Blogs & Articles: Don't Break Userspace! đź”— 33 weeks ago
- Category: Blogs & Articles | Bitcoin Magazine: Bitcoin News, Articles, Charts, and Guides
- Author(s): Shinobi
- Published: 19th January 2024 15:46
“Mauro, SHUT THE FUCK UP!It's a bug alright - in the kernel. How long have you been a maintainer? And you *still* haven't learnt the first rule of kernel maintenance?If a change results in user programs breaking, it's a bug in the kernel. We never EVER blame the user programs. How hard can this be to Understand?” -Linus Torvalds
Don’t break userspace. This is Linus Torvald’s golden rule for development of the Linux kernel. For those of you reading this who are not familiar with the nature of Linux, or operating systems in general, the kernel is the heart and soul of an operating system. The kernel is what actually manages the hardware, moving bits around between storage and RAM, between the RAM and the CPU as things are computed, and all of the little devices and pieces of the actual computer that need to be controlled at the hardware level.
Every application or program written for an operating system has to interact with the kernel. When you download Photoshop, or Telegram, everything that program is doing boils down to essentially calling the kernel. “Hey kernel, take what I just typed and process it and send it over a network connection to the server.” “Hey kernel, take the color shift I made to this pitch, take it out of RAM and send it to the CPU to modify it, then put it back in RAM.”
When the kernel is changed, in a somewhat similar fashion to Bitcoin, the chief goal of developers is to ensure that existing applications that assume a specific way to interact with the kernel do not break because of a change to the kernel. Sounds very familiar to Bitcoin and the necessity to maintain backwards compatibility for network consensus upgrades doesn’t it?
“Seriously. How hard is this rule to understand? We particularly don't break user space with TOTAL CRAP. I'm angry, because your whole email was so _horribly_ wrong, and the patch that broke things was so obviously crap. The whole patch is incredibly broken shit. It adds an insane error code (ENOENT), and then because it's so insane, it adds a few places to fix it up ("ret == -ENOENT ? -EINVAL : ret").
The fact that you then try to make *excuses* for breaking user space, and blaming some external program that *used* to work, is just shameful. It's not how we work.Fix your f*cking "compliance tool", because it is obviously broken. And fix your approach to kernel programming.” -Linus Torvalds
Linux is one of the most important, if not the most important, open source project in the entire world. Android runs on Linux, half of the backend infrastructure (if not way more) runs on Linux. Embedded systems controlling all kinds of computerized things in the background of your life you wouldn’t even consider run on Linux. The world literally runs on Linux. It might not have taken over the desktop as many autistic Linux users wanted to see happen, but it quietly ate almost everything else in the background without anyone noticing.
All of these applications and programs people use in the course of their daily lives depend on the assumption that Linux kernel developers will not break backwards compatibility in new versions of the kernel to allow their applications to continue functioning. Otherwise, anything running applications must continue using older versions of the kernel or take on the burden of altering their applications to interact with a breaking change in the kernel.
Bitcoin’s most likely path to success is a very similar road, simply becoming a platform that financial applications and tools are built on top of in such a way that most people using them won’t even realize or consider that “Bitcoin ate the world.” In a similar vein to Linux, that golden rule of “Don’t break userspace” applies tenfold. The problem is the nature of Bitcoin as a distributed consensus system, rather than a single local kernel running on one person’s machine, wildly changes what “breaking userspace” means.
It’s not just developers that can break userspace, users themselves can break userspace. The entire last year of Ordinals, Inscriptions, and BRC-20 tokens should definitively demonstrate that. This offers a very serious quandary when looking at the mantra of “Don’t break userspace” from the point of view of developers. As much as many Bitcoiners in this space do not like Ordinals, and are upset that their own use cases are being disrupted by the network traffic Ordinals users are creating, both groups are users.
So how do developers confront this problem? One group of users is breaking userspace for another group of users. To enact a change that prevents the use of Ordinals or Inscriptions explicitly violates the mandates of don’t break userspace. I’m sure people want to say “Taproot broke userspace!” in response to this dilemma, but it did not. Taproot activation, and the allowance for witness data to be as large as the entire blocksize, did not break any pre-existing applications or uses built on top of Bitcoin. All it did was open the door for new applications and use cases.
So what do we do here? To try and filter, or break by a consensus change, people making Inscriptions or trading Ordinals is to fundamentally violate the maxim of “don’t break userspace.” To do nothing allows one class of users to break the userspace of another class of users. There is fundamentally no solution to this problem except to violate the golden rule, or to implement functionality that allows the class of users’ whose userspace is broken now to adapt to the new realities of the network and maintain a viable version of their applications and use cases.
Not breaking the userspace of Bitcoin is of critical importance for its continued success and functionality, but it is not as simple as “don’t change anything.” Dynamic changes in user behavior, that require no change to the actual protocol itself, can have the same effect at the end of the day as a breaking change to the protocol. Are developers supposed to pick and choose which applications’ userspace is broken to maintain that of another application? I would say no, and go further to say that anyone advocating for such behavior from developers is demanding them to act irresponsibly and in a way that harms users of the system. So what is the answer here?
There is no answer except to push forward and continue adding improvements to the protocol that allow applications being broken by the behavior of certain users to function in the presence of emergent changes in users’ behavior. Otherwise, you are asking developers to throw out the golden rule and effectively play kingmakers in regards to what use cases are viable to build on top of Bitcoin.
If we go down that road, then what are we actually doing here? I can’t tell you what we’re doing at that point, but I can tell you it’s not building a distributed and neutral system anymore.