is complex and hypothetical, but folks like Eric S. Raymond believe the threat has enough legitimacy to be taken seriously. If they can really revoke permission to use their code, then it becomes effectively illegal to run Linux, and if it becomes effectively illegal to run Linux then the Internet pretty much dies, because a vast swathe of the servers that power the Internet run the Linux operating system, to say nothing of all the Android phones and various embedded systems out there. 

I don’t think this is very likely for two reasons. First, the social justice activists don’t want to overplay their hand, and you can see that in how Sharp walked back their original claim. Second, the legality of withdrawing code from the Linux kernel is complicated and untried; it’s never been tested in court and nobody knows for sure what would happen.[ref]It’s worth pointing out that legal uncertainty is a pretty significant threat in and of itself, however.[/ref] 

The nuclear deterrent may be grabbing all the headlines, but it’s not my long-run concern. I’m more worried about a subtler problem, the infiltration of post-meritocratic thinking into software that runs so much of the hardware that we all depend on. Linux is critical to our infrastructure. It’s not just powering the Internet. Linux is also the basis of Android, for example, so it’s the foundation for a big chunk of our smartphone ecosystem. It’s also in a lot of embedded hardware from cash registers to medical devices. Even if we avoid the nuclear option, you have to ask yourself: do you want potentially life-saving hardware to be developed and maintained by a community that has rejected merit as their guiding criteria?

It’s not an exaggeration or a conspiracy theory to assert either (a) that up until now Linux and the open source community had embraced merit as their ultimate guiding principle or that (b) the Code of Conduct is the spearhead of a philosophy that emphatically and explicitly rejects that position. In the past, I would have had to cite a definition of critical race theory from the UCLA School of Public Affairs and then link critical race theory to wider social justice theory, and basically you’d have had to trust me. But Ehmke has made my job much, much easier in this regard. Not only did she write the Code of Conduct, but she also wrote the Post-Meritocracy Manifesto. The manifesto begins:

Meritocracy is a founding principle of the open source movement, and the ideal of meritocracy is perpetuated throughout our field in the way people are recruited, hired, retained, promoted, and valued.

But meritocracy has consistently shown itself to mainly benefit those with privilege, to the exclusion of underrepresented people in technology…

It is time that we as an industry abandon the notion that merit is something that can be measured, can be pursued on equal terms by every individual, and can ever be distributed fairly.

The manifesto goes on to embrace ominous values such as the belief that “working in our field is a privilege, not a right,” a clear statement of intent to “no platform” all those who fail to affirm the dogma of the social justice activists. It’s just that in this case we’re not only preventing the undesirables from spreading their opinions, but also from contributing code to open-source projects.[ref]Worth mentioning: this is also a threat to their livelihoods, in many cases. Which is certainly a feature and not a bug as far as social justice activists are concerned.[/ref]

Obviously there are political dimensions here. Ehmke has stated that the Contributor Covenent–the basis for the Code of Conduct–is a political document.

The politics are problematic, in my view, to say the least. But while the politics can be complicated and ambiguous, the practical considerations seem a lot less murky. Let me ask you this: how comfortable would you be if your airline pilot or cardiac surgeon was trained by a post-meritocratic institution?

Don’t get me wrong: obviously software developers don’t have an immediate, life-or-death impact the way airline pilots or surgeons do. But that just means the degradation in quality from rejecting merit as the ultimate guiding principle will be more subtle and diffuse, but also have a correspondingly broader impact. Instead of diluting merit in the evaluation of pilots and surgeons, we’re going to dilute merit from the evaluation of software developers who write the code for air traffic control and teleoperated surgical robots. That still seems like cause for concern.

