The case for gatekeeping, or: why medieval guilds had it figured out
19 points by alper
19 points by alper
I have a friend who was enthusiastic about an IT guild idea years ago, so we had a lot of discussions about it. At the time the idea didn't get any traction from anyone we knew.
The medieval guild system gets a bad rap. It's usually remembered as a protectionist racket // a cartel of craftspeople colluding to keep prices high and competition low.
For the record, in many guilds there were two paths to become a master:
So path #1 was there mostly just to maintain the illusion of openness and meritocracy in practice. How to prevent a guild from devolving into that is a big question.
But the culture developed an allergy to gatekeeping so severe that suggesting contributors should meet any bar at all became politically radioactive. And that allergy made perfect sense when the failure mode was "talented person gets excluded by arbitrary social dynamics."
More like, an allergy to the mentions that gatekeeping already exists and, e.g., women are routinely excluded by social dynamics. Introducing explicit gatekeeping threatens the implicit forms.
I mean something closer to what the Debian project has done with its Web of Trust model for decades: existing trusted contributors vouch for new ones.
Disclaimer: things have improved since then, so this described my interaction with a guild, not with modern Debian.
But the story was, first time I tried to contribute to Debian without knowing any existing developers, the package I uploaded to the mentors server was ignored and automatically deleted because no one looked at it in the allotted two-four weeks.
I also remember the main Linux User Group in the location where I was growing up. It was a norm there to kick and ban people from their IRC channel for asking beginner questions. In the end, most of the channel traffic was just social chit-chat of the established members, with a very occasional deep technical discussion. That gatekeeping certainly achieved some goal but also ensured that all beginners would go elsewhere and the group stagnated.
Any guild that wants to be a force for good has to prevent those things from happening, somehow.
could amount to personally recreating all dependencies of the actual application you want to show as a masterpiece
Write a compiler, and a network stack, and a database etc. etc. That sounds very similar to my university curriculum.
Many FOSS-adjacent development-level places have an implicit economic gatekeeping, that they don't necessarily like to acknowledge, kind of inherent — being able to compile stuff at the relevant scale has a price… and also having the time to spend to become competent in the locally relevant technical details is an economic filter.
If you accept that this is what it is, I guess gatekeeping via «feeder» projects (too exotic to be pragmatically attractive if you don't like learning, too small and raw for a new participant to have plausible excuses to not have the expected lurking-acquired competence) could work, in the sense that some people with standing have pet interests too, and you can impress them there.
Of course, there will be still networking benefit from e.g. being a student in a place where a person with a standing teaches, and making an impression there.
Honestly, for many groups it made sense before and makes sense now to kick for questions below the level the group… although normal LUG positioning is an «entry» group, which also often means not getting interest from new non-beginner members…
And if you want to improve the headline economical-class-gatekeeping problem in comparison to its state in 2015 or 2005 or whatever, then I guess it would become an attempt to reform the economics at least at a common-market-bloc scale, which could be way more valuable if it succeeded, regardless of the gatekeeping problem part, but also much harder.
FWIW, to me this feels like a platform problem. My dim recollection of open-source in the pre-Github era is that it worked a lot like the author describes here:
The open source ecosystem used to have something like this, but it was organic. You'd show up on a mailing list. You'd lurk. You'd file a good bug report. You'd submit a small patch and wait. Over time, established contributors would come to recognize your handle and your judgment.
What changed? Well, Github and similar forges showed up with a big prominent “pull request” flow, and encouraged the idea that anyone could contribute without engaging first. Up until very recently, Github didn’t even allow disabling PRs on a repository.
The drive-by contributor flow has become so normalized at this point that I have seen people claim that a project isn’t truly “open source” if it doesn’t accept PRs on Github based purely on the quality of the specific code being added. “Real open source” apparently meaning that you are opting into running a free-to-join online community for your project.
I agree that a gatekeeping mechanism is needed, but I’ll admit to being skeptical of either an organized guild system or some kind of web-of-trust reputation system. I honestly think we need a shift in norms back to the idea that a project should control its membership, and that the best way to add new members is by building trust through engaging with the existing membership. And that platforms should support that model, even possibly make it the default.
Please read The European guilds: an economic analysis by Ogilvie, for why the medieval guilds actually didn’t have it figured out. It’s not as short a read as a blog, though.
https://press.princeton.edu/books/hardcover/9780691137544/the-european-guilds
There are organizations similar to guilds around today. You can't practice medicine or law in many jurisdictions without being licensed, and the certification is often grueling and expensive.
And many fields of engineering, at least in Spain, have the same, you need a license from a "guild", an insurance, ... There have been people pushing for it in software engineering for at least 20 years (I personally know some of them) but a majority of software developers weren't really into it because there will be a lot of self learners out (including people with some degree but not exactly related to IT, like physics). It's also something that requires a lot of politics to get approval. I think in my country this gatekeeping mechanism will not be in place until some other bigger countries start doing it.
I would love to see more approaches to bring back gatekeeping.
Not the first time there's been a need for social defence measures against technology and it won't be the last either.
What a modern guild would actually look like
Linux distributions, and other projects, that collectively collaborate to write, maintain and package software fulfill this role.
Like, I'm not sure how much people are aware of the interconnected relationships that exists between projects like KDE, Arch, Debian, Gnome, systemd and the Linux kernel? People are recognized by their contributions there, by the title you hold in the projects and the topics you engage with.
It already has a ladder system and a gatekeeping mechanism as well.
So… if anyone could create an account on lobsters, but only comment after collecting a number of upvotes with their submissions, could that work?
Incidentally, the best communities I have been part of were pretty gatekeep-y. I wonder how much my privilege affects me here, but they always made good impressions on me. There was a clear path to membership, and reputation systems in fora etc were the norm back then, so it felt natural. Also, having to work for it made membership special, and you wouldn't want to risk it by behaving recklessly without good reason.
So… if anyone could create an account on lobsters, but only comment after collecting a number of upvotes with their submissions, could that work?
I'm not sure that would be a good idea. Many folks who would be good commenters are not necessarily the same as people who are first to find and post interesting submissions.
Guilds may work, but for me one of the issues is the "long tail" problem.
There are a small number of people who do a large amount of work on Open Source. There is a much longer tail of people who do very little. AI has enabled that second group of people to greatly increase their output. That feels good to them.
The downside is that the first group is now inundated with garbage. This is essentially the spam problem from email.
For me, the problem solved by the guilds is reputation. If there is a way for Open Source maintainers to figure out peoples reputation, then it's much easier to triage PRs.
But when it takes 5 minutes to create a new online account, a good metric is to look at history. New accounts, unknown accounts, etc. are likely to be garbage.
That doesn't address the problem of how new people can bootstrap into having a good reputation. But the sheer volume of AI slop means that the Open Source maintainers have to protect themselves.
Perhaps the analogy could be made stronger with members-only source code? I'd call it, "clopen-source": open, but closed at the same time.
I'd like to see open source projects such as for instance Python make their licenses conditional for the major platform players. That could be quite interesting.
It beats bemoaning that Amazon, Google, OpenAI etc. benefit so much from open source projects without paying back.