If you're one of the people in the software freedom community who is attending O'Reilly's Open Source Software Convention (OSCON) next week here in Portland, you may have seen debate about O'Reilly and Associates (ORA)'s surreptitious Code of Conduct change (and quick revocation thereof) to name “political affiliation” as a protected class. If you're going to OSCON or plan to go to an OSCON or ORA event in the future, I suggest that you familiarize yourself with this issue and the political historical context in which these events of the last few days take place.

First, OSCON has always been political: software freedom is inherently a political struggle for the rights of computer users, so any conference including that topic is necessarily political. Additionally, O'Reilly himself had stated his political positions many times at OSCON, so it's strange that, in his response this morning, O'Reilly admits that he and his staff tried to require via agreements that speakers … refrain from all political speech. OSCON can't possibly be a software freedom community event if ORA's intent … [is] to make sure that conferences put on for the exchange of technical information aren't politicized (as O'Reilly stated today). OTOH, I'm not surprised by this tack, because O'Reilly, in large part via OSCON, often pushes forward political views that O'Reilly likes, and marginalizes those he doesn't.

Second, I must strongly disagree with ORA's new (as of this morning) position that Codes of Conduct should only include “protected classes” that the laws of a particular country currently recognize. Codes of Conduct exist in our community not only as mechanism to assure the rights of protected classes, but also to assure that everyone feels safe and free of harassment and hate speech. In fact, most Codes of Conduct in our community have “including but not limited to” language alongside any list of protected classes, and IMO all of them should.

More than that, ORA has missed a key opportunity to delineate hate speech and political speech in a manner that is sorely needed here in the USA and in the software freedom community. We live in a political climate where our Politician-in-Chief governs via Twitter and smoothly co-mingles political positioning with statements that would violate the Code of Conduct at most conferences. In other words, in a political climate where the party-ticket-headline candidate is exposed for celebrating his own sexual harassing behavior and gets elected anyway, we are culturally going to have trouble nationwide distinguishing between political speech and hate speech. Furthermore, political manipulators now use that confusion to their own ends, and we must be ever-vigilant in efforts to assure that political speech is free, but that it is delineated from hate speech, and, most importantly, that our policy on the latter is zero-tolerance.

In this climate, I'm disturbed to see that O'Reilly, who is certainly politically savvy enough to fully understand these delineations, is ignoring them completely. The rancor in our current politics — which is not just at the national level but has also trickled down into the software freedom community — is fueled by bad actors who will gladly conflate their own hate speech and political speech, and (in the irony that only post-fact politics can bring), those same people will also accuse the other side of hate speech, primarily by accusing intolerance of the original “political speech” (which is of course was, from the start, a mix of hate speech and political speech). (Examples of this abound, but one example that comes to mind is Donald Trump's public back-and-forth with San Juan Mayor Carmen Yulín Cruz.) None of ORA's policy proposals, nor O'Reilly's public response, address this nuance. ORA's detractors are legitimately concerned, because blanketly adding “political affiliation” to a protected class, married with a outright ban on political speech, creates an environment where selective enforcement favors the powerful, and furthermore allows the Code of Conduct to more easily become a political weapon by those who engage in the conflation practice I described.

However, it's no surprise that O'Reilly is taking this tack, either. OSCON (in particular) has a long history — on political issues of software freedom — of promoting (and even facilitating) certain political speech, even while squelching other political speech. Given that history (examples of which I include below), O'Reilly shouldn't be surprised that many in our community are legitimately skeptical about why ORA made these two changes without community discussion, only to quickly backpedal when exposed. I too am left wondering what political game O'Reilly is up to, since I recall well that Morozov documented O'Reilly's track record of political manipulation in his article, The Meme Hustler. I thus encourage everyone who attends ORA events to follow this political game with a careful eye and a good sense of OSCON history to figure out what's really going on. I've been watching for years, and OSCON is often a master class in achieving what Chomsky critically called “manufacturing consent” in politics.

For example, back in 2001, when OSCON was already in its third year, Microsoft executives went on the political attack against copyleft (calling it unAmerican and a “cancer”). O'Reilly, long unfriendly to copyleft himself, personally invited Craig Mundie of Microsoft to have a “Great Debate” keynote at the next OSCON — where Mundie would “debate” with “Open Source leaders” about the value of Open Source. In reality, O'Reilly put on stage lots of Open Source people with Mundie, but among them was no one who supported the strategy of copyleft, the primary component of Microsoft's political attacks. The “debate” was artfully framed to have only one “logical” conclusion: “we all love Open Source — even Microsoft (!) — it's just copyleft that can be problematic and which we should avoid”. It was no debate at all; only carefully crafted messaging that left out much of the picture.

That wasn't an isolated incident; both subtle and overt examples of crafted political messaging at OSCON became annual events after that. As another example, ten years later, O'Reilly did almost the same playbook again: he invited the GitHub CEO to give a very political and completely anti-copyleft keynote. After years of watching how O'Reilly carefully framed the political issue of copyleft at OSCON, I am definitely concerned about how other political issues might be framed.

And, not all political issues are equal. I follow copyleft politics because it's my been my day job for two decades. But, I admit there are stakes even higher with other political topics, and having watched how ORA has handled the politics of copyleft for decades, I'm fearful that ORA is (at best) ill-equipped to handle political issues that can cause real harm — such as the current political climate that permits hate speech, and even racist speech (think of Trump calling Elizabeth Warren “Pocahontas”), as standard political fare. The stakes of contemporary politics now leave people feeling unsafe. Since OSCON is a political event, ORA should face this directly rather than pretending OSCON is merely a series of technical lectures.

The most insidious part of ORA's response to this issue is that, until the issue was called out, it seems that all political speech (particularly that in opposition to the status quo) violated OSCON's policies by default. We've successfully gotten ORA to back down from that position, but not without a fight. My biggest concern is that ORA nearly ran OSCON this year with the problematic combination of banning political speech in the speaker agreement, while treating “political affiliation” as a protected class in the Code of Conduct. Regardless of intent, confusing and unclear rules like that are gamed primarily by bad actors, and O'Reilly knows that. Indeed, just days later, O'Reilly admits that both items were serious errors, yet still asks for voluntary compliance with the “spirit” of those confusing rules.

How could it be that an organization that's been running the same event for two decades only just began to realize that these are complex issues? Paradoxically, I'm both baffled and not surprised that ORA has handled this issue so poorly. They still have no improved solution for the original problem that O'Reilly states they wanted to address (i.e., preventing hate speech). Meanwhile, they've cycled through a series of failed (and alarming) solutions without community input. Would it have really been that hard for them to publicly ask first: “We want to welcome all political views at OSCON, but we also detest hate speech that is sometimes joined with political speech. Does anyone want to join a committee to work on improvements to our policies to address this issue?” I think if they'd handled this issue in that (Open Source) way, the outcome would have not be the fiasco it's become.