I always sort of resent having to add this caveat, because I think it should be obvious, but I emphatically support the purported goal of equality and fairness. What I don’t support is the willingness to sacrifice merit to get there, much less the assumption implied by the belief that we need to do so. I don’t think we need to reject meritocratic ideals to accept women, gays, blacks, or any other minority as coders for the simple reason that I think women, gays, blacks and all other minorities are just as capable–given equal opportunity–of contributing equally meritorious code. If they cannot due to unequal opportunity,[ref]For example, it’s hard to be a great coder if you don’t have access to computers and the Internet.[/ref] then it seems more like a betrayal than a helping hand to jettison merit as a benchmark instead of undertaking the much harder work of rectifying unequal opportunity to ensure that everyone has the same chance to develop their skills and talents. Equality and merit are not mutually exclusive. On the contrary, it is only in the context of a meritocracy that we can ultimately have equality. Not only should we be unwilling to accept anything less than a meritocracy for vitally important infrastructure, but also for a just society.

14 thoughts on “The Post-Meritocracy Comes to Linux”

  1. It seems to me that you’d be able to find substantial areas of agreement with her critique of meritocracy as practiced in the open source community. Does that seem accurate to you?

  2. My fear is that she’s doing a few things which seem excessively provocative, but that her critique may be not only basically sound but consonant with the views of many of the people she’s making enemies. It seems clear from some of what she’s writing that she doesn’t think the open source community operates on meritocracy at all; instead, she thinks it (at least often) applies a terrible measure of merit to reinforce the power of those who have it. In doing so, they’re trading off certain sorts of technical competence against other genuinely valuable skills, largely because the people making the trade off have those technical skills but not the others. As I would use the word, that’s clearly not meritocracy, but if most of the people in her community who appeal to meritocracy are doing it, I can understand why she’d use that word to designate what she’s against. Indeed, it seems to me that a lot of her arguments hinge on assembling and valuing the team which will produce the best outcome-that’s exactly what I’d take “merit” to mean, so it seems like she’s reacting to inaccurate uses of words by perpetuating them.

    I don’t get why she seems to think that expulsion is the sole available means of redress, but if my more generous take is right, you can see why it might make a lot of sense to value what she calls “post-meritocracy” even in critical cases. A surgeon with good patient communication skills can help patients make better decisions which might well matter more than small differences in technical skills at surgery or board scores. A pilot able to recognize when someone on their team is frustrated or misunderstands might be able to avoid critical issues in ways that might more than compensate for worse easily-measured piloting skills. Those seem like the sorts of points she’s making. I don’t think her attacks make sense as a critique of the goal of choosing the best contributors, only as a way of recognizing that we need some humility in our epistemology about how that’s determined.

    I think there are legitimate grievances against her view, but I’m not clear why she seems so keen on exaggerating them and presenting her position as less reasonable than it seems to be. Perhaps it’s an effective means of mobilizing frustrated allies?

  3. Emhke seems perfectly capable of making the points she wants to make, Kelsey. And the things you describe aren’t at all the points she’s making. She’s not trying to broaden the term “meritocracy” for a meaning that includes social or emotional intelligence. She’s tearing it down as a patriarchal construct that oppresses minorities. You can agree or disagree with her fairly clearly communicated points if you like, but inventing arguments that she hasn’t bothered making and attributing them to her seems pointless to the discussion. In short, I don’t buy your claim that her position is more reasonable than she is presenting it. Her communication isn’t at all unclear. And I rather suspect she’d resent your remapping of her argument in ways that neuter her central emphasis on privilege and “fairness”.

  4. These are the sorts of things in her manifesto which seem to me to appeal to the value of merit as she sees it:

    “-We believe that interpersonal skills are at least as important as technical skills.
    -We can add the most value as professionals by drawing on the diversity of our identities, backgrounds, experiences, and perspectives.

    -We acknowledge the value of non-technical contributors as equal to the value of technical contributors.”

    Talk of skills being equally important and valued in the context of adding value sounds like an appeal to making the project better. That’s what I would mean by “merit”.

    Come to think of it, I would include professional ethics in “merit” as well. That ought to include elements intended to foster the health of the discipline as well as broader social good. A talented surgeon with no qualms about using her talents for evil sounds better suited to employment with Hydra than my local hospital.

