415 karma · joined July 27, 2011
Basically both are concerned with different parts of Authentication, initial vs on-going (though 2 legged Oauth2 is also an initial authentication step).
The line between authentication and authorization can get quite fine, especially if the authorization policy is as simple as "this set of services can talk to me" if you use mTLS and a fixed set of trusted services in a trust-store.
It reminds me of not being able to pick a milk without getting scorn from somewhere (real milk ~ cows, oat milk ~ sugar, macadamia milk ~ env impact, soy milk ~ hormones, etc). All the while we're watering alfalfa to feed cows for beef.
Oh boy the bikes are less repairable, even though half the city dwellers could be replacing cars with electric bikes. A car has way higher repair bills and up front purchase costs not to mention the ongoing costs.
It really feels like a substantive step forward in terms of computer utility kind of like spreadsheets, databases, apps. We'll see how far it takes us down the line of human replacement though.
Covered a lot of things I was thinking reading the original article.
But I also think the fact that they both get abbreviated to "auth" also causes a lot of bad mental modeling and poor communication.
The big problem these deep inheritance hierarchies cause is mediocre programmers often think they're writing good code (often high focus on maximal reuse) but the interfaces and clarity are terrible.
Also great time to plug a related post, What color is your function:
https://journal.stuffwithstuff.com/2015/02/01/what-color-is-...
1. Move to front lists are at most 2x the optimal list order search time (and are often much better than any static list order). [1]. A similar conjecture exists for Splay trees but is unproven called the dynamic optimality conjecture.
2. Randomized algorithms (e.g quicksort) often have the same worst case time as a non-randomized one but often are much faster in practice than their non-randomized variants.
1: https://www.dgp.toronto.edu/public_user/JamesStewart/378note...
I guess we'll all find out if/when people use the model to get it to disclose other users information but I won't be holding my breath.
e.g. before going to production with real users these cases should error as hard as possible, including potentially logging crashing the program to make the program errors super visible. Then in production it should largely log these things in error logs and keep chugging along.
And not like an environment variable with conditions. Something as first class as changing the standard exception and fatal calls behaviors.
"While I'm excited about the potential improvements to traffic safety, congestion and waste of life that self driving cars can bring. I still yearn for improvements to public transit and micromobility that can eliminate cars from the road altogether and yield major improvements in density, efficiency, safety and pollution (noise, tires)"
I'm sure you know all this and you were trying to make a high level point, but this old school diagram from the java content repository days shows a nice breakdown of functional differences between the two that lead to very different outcomes: https://docs.jboss.org/author/display/MODE28/images/author/d...
"FoundationDB: A Distributed Key-Value Store" [https://news.ycombinator.com/item?id=36572658]
"FoundationDB Record Layer" [https://news.ycombinator.com/item?id=18906341]
"Apple Acquires FoundationDB [https://news.ycombinator.com/item?id=9259986]
"How FoundationDB works and why it works" [https://news.ycombinator.com/item?id=37552085]