I think your use of "thin client" here is literally redefining a thin client to be the opposite. The thick client software of MS Word does all the things locally, which is part of what makes it responsive.

What could have made it responsive.

To be competitive with SaaS, open source software needs to have a web version because there’s so much less friction to use web software compared to desktop software. So desktop thick clients with native UI need to be maintained in addition to, not instead of, a web UI. Most open source options barely have the resources to build a web UI, asking also for a thick client for each user operating system platform seems unrealistic.

OSS is unfortunately an overloaded term. But assuming it is about FOSS, than software that runs on the user’s machine gives so much more freedom. Can software even be free for a consumer, when it is really unpractical to run it yourself, because it needs a server?

Then you have to deal with automatic updates, multiple platforms, data synchronisation, etc.

Rent-charging isn't the only reason to use the web, especially for an application like this.

While you have valid points, from a dev perspective, desktop UX is just as much shifting sand and land mines, which fuels the move to the browser for better or worse

Ah yes, another web-grumbler, here to tell us to abandon the connected platform available on all devices that is hugely successful.

It's just your opinion that the web isn't great for UI. It doesn't have much evidence (and you certainly aren't making any testable assertions). If the web was so bad you'd think native apps would be ruling the planet, would be vastly more popular & loved & be what everyone is trying to ship. But that's not the case. The web is everywhere and companies have full faith in the web and it's working.

> But the web isn't great for building input UX (it's all bloat + jank and you have to completely reimplement the input elements to do anything useful, on top of shifting sand)

Theres a school of thought that wants a complete & total solution, handed to them, that does everything, that everyone uses.

These people are ideologically going to hate the web. Even though there are countless really really good well made component libraries out there, the fact that it's not all baked in upsets a certain type a lot. But, this is the strength of the web. The web is free to shift & change & emerge new patterns & ideas, is free to have new and different components emerge. This is a strength, this is why the ecosystem keeps steamrolling everyone else: everyone else is singular & tries to be complete, and the web is legion & perpetually innovating on the edge in a way classic platforms never can. Some people see weakness, but the ideas of the low level Web Extensibility Manifesto keep the web buoyant & iterating for developer and user happiness in ways no other system has remotely been able to achieve. https://github.com/extensibleweb/manifesto

The other good news is that the OpenUI community group got formed & is upstreaming into HTML more components. Popovers and dialogs are both about to be fully usable by anyone. There's a ton of drive to improve the core elements available on the web. While also not harming that extensibility. https://open-ui.org/

> the data can be shared or backed up somehow sure; that's just file hosting with a collaboration layer maybe.

Being connected software is the new base expectation, is table stakes. You have to have connected software to succeed. And the desktop doesn't have this. It's all special snowflake systems, all build it yourself. It's the default for software to be connected on the web. And on the web it's the default that it'll run on you desktop, your phone, your tv, your fridge, and all your neighbors friends and coworkers device.

For something like a notion clone here, I think most users would have an enormous enormous expectation that this kind of groupware be on the web. Trying to reinvent a collaborative knowledge capture paradigm on desktop sounds absurd & sounds like folly. The desktop has no starting place for developers nor users.

When I think of performance specifically, I think of how much opportunity the web has. Any popular software platform is going to have an absolutely gobsmackingly huge amount of really bad software written for it. And that challenge is magnified greatly in the connected age where vastly more data lives outside of your computer than on it, where your computer is at best a local cache. We will always be beset by unideal software, by bad architectures, by poor data remoting practices. But that doesn't make me sad or want to go run away to the many different native platforms, to go trip over all the same data syncing & data architecture problems there. It makes me excited to see us iterate and improve, makes me so happy there are brave people cranking out Svelte and Stencil and Lit, makes me happy we are moving forward speccing out signals and observables (although having two similar but different async data constructs maybe is a misfire, IMHO), makes me happy we are getting wasm and still so early but wading slowly off the main thread and into workers. There's so much opportunity here to iterate & improve & grow, even with what we have, and there's been such strong help & improvement ongoinglg given to the platform layer from such a wide coalition of spec authors and passionate wicg.io community members. World wide web on & on.