These are fair points, but they are missing the things that have changed in the DAWscape.
In 1999, and maybe even 2011, a DAW that couldn't do elastic time was still considered to be a major player. In 2011, a DAW that couldn't do clip launching was still OK. In 2011, the idea of cross-routable modulation (like Bitwig now does) was a wild and crazy idea (mostly). In 2011, the idea of a modular environment within the DAW was floating around (Logic had done it for a while) but was hardly mainstream. These things are now either already or becoming more or less mandatory to be considered a major DAW.
Then there is the matter of control surfaces, which keep evolving. In 2011, the idea of a controller that was essentially a programmable grid of illuminated pads was mostly the domain of experimental performance (Monome). And every month or two, a new controller appears that claims to support Mackie Control Protocol, but has in fact bastardized it in some small but significant way. There's also the touch-based controllers, frequently using OSC, and their constantly changing suggestion of what should be possible.
And then there's plugin APIs, which also keep evolving, and we are under enormous expectation to ensure that every single one of many thousands of little code blobs written using one of a half-dozen plugin APIs and by developers with a range of experience that goes from nothing to world expert should all just work.
Over the years, we've also seen constant changes on Windows in terms of the OS-level audio APIs, each change bringing with it new possibilities and new problems. There's no POSIX for this. And, of course, a slow but steady evolution in audio interface hardware, which doesn't often impinge on the DAW itself, but sometimes does.
As for performant, browsers don't even have RT-style constraints, and there's plenty of observations about "snappiness" in this domain too. It's just a much smaller niche, and developer culture (as testified to by posts here on HN) is now very dominated by web-centric thinking and experience. As a result those sorts of things aren't really part of the culture in the way that "wow, Chrome is so fast" etc. has become. Look at the praise Reaper receives because of its (apparent) ability to handle more plugins with less CPU cycles.
As to why new browsers are rare ... I think it's because (1) browsers are not really fun at most levels, and (2) adoption is hard when there's a default on just about every platform. The fun part is important though. Many developers enjoy "messing around with audio" and it provides a kind of gratification that is rare (video stuff would be similar). It's not that cooking up a standards compliant and crazy fast and beautiful CSS implementation isn't without its joys, but there's so little point doing that as a standalone project. Contrast that with the steady drip, drip, drip of developers who want to try their hand at a synth, or an audio file editor, or FX processor and eventually think "oooh, how about a DAW". There's not really any evolutionary pathway for browser development: you're either writing a whole browser, or you're not. Audio gives you a way in, and then the sky's the limit.
In 1999, and maybe even 2011, a DAW that couldn't do elastic time was still considered to be a major player. In 2011, a DAW that couldn't do clip launching was still OK. In 2011, the idea of cross-routable modulation (like Bitwig now does) was a wild and crazy idea (mostly). In 2011, the idea of a modular environment within the DAW was floating around (Logic had done it for a while) but was hardly mainstream. These things are now either already or becoming more or less mandatory to be considered a major DAW.
Then there is the matter of control surfaces, which keep evolving. In 2011, the idea of a controller that was essentially a programmable grid of illuminated pads was mostly the domain of experimental performance (Monome). And every month or two, a new controller appears that claims to support Mackie Control Protocol, but has in fact bastardized it in some small but significant way. There's also the touch-based controllers, frequently using OSC, and their constantly changing suggestion of what should be possible.
And then there's plugin APIs, which also keep evolving, and we are under enormous expectation to ensure that every single one of many thousands of little code blobs written using one of a half-dozen plugin APIs and by developers with a range of experience that goes from nothing to world expert should all just work.
Over the years, we've also seen constant changes on Windows in terms of the OS-level audio APIs, each change bringing with it new possibilities and new problems. There's no POSIX for this. And, of course, a slow but steady evolution in audio interface hardware, which doesn't often impinge on the DAW itself, but sometimes does.
As for performant, browsers don't even have RT-style constraints, and there's plenty of observations about "snappiness" in this domain too. It's just a much smaller niche, and developer culture (as testified to by posts here on HN) is now very dominated by web-centric thinking and experience. As a result those sorts of things aren't really part of the culture in the way that "wow, Chrome is so fast" etc. has become. Look at the praise Reaper receives because of its (apparent) ability to handle more plugins with less CPU cycles.
As to why new browsers are rare ... I think it's because (1) browsers are not really fun at most levels, and (2) adoption is hard when there's a default on just about every platform. The fun part is important though. Many developers enjoy "messing around with audio" and it provides a kind of gratification that is rare (video stuff would be similar). It's not that cooking up a standards compliant and crazy fast and beautiful CSS implementation isn't without its joys, but there's so little point doing that as a standalone project. Contrast that with the steady drip, drip, drip of developers who want to try their hand at a synth, or an audio file editor, or FX processor and eventually think "oooh, how about a DAW". There's not really any evolutionary pathway for browser development: you're either writing a whole browser, or you're not. Audio gives you a way in, and then the sky's the limit.