Would you describe today's web browsers as “innovative?”

The good think is that there's a lot more of variety on apps and plugins than there are different viable mainstream browsers, and I can use those for specific functions at different times. Bundling everything within the same application would force me to have them in my interface all the time.

Why do we still use IP/TCP when many of the same people who helped develop IP/TCP learned important lessons from it and moved on to new ideas such as RINA? We all know that including a port number in a network address violates the principles that the levels in a network stack should be isolated, so why have we allowed 40 years to go by, without doing anything to address this mistake?

We know there is a need for structured documents, and we have endless serialization formats for various ontololgies. We also know there is a need for a GUI that works over networks. We all know what a struggle it has been, for the last 26 years, to make HTML serve both purposes. Isn't it time we get rid of HTML and replace it with different technologies that can specialize in either being a GUI, or in delivering structured data?

Ethernet was introduced in 1980 and became the dominant wire for corporate and data center networks, despite more efficient formats being possible. Why is there such overwhelming conservatism in this area?

Polyglot programming has become the norm on servers, but the client has become a monoculture where Javascript killed off the other competing technologies (Flash, Swing, etc). An innovative browser would be one that gave us a virtual machine in the client that could support polyglot programming on the client. Instead, most "innovation" in 2015 is focused on making Javascript incrementally better.

"Browser" has become almost synonymous with the HTTP protocol, plus WebSockets (which still uses HTTP for the handshake). Wouldn't an innovative browser merely grant us a shell for handling IP/TCP, into which we could drop whatever runtime we wanted? That would enable polyglot programming on the client, and open the door to new categories of software being handled by the "browser". In fact, all software could then be handled by the "browser" as the "browser" would then become the most obvious way to enable desktop software.

For decades now, at least since the 1960s, programmers have been seeking ways to make their software multi-platform. Back in 1990 Patrick Naughton and James Gosling started working on Java, guided by the slogan "Write once, run anywhere". That slogan should still be our goal. What are we doing to move the society forward to the era when that slogan can be true?

In the 1830s and 1840s it was common for the steam engines in locomotives to explode. Early steel bridges sometimes collapsed, because engineers did not know what strain the steel could take. Our society moves forward when important technology becomes so mature and stable that the whole of society can depend on it. At what moment will computers become as reliable as locomotives and steel bridges? And what are we each doing to get us there?

That seems to be a nice-sounding way of saying "breaking something that used to work", which is a significant problem with far too much "innovation" in modern browsers.