UI automation is necessary for accessibility [LWN.net]

UI automation is necessary for accessibility

Posted Jul 13, 2026 11:59 UTC (Mon) by farnz (subscriber, #17727)
In reply to: UI automation is necessary for accessibility by paulj
Parent article: An update on the scraper situation

Your last paragraph is self-contradictory; legal action involves a centralized "recovery" authority (the legal system) getting involved to force the reversal of a payment, but you're also saying that the existence of that authority is a malfeature that should not exist. It can't be both.

Without it, though, you face the fraud problems; sophisticated users can avoid being bitten too badly by fraud, but you're going to have to solve the mass-market chicken-and-egg problem somehow. Most people won't risk their money without a decent chance of a high reward, and without most people being willing to use whatever "fraud is your problem" payments network you create, sites have no incentive to require people to use it (since by doing so, they lose all their users).

And that, in a nutshell, is the issue that any micropayments system has to solve somehow. How do you bootstrap popularity of a micropayments system if payees won't come on board en-masse until there's a significant number of people willing to pay with it, and payers won't come on board until there's either very low risk of fraud, or payees have signed up en-masse?

After all, if there is a system that is completely viable for small payments, there would be a site like Patreon, Ko-fi, Substack, or similar that accepts it - but none of the big "pay your content creator" sites do accept alternatives to cards.

to post comments

UI automation is necessary for accessibility

Posted Jul 13, 2026 12:14 UTC (Mon) by paulj (subscriber, #341) [Link] (15 responses)

To be clear, I wrote in reply to the possibility /micro-payment system/ must have some kind of recovery mechanism, even though (as you stated) that is technically very hard to solve, which kind of falls out of your comment.

I agree, it is technically hard to solve, and I agree with you it is a social problem. So I'm pointing out that there are micro-payment systems which just ignore that problem, and it is indeed left as a social problem to solve. In which case problem #3 is solved - it can't happen. Problem #2 is potentially solvable, depending on context. Problem #1 is partly solved - the resources were paid for, and the server of the resources was not (financially) harmed.

The implication in your comment was there was a problem that the micro-payment system itself had to solve somehow, even though it was hard to solve / unsolvable at the technical level. I'm saying micro-payment systems definitely already exist, with reasonably non-trivial use, and they do so without having tried to tackle the problem you (possibly) implied needing solving.

I.e., the micro-payment system can go ahead and be used, and that social issue of recoverability can be left for later solving (by probably a plethora of different social solutions, which may evolve). It's not an obstacle as such.

I can pay for a variety of things (digital services particularly) with said micro-payment system. I can buy (e)SIMs, VPNs, email hosting, VPs, Proton accounts and other similar services, AI hosting, contract developors, donate to various orgs (inc. FSF), etc., etc.. It's already bootstrapped and being used. Maybe it's not perfect enough for you, but it's good enough for a chunk of users.

UI automation is necessary for accessibility

Posted Jul 13, 2026 12:17 UTC (Mon) by farnz (subscriber, #17727) [Link] (14 responses)

If you don't solve these problems to a point where, as an unsophisticated user, my funds are not at risk if I'm hacked, then you're limited to being something for sophisticated users, which means that the web as a whole can't use the system, since most users aren't sophisticated.

Any such system that's good enough for mass adoption would be in a good position to replace PayPal, Venmo, Patreon and Ko-fi; once you've done that, then you've got a system that's viable for the mass market. Until then, you have a minority use system that's suitable for sophisticated users, but excludes most of the world.

UI automation is necessary for accessibility

Posted Jul 13, 2026 12:52 UTC (Mon) by paulj (subscriber, #341) [Link] (13 responses)

Nothing would stop a PayPal like entity acting as a gateway to whatever underlying micro-payment system. If a PayPal like frontend is what is needed for "normal" users, nothing technical stops that.

The fundamental issue, as you already stated, is that the blocker is a social issue. Though, it's just /not/ the one you stated. Rather, the reason we can't have a global, ubiquitous, universal access, distributed micro-payments systems is the same reason that the existing PayPals, Venmos, etc., can't act as a gateway to such a system: The governments of the world do not want it - they only want something they can control.

And so long as there are different and unaligned blocks of states in the world, we can never have that global, ubiquitous, universal access system.

And even if all were aligned, they would not allow us to have a /private/ system, and quite probably also not a universal-access system.

However, at the technical level, the micro-payments system(s) are built and are in-use. I wish LWN would integrate something with the accounts system, so I could add "Pay for LWN subscription" to my list of digital stuff I can pay for with distributed Internet money.

