My only use case so far has been fixing broken builds when a package has build-)ldependencies that don't actually work (e.g. a dependency of a dependency breaks stuff). Not super common, but it happens.

But pyproject.toml supports neither locking nor constraints.

How does HATEOAS deal with endpoints that take arguments? E.g. I have an endpoint that merges the currently viewed resource with another one? Does it require a new (argumentless) endpoint showing a form where one can enter the second resource? Wouldn't it be quite inefficient if you have to now do two (or more) requests instead of just one?

I'm taking a broad approach? The article is literally about the FCC. You know, the Federal Communications Commission. That applies to the entire country.

Indeed two companies is not really competition. So why not focus on that, instead of reducing choice, which may lead to even less competition by making differentiation harder?

If there is no reason for caps, why wouldn't one of these companies simply remove them, giving them a competitive advantage, and making them more money? Why would one company reject making more?

Maybe capless actually costs them more due to bad infrastructure, and they don't see consumer demand for it? Forcing them to go capless would in that case result in higher prices.

Maybe they form a cartel and have collectively decided to keep caps. But why, if it doesn't actually cost them more to remove the caps? And if it does, then prices would again rise if forced to go capless.

I would like to know how you figure that load of horseshit. The average customer never even hits the data cap, so it’s not like it’s just cutting people off so others can get on.

Well no, of course not. That would upset people. What it does is make people afraid of hitting the limit, which makes them concious of data use and reduces it, even if it does not actually hit the cap.

Very few places have any choice about what service they get.

Most of the country has no choice, so remove choice from those that do, to make it even? Shouldn't we rather make it even by giving everyone choice? How about instead e.g. forcing ISPs to offer capless plans, while still allowing for capped, but cheaper, plans for those that prefer it?

I’m confused where you believe consumers are given choice here.

I'm confused by you being confused. Consumers can pick a subscription with a data cap, or they can pick one without. Maybe you can clarify what you are confused about?

Clearly this is a marketing issue, not a technical one.

Why not both? Marketing can be a great way to work around technical issues, e.g. by steering consumer behaviour in a way that avoids the technical issues.

Also, just because one network has sufficient spare capacity to not steer users to reduce data usage does not mean that every network does that. In fact this is where choice comes in: I can pick a provider which spends more money on the network, resulting in a higher costs, but also higher caps. Or I can pick a provider that spends less on networks, resulting in lower costs, but needing caps to make sure the limited bandwidth is sufficient for all customers.

The industry has grown up since then, technically speaking, and there is no cause for data caps except to line the pockets of ISPs.

You mean except the reason I gave, and you ignored?