[tpmx]

I still think that was a necessary decision. Jon Von Tetzchner (one of the two Opera founders, and a previous long time CEO) strongly disagreed, then from the outside, but later did the exact same thing with Vivaldi. (Jon is a fantastic mensch, btw. One of the best CEOs I've ever had.)

Google had very purposely raised the bar by putting like 5x-8x more competent engineers than the Opera core (non-platform/UI-specific stuff) team had, working on inventing and implementing random new web standards that they then promptly started using on google.com properties. Think e.g. 500-800 engineers compared to 100. We simply couldn't do the same. Then this ratio started growing until it was obvious that it would eventually become an existential threat.

They used their financial success in one business area (search ads) to become dominant in another area (browsers) in a clever and perhaps not entirely legal way.

[nonbirithm]

The irony was how ages ago everyone was waiting for a browser to come along and unseat IE, which was not a good browser.

Now Chrome works well enough as a browser and has so much market share that it exerts too much control over its users and the Web, and it's unlikely that anything will be capable of obsoleting it in the near future.

[fiddlerwoaroof]

I think the trick might be to make a browser that supports a reasonable subset of html/css/js + some killer feature. (Much easier said than done, though: but some kind of mobile app platform like WeChat might be a good way to get a foothold)

Or just fork webkit/blink and aggressively refactor.

[6510]

The killer feature could be a file type named after the kinds of websites people make. Say .blog or .forum or .store or .social perhaps even .aggregator or .hn (lol) each would work like a straight jacket or like a website on a platform (.blog like blogger, .store like shopify etc) it could be distributed and insanely fast. Every other link just opens in the normal browser.

[fiddlerwoaroof]

It's interesting to think about this sort of thing: interactive forms generated from OpenAPI specs? Some way to plug-in handlers for specific mime-types? etc.

[easton]

This is the dream, honestly. Dynamic form creation from specs. PowerShell has something like it (for a function it will create you a ugly form that you can render), but if you could deliver a OpenAPI spec and have Bootstrap (or whatever) UI on the frontend automatically generated.

[kelnos]

The Swagger UI[0] is more a documentation/exploration tool for APIs, but it shows enough of what you can do that I don't think it'd be all that difficult to generate something that you'd want to put on your own website.

[0] https://swagger.io/tools/swagger-ui/

[girvo]

This is not quite what you mean, but I've used it before to decent success:

https://github.com/rjsf-team/react-jsonschema-form

[lmz]

Surely IE6 already had a "reasonable subset" and devs hated it for holding back progress...

[fiddlerwoaroof]

The main reason I disliked it was it wasn't standards compliant for the parts it did implement: you couldn't write code to-spec and expect it to work on both IE6 and Firefox, without testing each browser separately.

[pjmlp]

Thankfully Google has fixed that problem, ChromeOS is now the stardard that drives the Web. /s

[underwater]

There's a form of regulatory capture where big companies benefit from having more stringent regulations because they are the only entities with enough resources to comply.

It certainly feels like Google is using the same playbook for web standards. They've created enough churn that no one else can ever hope to catch up in implementing those standards.

When Google products stop working on some competitors browser they can simply say "it's just web standards" and feign ignorance.

[captainmuon]

I would very much like to use a browser that didn't support all the random new web standards (like, I don't know, WebUSB) and rendered most "content" sites good enough. You could still just embed a WebKit/Blink window in order to run web apps.

I really liked old Opera and even Edge for their engines that made the web pages feel snappy and somewhat different than the WebKit monoculture.