Package Managers Keep Using Git as a Database, It Never Works Out

That works while the size of the community is ~100-200 people, when everyone knows everyone else personally. It breaks down rapidly after that. We compensate for that with hierarchies of governance, which give rise to written laws and bureaucracy.

New tribes break off old tribes, form alliances, which form larger alliances, and eventually you end up with countries and counties and vovoidships and cities and districts and villages, in hierarchies that gain a level per ~100x population increase.

This is sociopolitical history of the world in a nutshell.

No it does not. This sentiment, which many people have, is based on a fictional and idealistic notion of what small communities are like having never lived in such communities.

Empirically, even in high-trust small villages and hamlets where everyone knows everyone, the same incentives exist and the same outcomes happen. Every single time. I lived in several and I can't think of a counter-example. People are highly adaptive to these situations and their basic nature doesn't change because of them.

Humans are humans everywhere and at every scale.

Commons can fail, but the whole point of Hardin calling commons a "tragedy" is to suggest it necessarily fails.

Compare it to, say, driving. It can fail too, but you wouldn't call it "the tragedy of driving".

We'd be much better off if people didn't throw around this zombie term decades after it's been shown to be unfounded.

that seems like an unreasonable bar, and less useful than "does this system make ToC less frequent than that system"

But it appears we cannot avoid getting into the weeds a bit…

> Governments and corporations don't exist in nature.

This is not as simple as you seem to think.

The claim “don’t exist in nature” is vague, because the word “nature” in common speech is vague. What is “natural”? Is a beehive “natural” Is a house “natural”? Is synthetic water “natural”? (I claim that the concept of “nature” concerns what it means to be some kind of thing. Perhaps polystyrene has never existed before human beings synthesized it, but it has a nature, that is, it means something to be polystyrene. And it is in the nature of human beings to make materials and artifacts, i.e., to produce technology ordered toward the human good.)

So, what is government? Well, it is an authority whose central purpose is to function as the guardian and steward of the common good. I claim that parenthood is the primordial form of human government and the family as the primordial form of the state. We are intrinsically social and political animals; legitimate societies exist only when joined by a common good. This is real and part of human nature. The capacity to deviate from human nature does not disprove the norm inherent to it.

Now, procedurally we could institute various particular and concrete arrangements through which government is actualized. We could institute a republican form of government or a monarchy, for example. These are historically conditioned. But in all cases, there is a government. Government qua government is not some arbitrary “construct”, but something proper to all forms and levels of human society.

> "Tragedy of the commons" is a general coordination problem.

We can talk about coordination once we establish the ends for which such coordination is needed, but there is something more fundamental that must be said about the framing of the problem of the “tragedy”. The framing does not presume a notion of human beings as moral agents and political and social creatures. In other words, it begins with a highly individualist, homo economicus view of human nature as rationally egoist and oriented toward maximizing utility, full stop. But I claim that is not in accord with human nature and thus the human good, even if people can fall into such pathological patterns of behavior (especially in a culture that routinely reinforces that norm).

As I wrote, human beings are inherently social animals. We cannot flourish outside of societies. A commons that suffers this sort of unhinged extraction is an example of a moral and a political failure. Why? Because it is unjust, intemperate, and a lack of solidarity to maximize resource extraction in that manner. So the tragedy is a matter of a) the moral failure of the users of that resource, and b) the failure of an authority to regulate its use. The typical solution that’s proposed is either privatization or centralization, but both solutions presuppose the false anthropology of homo economicus. (I am not claiming that privatization does not have a place, only that the dichotomy is false.)

Now, I did say that the case with something like github is analogical, because functionally, it is like a common resource, just like how social media functions like a public square in some respects. But analogy is not univocity. Github is not strictly speaking a common good, nor is social media strictly a public square, because in both cases, a private company manages them. And typically, private goods are managed for private benefit, even if they are morally bound not to harm the common good.

That intent, that purpose, is central to determining whether something is public or private, because something public has the common benefit as its aim, while something private has private benefit as its aim.

You’re fundamentally misunderstanding what tragedy of the commons is. It’s not that it’s “zero-cost” for the participants. All it requires a positive return that has a negative externality that eventually leads to the collapse of the system.

Overfishing and CO2 emissions are very clearly a tragedy of the commons.

GitHub right now is not. People putting all sorts of crap on there is not hurting github. GitHub is not going to collapse if people keep using it unbounded.

Not surprisingly, this is because it’s not a commons and Microsoft oversees it, placing appropriate rate limits and whatnot to make sure it keeps making sense as a business.

Probably the worst thing MS did was kill GitHub’s nascent CI project and replace it with Azure DevOps. Though to be fair the fundamental flaws with that approach didn’t really become apparent for a few years. And GitHub’s feature development pace was far too slow compared to its competitors at the time. Of course GitHub used to be a lot more reliable…

Now they’re cramming in half baked AI stuff everywhere but that’s hardly a MS specific sin.

MS GitHub has been worse about DMCA and sanctioned country related takedowns than I remember pre acquisition GitHub being.

Did I miss anything?

As for how the site has become worse, plenty of others have already done a better job than I could there. Other people haven't noticed or don't care and that's ok too I guess.

It's a very hacky feeling addon that RKE2 has a distributed internal registry if you enable it and use it in a very specific way.

For the rate at which people love just shipping a Helm chart, it's actually absurdly hard to ship a self contained installation without just trying to hit internet resources.

Using the slowest clone method they measured 8s for a 750 MB repo, 0.45s for a 40MB repo. appears to be linear so 1.1s for 100MB should be a valid interpolation.

So doing 30 of those per second only takes 33 cores. EPYC servers have 384 cores.

And remember we're using worst case assumptions in a 3 places (full clone each time, using the slowest clone method, and numbers from old hardware). In practice I'd bet a fastish laptop would suffice.