Also ActiveX, which brought a lot of "desktop only" capabilities to the web, at the cost of the ultimate lock-in.

We should probably be glad that Microsoft basically stopped all IE investment when the dot-com crash happened (IE6 came out 2001), if they had kept investing the internet might still be dominated by MS

Probably had more to do with the DoJ antitrust suit than the dot com crash. Microsoft approached anything that sniffed of antitrust violations with trepidation for a good 5-7 years.

But yes, ultimately a good thing.

Yes, this is often forgotten. I've always assumed it's the biggest reason for Microsoft basically doing less & being more or a cash cow during the Ballmer years (similar to Apple's current philosophy with a CFO as CEO).

Though the phone fail was a huge fail. Those initial Windows phones were really awesome. These were the ones that competed with Blackberry before Apple took over the market.

Microsoft phones only died because they were bad at making an App store. Microsoft made the exact same mistake with the later Zune HD. Despite having a massively superior IMO experience for the user, Windows phone was doomed.

The lock-in is mostly because competitors were not good enough to implement a similar solution, not because of artificial restrictions.

As an example of that lock-in... It's 2023, and my Fortune 250 corporate laptop image still has IE (along with Edge, Firefox, AND Chrome), because we STILL have web applications that were tied to ActiveX, and still need it.

As crazy as that sounds, IE had a better development team until IE6 basically killed off its competition. It was more stable than equivalent browsers, it had more features, I think it was even developed faster than competitors.

Which sounds a lot like Chrome, except closed source. Those considering how big and cumbersome browsers are... does open source that matter that much in the grand scheme of things? Just browser complexity is a HUGE moat, there are probably 5-10 organizations that can sustain leading edge browser development 10+ years.

OS complexity isn't much of a moat, linux works fine. The "problem" is that the browser projects are organised as big monoliths with limited internal competition for the components. Although it isn't really a problem yet. Chrome works fine, and switching web browsers isn't hard to do.

But if community-developed web browsers were required the complexity wouldn't slow the process down much. It just happens Google does a really good job.

> OS complexity isn't much of a moat, linux works fine.

I said browser complexity, not OS complexity.

What difference do you see between a browser and an OS? A browser is an OS. They perform the same role. A context to run whatever 3rd party software.

If a project is too complex for a single organisation, then break it up into 2-3 projects. That is how OSS software handles complex tasks. Someone could do a real LLVM-style browser project then we don't all have to use javascript which'd catapult us all into the 21st century.

> What difference do you see between a browser and an OS? A browser is an OS. They perform the same role. A context to run whatever 3rd party software.

This is where everything has gone screwy. A browser should only be rendering and displaying html files. Browsers should be minuscule in size, and /definitely/ not run any 3rd party software.

> This is where everything has gone screwy. A browser should only be rendering and displaying html files. Browsers should be minuscule in size, and /definitely/ not run any 3rd party software.

You can try to close the door on that stable, but the horse that escaped has died of old age it's happened so long ago.

Says who? This sounds more like preaching than anything else.

Blame the OS vendors (Microsoft has 70% of the blame, Apple 25% and even Linux distro share a 5% amongst themselves).

They killed[1] cross platform apps and they made software updates hard, so everyone just flocked to the web.

[1] Not through a single blow, death by a million cuts.

>They killed[1] cross platform apps

No they didn't! Extremely lazy devs deciding that they should just be able to deploy javascript and html anywhere and using Electron "killed" native apps. Why would anyone ever invest even 1% more into a native app when you can just hire intern javascript kiddies, pay them terribly, oversee them with one or two expensive javascript bros, and all your users will blame the slowdown on the operating system and Microsoft as they always have?

Native apps cost far more than just 1% over web apps.

I'd argue each native version costs about as much as an web app.

The real shame was Gtk and Qt not getting the support they deserve.

>This is where everything has gone screwy

I, on the other hand, find that "browser as OS" is one of the best things that has happened in recent software history. I can run Photopea anywhere, on any platform that has a decent browser. Absolutely magic!

The web was invented in 1989, became a public project in 1991, and JavaScript support for it appeared in 1995, Java Applets the same year, with Flash and ActiveX following in 1996. So it was a static document model for ~6 years, and an app platform for ~27 years.

PDFs and work documents also were infected with scripting languages in the 90’s. It was just a weird time.

Kind of sounds like EA Sports Madden.

The IE block model was better though, and everyone is using it today with box-sizing: border-box. I still maintain that the pragmatic and sensible thing to do was to just modify the standard – "best viewed in IE" was a self-inflicted wound.

Also remember the context: Netscape was just as gung-ho with standards as IE and everyone was kind of figuring out what did and didn't work on the fly. To just ossify a standard at some point just made no sense. Good standards respond to what's going on in reality.

The standards are there for a reason (interoperability), so following them makes sense even if a browser vendor reasonably thinks it should have been done differently. If everyone just implemented what they think is best, no standard would be required in the first place.

Apart from that, IE6 also had massive rendering bugs and no CSS2 support. They only fixed the worst stuff five years later, with IE7. Of course web developers still had to wrestle IE6 for years after that, as its market share declined only slowly.

In this case, the standard reduced interoperability. As a general point you're correct of course, but you need to consider the context and specific situation.

No one implemented standards back then, and IE WAS the de-facto "standard" as it had something like ~98% market share. I'm not saying standards should have been made bug-compatible with IE or even that they should have gone out of their way to accommodate IE, but stuff like the box model which was 1) actually better (or, at the least, equivalent), and 2) a huge pain for everyone could have just been in a "CSS 2.2".

It would have reduced headaches for everyone, especially non-IE users as the people who suffered the most from this weren't the IE users and devs who only cared about IE, they were the Phoenix/Firebird/Firefox and Opera people and devs who cared about them. One has to wonder if uptake of things like Firefox or Opera wouldn't have been much faster in an alternative universe where these changes were made.

Mindlessly following standards just makes no sense. Also see GNU and POSIX_ME_HARDER/POSIXLY_CORRECT.

Sure, and the XMLHttpResponse idea led to everything we have today. But lots of other things didn't work. document.all wasn't any better than Netscape's document.tags but they kept at it until they finally got around to supporting standards. They had major long-term bugs in how margins were figured. The layout model struggled if you used absolute positioning between two elements that were not positioned. Those are just the scars I still have, there were plenty more.

> The JS and CSS and HTML box model worked differently

This was required for compatibility with Netscape. Developers coded for the Netscape box model when it was dominant, so Microsoft had to follow the same model to prevent the pages from breaking when viewed in Internet Explorer.

Neither Mozilla nor Opera did that, so apparently they didn't have to.

> Neither Mozilla nor Opera did that, so apparently they didn't have to.

Both of them did that. The box model in question was invented by Mozilla (internal code name for Netscape), and Opera 7 also implemented it.