UI automation is necessary for accessibility

Posted Jul 13, 2026 13:00 UTC (Mon) by daroc (editor, #160859) [Link] (1 responses)

There are almost certainly complications on our end to receiving payment in that way, but I don't think we would be opposed in principle — letting people pay us is always a good idea. The difficult part would probably be integrating things with our payment processor.

Could I ask what form of distributed internet money you prefer, just so that I know what to look for when investigating the possibility?

UI automation is necessary for accessibility

Posted Jul 13, 2026 14:33 UTC (Mon) by paulj (subscriber, #341) [Link]

My preferred one would be Monero. It's technically beautiful. It's private. Transaction fees are low enough to be viable for ¢ level payments. BTC is obviously the big one - but higher fees.

I don't know about your payment processor. There are open-source distributed Internet money payment processing platforms you can self-host, and integrate yourself into your site. E.g., BTCPayServer, MoneroPay (moneropay.eu) - I don't know how good those are. There are also service providers, like mixpay.me. Some of the big payment processors are starting to get into accepting some crypto payments - but usually heavily constrained by jurisdiction, which may make it less attractive for the payer even if it might be more convenient for the payee.

UI automation is necessary for accessibility

Posted Jul 13, 2026 13:10 UTC (Mon) by farnz (subscriber, #17727) [Link] (10 responses)

But that in turn is the core of the problem - there's nothing unsolved in the technical arena there.

What we lack is the PayPal-like frontends that make it usable for the ordinary person - and I don't buy your "government blocks it" conspiracy theory, given that entities like Coinbase, Crypto.com, Binance, Kraken and others are perfectly permitted to trade in various cryptocurrencies, and there exist entities like MetaMask on top that also supply cryptocurrency wallets in the open and legally.

UI automation is necessary for accessibility

Posted Jul 13, 2026 13:59 UTC (Mon) by paulj (subscriber, #341) [Link] (9 responses)

It's not a conspiracy theory. Governments will not allow global, universal-access distributed payments system to integrate easily with their own financial systems (inc. the kinds of retail-user friendly entities you gave examples of). Firstly, because there are a number of unaligned blocks of states, and at least 2 of those blocks wish to heavily limit (if not entirely block) access to/integration with the payments system in their own blocks to certain other blocks/states. Secondly, because nearly all are implacably opposed to entities/people within their blocks having access to distributed, universal-access payments systems.

Those are not conspiracy theories, those are basic facts of regulations that governments have brought in and are bringing in more and more regulations to strengthen.

All the entities you mention are legally or de facto unavailable to many people across the world. MetaMask is open-source and can't be stopped even if Consensys disappeared tomorrow, but it does not integrate with the traditional financial system either (AFAIK).

We can not have a global, universal access, distributed micro-payments because of those social reasons. What we have today are a number of 'islands' of gated-access micro-payments systems, with high-friction or even fully-blocked channels between them; and running in parallel alongside those systems are a couple of much smaller - but growing - truly distributed, universal-access micro-payments systems. At best, if somehow all those blocks were aligned, we could have a global, semi-distributed, *gated-access* micro-payments system, given the stance of many governments; but then we'd still have the truly distributed, universal-access micro-payments systems running alongside them.

UI automation is necessary for accessibility

Posted Jul 13, 2026 14:02 UTC (Mon) by farnz (subscriber, #17727) [Link] (8 responses)

But, while we have these entities, they're worse for unsophisticated users than things like PayPal (which is also not a "global, universal-access payments system" by your standard, since it's blocked in some states).

And I'm saying that unless you can displace at least one of PayPal, Ko-fi, Patreon or Venmo, you've not got a system that's suitable as a micropayments system; you've got something that sophisticated users who can manage their own fraud risk can use, but not something suitable for the mass market.

UI automation is necessary for accessibility

Posted Jul 13, 2026 15:12 UTC (Mon) by paulj (subscriber, #341) [Link] (7 responses)

> unless you can displace at least one of PayPal, Ko-fi, Patreon or Venmo, you've not got a system that's suitable as a micropayments system;

The point is we have no universal micropayments system because we have high regulatory friction.

A universal micropayments system would not displace the examples you give, rather those examples would become gateways to such a universal system - in addition to allowing those who wished to directly interact with said universal micro-payments system with software under their own control.

