"The engineering equivalent of a car made of duct tape"
Kind of. But do you have a better proposition?
Kind of. But do you have a better proposition?
I just came up with that in 2 minutes, so it might not be perfect, but you can see how with a little bit of work there's much better solutions than "I check for not-evil domain in list!"
So now, if a website leaks its private key, attackers can exfiltrate cookies from all of its users just by making them open an attacker-controlled link, for as long as the cookie lives (and users don't visit the website to get the rotated key).
> If the web server needs a cookie, it should request one
This adds a round-trip, which slows down the website on slow connections.
> the client can submit again to "reply" to this "request"
This requires significantly overhauling HTTP and load-balancers. The public-suffix list exists because it's an easy workaround that didn't take a decade to specify and implement.
This attack already exists in several forms (leaking a TLS private key, DNS hijack, CA validation attack, etc). You could tack a DNS name onto the crypto-cookies if you wanted to, but DNS is trivial to attack.
> This adds a round-trip, which slows down the website on slow connections.
Requests are already slowed down by the gigantic amount of cookies constantly being pushed by default. The server can send a reply-header once which will tell the client which URLs need cookies perpetually, and the client can store that and choose whether it sends the cookies repeatedly or just when requested. This gives the client much more control over when it leaks users' data.
> This requires significantly overhauling HTTP and load-balancers
No change is needed. Web applications already do all of this all the time. (example: the Location: header is frequently sent by web apps in response to specific requests, to say nothing of REST and its many different request and return methods/statuses/headers).
> The public-suffix list exists because it's an easy workaround
So the engine of modern commerce is just a collection of easy hacks. Fantastic.
An attacker who gets the TLS private key of a website can't use it easily, because they still need to fool users' browser into connecting to a server they control as the victim domain, which brings us to:
> You could tack a DNS name onto the crypto-cookies if you wanted to, but DNS is trivial to attack.
It's not. I can think of two ways to attack the DNS. Either 1. control or MITM of the victim's authoritative DNS server or 2. poison users' DNS cache.
Control/MITM of the authoritative server is not an option for everyone (only ISPs/backbone operators), and according to Cloudflare: "DNS poisoning attacks are not easy" (https://www.cloudflare.com/learning/dns/dns-cache-poisoning/)
> Requests are already slowed down by the gigantic amount of cookies constantly being pushed by default
Yes, although adding more data and adding a round-trip have different impacts (high-bandwidth high-latency connections exist). Lots of cookies and more round-trips is always worse than lots of cookies and a fewer round-trips.
> The server can send a reply-header once which will tell the client which URLs need cookies perpetually, and the client can store that and choose whether it sends the cookies repeatedly or just when requested.
Everyone hate configuring cache, so in most cases site operators will leave it to a default "send everything", and we're back to square one.
> No change is needed.
I was thinking that servers need to remember state between the initial client request and when the client sends an other request with the cookies. But on second thought that's indeed not necessary.
> So the engine of modern commerce is just a collection of easy hacks. Fantastic.
I'm afraid so
Cache poisoning is the easiest method, and contrary to whatever Cloudflare says, it's trivial. The DNS transaction number is 16-bits. All you have to do is flood the shit out of the resolver with spoofed packets and eventually one of the transaction numbers will hit, and your attack is successful. It's low-bandwidth, takes at most a couple hours, and nobody notices. This is one of the many reasons you can't just trust whatever DNS says.
The choice of what HTTP messages to cache is not always a choice, as is the case with HSTS. But it could be made one if testing of this proposal (which again, I came up with in 2 minutes) showed better results one way or another.
But all this is moot anyway cuz nobody gives a crap.
End of random rant.
But then you would loose plattform independency, the main selling point of this atrocity.
Having all those APIs in a sandbox that mostly just work on billion devices is pretty powerful and a potential succesor to HTML would have to beat that, to be adopted.
The best thing to happen, that I can see, is that a sane subset crystalizes, that people start to use dominantly, with the rest becoming legacy, only maintained to have it still working.
But I do dream of a fresh rewrite of the web since university (and the web was way slimmer back then), but I got a bit more pragmatic and I think I understood now the massive problem of solving trusted human communication better. It ain't easy in the real world.
This all just drives a need to come up with ever more tacked-on protection schemes because browsers have big targets painted on them.
Anyway, in your scenario the controller would be essentially a one off and you'd be better off writing a native app to interface with it for the one computer this experiment will run on.
And that's before realizing it's already a bad idea with existing devices because they were never designed for giving untrusted actors direct access.
Not unlike the programming language or the app (growing until it half-implements LISP or half-implements an email client), the browser will grow until it half-implements an operating system.
For everyone else, there's already w3m.
