This sounds more like a workaround to a website with poor navigation and UX, but ultimately you are still at the mercy of the website, eg if they have /articles/<title> links but throw 404 when you try to access /articles directly. It's up to the website to choose how navigation works, and I don't think we should constrain the structure of website urls just to support these hacky workarounds.

Now there are other reasons for hierarchal urls mentioned in the thread, like providing better semantic meaning when representing content that is already hierarchal. Though I find that that is rarely the case. For example, you might think to represent blog posts like myblog.com/articles/<title>. But you could also do it like myblog.com/<article year>/title (eg myblog.com/2017/why-choose-short-urls). I've seen both formats, and both are valid to a certain degree. So choosing one is a bit arbitrary. But if you allow both formats, that just pushes the ambiguity onto the user. For example if the user wants to bookmark the article, which url do they bookmark?

From my experience, data follows graph structures, so trying to force a hierarchal structure and doing things like encoding paths in urls, feels too arbitrary and adds unnecessary ambiguity.

> It's up to the website to choose how navigation works

Yes, and they should live by their choice and implement whatever is implied by their semantics.

URLs also have their own affordances and I expect to be able to act upon them.

Url hacking is a clunky and extremely uncommon method of navigation. It should be the least of the website's concern. We don't design our doors based around the 0.01% of people that get in by kicking them down.