    Would you differ with me about what merit is in any of those cases?

  5. Yes. I differ with you about what merit is in all of those cases. The meritocracy established by the OSS movement decades ago is measurable, testable and verifiable. And backed up with a track record of producing the backbone of the modern internet and cell phone systems. It says merely and only “does the contributed code work?” That question is deterministic and provable. Each of the “values” in the manifesto are the opposite. They are unverified statements that even if provable, certainly haven’t been proven. Let’s take that first one (skipping the “we do not believe” header) as an example of what I mean.

    “We believe that interpersonal skills are at least as important as technical skills.”

    Are they really? If I’m building a light-weight router that will work day-in and day-out for decades without dropping a packet do I want the engineer who talks nice or the one who produces the most solid piece of work possible that does only exactly what it needs to do and no more? Interpersonal skills may indeed be more important than technical skills in some contexts. But in producing OSS software? Interpersonal skills are a baseline, at best. Someone with none will be more aggravation than they are worth, sure. But even rudimentary communication skill is adequate when you’re working on a distributed team with a known, technical platform, that doesn’t actually live in the same country and speak English mostly as a second language.

    I think these statements of “value” are better characterized as principles. They represent “values that one holds”, not the “value that one creates with the work of ones mind and hands”. It’s a tricky rhetorical alchemy they’ve used, but it’s fundamentally incompatible with the OSS ideal of meritocracy. And designed to be such, I believe. It’s taken as a given that, of course, all right-thinking people will agree with these statements of principle and see them as “valuable”. And clearly those who wrote the manifesto think we will live in a better world if we trash the idea of “merit” and instead enter the world defined by these principles. That way, we can be “fair” and “diverse” and lots of other principles that may or may not actually produce “value”. But that will certainly assuage the political machinations of an energetic few who will then get to tell everyone else whether they measure up or not.

  6. Jacob-

    I like your point about a team that “speak[s] English most as a second language.” One of the things that I wish people would understand about these crusades to stamp out all inequality is how futile the effort really is. It’s basically impossible to see an elevation of communication skills as something other than a standard biased against people who speak English as a second language. It is thus not only potentially divisive, but certainly open to accusation of being a form of latent colonialism.

    The reason a critique like this is important is the demonstration that the purported ideals of social justice activists are at best more complex and hard to embody than let on and at worst an after thought in what is essentially a power grab. Their methods can’t achieve their stated ends, either due to some intrinsic defect in their ideology (that’s where I put my money) or just plain old sloppiness in execution.

  7. The line “We believe that interpersonal skills are at least as important as technical skills.” stood out to me as well. I’m not a software developer, and so not a developer who contributes remotely as part of a distributed team, but I do know what code is and does and I’ve screwed up enough code in my life to know that this statement about interpersonal skills is simply false in this context. I also can’t help but look with disgust at the words “at least” in that statement. She is just dying to have the chance to make the claim that in open source software development, *technical skills do not matter as much as interpersonal skills.* There is no way that is true.

  8. I think the idea of meritocracy was thoroughly disproven in the case of the Randolph versus Eddie Murphy’s character in trading places.

  9. Jacob, you wrote: “The meritocracy established by the OSS movement decades ago is measurable, testable and verifiable. … It says merely and only ‘does the contributed code work?’ That question is deterministic and provable.”

    That doesn’t sound right to me. Even in code no other human will ever see, there are other values, like efficiency, robustness, and security. If anyone else will ever look at the code, or even its author look at it later, transparency, reusability, and clarity of commenting are also important. If consumers will ever deal with the end product, a whole host of interface issues arise which are separate from the question of simply whether it’s possible to accomplish the intended purpose. And, of course, the quality of the code is only a part of what goes into a project–identifying and dividing up the various tasks needed for a project, managing the community, and documenting the software are also crucial if it’s to be of use in the world. If what you’re saying is that OSS has a history of a “meritocracy” which ignores the value of all of these things, why should anyone value a meritocracy which ignores most of the components of merit?