WebUSB I don't use or would miss it right now, but .. the main potential use case is security and it sounds somewhat reasonable
"Use in multi-factor authentication
WebUSB in combination with special purpose devices and public identification registries can be used as key piece in an infrastructure scale solution to digital identity on the internet."
You have sites now that let you debug microcontrollers on your browser, super cool.
Same thing but with firmware updates in the browser. Cross platform, replaced a mess of ugly broken vendor tools.
Your micro-controllers should use open standards for their debugging interface and not force people to use the vendor website.
Yes. Regards, CIA, Mossad, FSB etc.
You remove that, and videoconferencing (for business or person to person) has to rely on downloading an app, meaning whoever is behind the website has to release for 10-15 OSes now. Some already do, but not everyone has that budget so now there's a massive moat around it.
> But do we need e.g serial port or raw USB access straight from a random website
Being able to flash an IoT (e.g. ESP32) device from the browser is useful for a lot of people. For the "normies", there was also Stadia allowing you to flash their controller to be a generic Bluetooth/usb one on a website, using that webUSB. Without it Google would have had to release an app for multiple OSes, or more likely, would have just left the devices as paperweights. Also, you can use FIDO/U2F keys directly now, which is pretty good.
Browsers are the modern Excel, people complain that they do too much and you only need 20%. But it's a different 20% for everyone.
Same as your camera/microphone/location.
If you give USB access, it is not really a website anymore, rather a app delivered through the web. I don't see a fundamental difference in trust.
I rather am able to verify the web based version easier and I certainly won't give access to a random website, just like I don't download random exes from websites.
Performance is lower, yes and well ... like I said, it is all a big mess. Just look at the global namespace in js. I still use it because of that power feature called plattform independence. What I release, people can (mostly) just use. I (mostly) don't care which OS the user has.
But do we need audio, images, Canvas, WebGL, etc? The web could just be plain text and we’d get most of the “useful” content still, add images and you get a vast majority of it.
But the idea that the web is a rich environment that has all of these bells and whistles is a good thing imo. Yes there’s attack surface to consider, and it’s not negligible. However, the ability to connect so many different things opens up simple access to things that would otherwise require discrete apps and tooling.
One example that kind of blew my mind is that I wanted a controller overlay for my Twitch stream. After a short bit of looking, there isn’t even a plugin needed in OBS (streaming software). Instead, you add a Web View layer and point it to GamePad Viewer[1] and you’re done.
Serial and USB are possibly a boon for very specific users with very specific accessibility needs. Also, iirc some of the early iPhone jailbreaks worked via websites on a desktop with your iPhone plugged into usb. Sure these are niche, and could probably be served just as well or better with native apps, and web also makes the barrier to entry so much lower .
I think the giant major downside, is that they've written a rootkit that runs on everything, and to try to make up for that they want to make it so only sites they allow can run.
It's not really very powerful at all if nobody can use it, at that point you are better off just not bothering with it at all.
The Internet may remain, but the Web may really be dead.
But people do use it, like the both of us right now?
People also use maps, do online banking, play games, start complex interactive learning environments, collaborate in real time on documents etc.
All of that works right now.
What do you mean, you can run whatever you want on localhost, and it's quite easy to host whatever you want for whoever you want too. Maybe the biggest modern added barrier to entry is that having TLS is strongly encouraged/even needed for some things, but this is an easily solved problem.
I don't see how that solves the issue that PSL tries to fix. I was a script kiddy hosting neopets phishing pages on free cpanel servers from <random>.ripway.com back in 2007. Browsers were way less capable then.
It's not even broken as the edge cases are addressed by ad-hoc solutions.
OP is complaining about global infrastructure not having a pristine design. At best it's a complain over a desirable trait. It's hardly a reason to pull the Jr developer card and mindlessly advocate for throwing everything out and starting over.
We live in world where whatever faang adopts is de facto a standard. Accessible these days means google/gmail/facebook/instagram/tiktok works. Everything else is usually forced to follow along.
People will adopt whatever gives them access to their daily dose of doomscrolling and then complain about rather crucial part of their lives like online banking not working.
> And of course, if the new solution completely invalidates old sites, it just won't get picked up.
Old sites don't matter, only high-traffic sites riddled with dark patterns matter. That's the reality, even if it is harsh.
Try 90s! We had to fight off ActiveX Plugins left and right in the good olde Internet Explorer! Yarr! ;-)
This might be what's needed to break out of the current local optimum.
which is still much too new to be able to shut down the PSL of course. but maybe in 2050.
CORS lets sites define their own security boundaries between subdomains, with mutual validation. If you're hosting user content in a subdomain, just don't allow-origin it: that is a clear statement that it's not "the same site". PSL plays absolutely no part in that logic, it seems clear to me that it's at least in part intended to replace the PSL.
Do other sites (like google's safety checks) use CORS for this purpose? Dunno. Seems like they could though? Or am I missing something?