I disagree. We add complexity largely (but not exclusively) because they fix problems. The entire point of a web browser exists to solve the very complex problem "How do we run arbitrary code on client hardware in a way that's safe for all parties?"
What’s about sharing documents which is still it’s primary purpose. And most of the current complexity stems from the sprawling JS API surface. We know how to sandbox software (VM and malware analysis,…), browser’s complexity is mostly corporate driven.
I had forgotten the original purpose of web browsers was to view hyperlinked documents. It seems like we abandoned that as the primary use case over 20 years ago, with technology such as Flash, Silverlight, and Java Weblets.
I don’t think that has been its primary purpose for a long time. If you want to share a document for the most part you email it. It has long since moved from documents to applications.
VMS was a sandbox-based OS released in the 70s that solved "how do we run arbitrary code on client hardware in a way that's safe for all parties?" Browsers are not that.
Browsers do have this component, but that's not the reason they exist.
I wonder how productive it is to mention that though. Sure, that may not be the reason they exist and came to be (something about exchanging documents) but it may be how they are mostly used today.
You are insisting on justifying all of the complexity of browsers by claiming the same functionality that people got in the 70's with a small OS that was hand-written in assembly.
Sorry, but you don't seem to want a productive discussion. You just have a preconceived idea that turns out to be wrong.
I am not insisting anything. I wanted to point out that the "reason why browsers exist/came to exist" can be relevant in this discussion but is not a direct logical contradiction to their current complexity.
If I say, the pdf came to be in order to be a simple way to digitally represent printed/printable documents that may be correct but wouldn't actually express anything about the current complexity of the pdf format. Admittedly perhaps a bad example due to how similar it is.
It's possible I misunderstood though. If so, I apologize.
Some subset of added complexity is justified as the problems we’ve tried to solve have gotten more complex.
Another subset is added complexity for the reasons I listed (no good feedback loops to prevent it), which is imo not good. I also feel that this subset is larger than the first subset.