    Perhaps it would help to illustrate with an example. A couple years ago, I was setting up a website and wanted a forum with a few more features than the default WordPress options, so I chose Discourse. Its goals are fantastic, and it seems to work very well now that I have it set up. But the setup process was almost traumatically unpleasant. Instead of going to my WordPress dashboard, choosing Discourse from a list, and clicking “install” (which is the install process for all of the other forum software I considered), I had to contract with a different company to provide separate server space for the forum because it would only run on servers configured in certain ways. I had to dredge up my memory of Unix because there is no GUI for installation, and then deal with the fact that the only documentation available was a version or two out of date, and a few things had changed. I ended up trawling forums to find the information I needed, only mostly succeeding, and then trial-and-erroring the rest over a period of almost three weeks before I finally got it working correctly. I’m now terrified of every update lest it break something and send me back down that path. That’s the sort of user experience you get when you value only whether code works.

    Would anyone disagree that user interface choices, clear documentation, and the maintenance of a helpful community provide actual value to end users? I don’t see how any of that gets captured in your idea of merit. I’m not going to defend the claim that interpersonal skills are just as valuable as technical ones, because I don’t know what sense it makes to try and find a balance between two things which are both necessary. But I would say that if you can’t understand your users’ needs well enough to meet them, all the technical skill in the world won’t produce a useful product. Similarly, if the task is large enough that no one can do it alone, no amount of technical skill can finish the job if you drive away all potential collaborators. That seems to me like the non-hyperbolic way of saying what she’s getting at with that claim, and it seems utterly uncontroversial.

    Certainly, she’s also importing a whole political element, and that’s alienating. Indeed, I would suggest that choosing to explicitly reference politics reduces the merit of her point, displaying a failure of interpersonal skill. Similarly, the issue about those who don’t speak English seems to me like an interesting one, and she seems less aware of the social justice implications of her position than I would like. But, given the consistent complaints we hear about how women are forced to do unrecognized emotional labor and the fact that OSS projects seem often to rely on essentially unacknowledged managers doing exactly that to maintain the enthusiasm of the work force which keeps them contributing, it really seems like she’s identifying a genuine problem. It’s just also bound up in all sorts of problematic other stuff.

    Nathaniel, I don’t see any indication in her manifesto that she expects to stamp out all inequality, only that she intends to address some inequalities which are salient to her. Does that also seem futile to you?

  10. Kelsey: It’s not about what seems right to you. It’s about what seems right to the OSS community. And until now, that has been “does the code do what it is intended to do”. Which, incidentally, includes efficiency, clarity, and security. But has not included things like “open to regular users”. Which is something I hate about OSS, frankly, and the main reason I don’t use it. At any rate, their criteria has been clear-enough in the community that they haven’t needed anything else for some decades.

    Since it’s obvious that neither of us is part of that community, I’m not sure why we’re arguing about it. You seem to think the anti-meritocracy folks are just dandy and their call on principle is a benign extension of essential programmer skills. I think it’s more likely a power grab using values not currently in the equation for OSS so that those who can’t cut it in the current valuation can find a way to elevate their voices over those who are better programmers than they are. My valuation is based on being a programmer for nearly 25 years. I’m open to proof I’m wrong, but I want actual proof and not statements of principle that aren’t backed up by actual data.

    As for your conclusion, I don’t see how you can claim she has identified a genuine problem when the only problem identified in the manifesto is that the current system benefits “those with privilege” and that benefits aren’t “distributed fairly”. The lovely thing about privilege and fairness is that the terms have whatever meaning the speaker wants them to have and are thus impossible to disprove. Handy attributes for someone grasping unearned power. Less handy when trying to build software…

  11. Jacob, you’re right that merit in software development isn’t about what seems right to me, but it’s also possible for the OSS community to be mistaken about it. Indeed, you admit that their definition leads to consequences you hate, and I assume you also agree that it leads to their software being much less successful than it would have been otherwise. That seems like the proof you seek that there’s a problem. So, I’ll ask again, if their definition of merit undermines the success of their projects, why should we value that understanding? The ways I’ve seen genuine meritocracy defended involve appeals to its justice and success. This perversion of meritocracy seems problematic on both counts, and that seems like a consensus position.

    I’m not saying I don’t get why there’s lots of opposition to this woman—I’ve already criticized her approach several times (which makes your claim that I see her group as just dandy particularly galling). I think it has just the problems you mention: because it’s untethered to clear standards of evidence, it seems like a political move for power rather than a principled position. But saying simultaneously that you hate that OSS projects devalue understanding of other people (users) and also that there’s no meaningful criticism coming from people saying they ought to value that doesn’t make sense to me.

  12. “I assume you also agree that it leads to their software being much less successful than it would have been otherwise.”

    Why would you assume that? I certainly don’t. More effort spent on user experience would have to take effort from other areas. Just because they made trade-offs with consequences I dislike doesn’t mean those decisions were wrong. Frankly, I’m trying to imagine what it would look like in a world where Linux wasn’t hampered by it’s “much less successful” model. Do you seriously claim that Linux has been hampered all these years by its adherence to meritocracy?!? That begs for some backing data…

    “But saying simultaneously that you hate that OSS projects devalue understanding of other people (users) and also that there’s no meaningful criticism coming from people saying they ought to value that doesn’t make sense to me.”

    Well, there’s lots of criticism. Some of it is meaningful, some not. I believe that the anti-meritocracy folks are in the “not” category for reasons I’ve tried to articulate. The thing I find weird is that you’d think that just because I hate the results of their emphasis that I’d think they should therefore change it. Yeah, that’d be nice for me. But I’m not the only person on the planet. And it is without question that the Linux community, for starters, has produced outstanding software that is so good that millions ignore the pain of learning to use it. Every day. And those people use OSS projects to sustain the modern internet and cell phone infrastructure. That’s an incredible achievement and anybody who wants to change the processes and values used to attain those accomplishments should step lightly and respectfully and not go in guns blazing to collect scalps for their pet political cause. I’m not willing to say the OSS emphasis on meritocracy is the best system. Nor am I saying that everyone should adopt that same emphasis. But what I very much *am* saying is that it worked very well in the OSS community and managed to support incredible achievements doing something nobody has ever done before–coordinated thousands of contributors from hundreds of countries to produce things that are actually useful. That they’re also free (for some values of free, anyway) is the cherry on top. People can come up with their own theories and principles and advocate that we try them out. But setting out to destroy one of the most successful organizing frameworks in the world today strikes me as foolish, at best. Yeah, someone may come up with a better way. But how about we try it out some and build experience and evidence about how and why it works before we go in with big boots on and tell people they’re doing it wrong???

  13. Jacob, that’s an admirably humble epistemic position you’re taking in refusing to extrapolate from your own preferences to grounds for success. I wasn’t expecting it, which is why I made the assumption I did.

    I can think of some evidence for the claim that the OSS community is doing it wrong: markets. Companies with a profit motive, on average, value these softer skills far more than OSS projects. Linux is free, but still has a smaller desktop market share than Windows or MacOS. There is plentiful evidence that emotional work is undervalued in our culture generally (one example:, and it seems to be entirely absent from the proposed metric of OSS merit. That seems like sufficient reason to try an alternative, at least for some OSS projects. I’m totally with you on not wanting jackbooted thugs to force things on every project, but even the exaggerated campaign Ehmke is waging seems to be targeted to a few projects and to proceed via persuasion, rather than force. Comparisons to literal murder seem a tad histrionic.

    And let’s return to that standard for a moment. Earlier, you wrote that it was simply “does the code work”, and that it was deterministic and provable. But, later, you claimed that this includes efficiency, clarity, and security. Do you see the tension there?

Comments are closed.