That none of those examples you give easily be integrated with (and are not universally accessible - e.g., I don't think I can use Venmo, and possibly you can't either if I remember correctly about where you're based), is because of a lack of a suitable universal micro-payments system. Which is cause of high regulatory friction (including severe fragmentation). Which is a social problem. ;)

UI automation is necessary for accessibility

Posted Jul 13, 2026 15:19 UTC (Mon) by paulj (subscriber, #341) [Link]

And to be clear: I can't buy an eSIM with Venmo - cause I can't access Venmo. Even if I have access to something like PayPal, many other people do not and it could well be that the person/entity I want to make a financial transaction with does not have access.

For a large chunk of the world, it is very difficult to conclude financial transactions with them, via the "retail friendly" way - and the methods available are not suitable for micro-payments (and are not recoverable either!). For a certain section of the world there is no available method.

*Other* than the distributed Internet money methods - those work well. And the friction of the other methods is such that even many "normal users" will be willing to overcome the UI-overheads. We /could/ easily have the retail friendly UIs, and integration with the types of entities you list, if not for the high regulatory friction.

So it's all social, as far as I'm concerned.

UI automation is necessary for accessibility

Posted Jul 13, 2026 15:31 UTC (Mon) by farnz (subscriber, #17727) [Link] (5 responses)

You're looking too far into the future - I'm saying that if there exists a system that's suitable for micropayments, there would be a Ko-fi, PayPal (friends and family only - I'm not expecting the full gamut of PayPal services to be taken over by it), Venmo, Zelle, or similar service competitor built around that system, and demonstrating that the system is suitable by being a friend-to-friend money transfer system for people in the USA (only - I'm not expecting this to be a global system) that outcompetes the existing payment systems.

If you can't build a system that's better than the existing payment systems for one country, then your assertion that your system is better than the existing payment systems is suspect; why hasn't someone used the network to outcompete the existing players?

UI automation is necessary for accessibility

Posted Jul 13, 2026 16:10 UTC (Mon) by paulj (subscriber, #341) [Link] (4 responses)

Well,

a) even within a country there is very high regulatory friction (true for at least 2 of the biggest blocks in the world)

b) despite that friction, there are a number of places that do accept universal, distributed, micro-payments. And I find it as easy, if not easier, to use than the "retail friendly" stuff.

E.g., I topped up my kid's eSIM and I can just open the app on my Linux desktop, type my password (optional - my desktop is already obviously secured, but belt and braces), copy the address and amount from the payer's site (or the payer's processor) and send the payment. And it clears within a few minutes.

With the "retail friendly" stuff you prefer, the flow is similarish, but I type stuff (CC card) into the processor form, plus I have to type more stuff to satisfy algorithms, then I have to wait while it redirects through 1 or more websites - which sometimes fails - then I have to open an app on my Android phone, so I have to authenticate to my phone, then open the app and authenticate on it with another PIN, then I have to confirm the push notification - which sometimes will have failed to arrive. Then I have to go back to the website and click on a button there, and then wait for it to redirect back from the processor iframe or page to the original site. Also, sometimes, for whatever reason, the bank decides something is suspicious and blocks stuff and I have to phone them - sometimes that can take quite a while. Sometimes I'm abroad, and this can be a real PITA because of timezones (plus, cost of international phone calls - maybe I was buying the eSIM cause that was my Internet and phone access!).

I don't really see the major UI win for the "retail friendly" thing, the distributed Internet money micro-payments thing is actually a /lot/ easier and _far more reliable_ for me. But YMMV.

Also, because I want to have a phone that isn't a complete blob of god-knows-what crap from some big mobile company and all the companies that paid them to push their crap into said blob, and instead run a bit-more-open/far-fewer-blobs OS on my phone, this "retail friendly" app sometimes has failures after updates and I have to spend an hour on the phone to the bank to get them to clear some weird failure condition that so I can reauthenticate my phone. But that's another discussion.

Finally, the internet, distributed micro-payments thing is still very young. It is kind of normal that early technology still doesn't have optimal UI or adoption. If "it must have optimal UI and UX" is a precondition for believing anyone should use it, then... it's catch-22 (esp. if a great part of the UX hinges on adoption).

Again, perfect is the enemy of good. Despite the lack of perfection, there are many use-cases where distributed Internet money beats the "retail friendly" stuff. And in time, it can go from "good enough for these use-cases" to "good enough for more", and with enough time, maybe we see enough adoption that the regulatory friction becomes unsustainable (particularly as our generation, which is driving the regulation because of vested interests, ages out of holding the reins of power).

UI automation is necessary for accessibility

Posted Jul 14, 2026 8:47 UTC (Tue) by farnz (subscriber, #17727) [Link] (3 responses)

You are, however, by definition a "sophisticated user" taking responsibility for your own fraud and security risks.

The whole point of regulation is that unsophisticated users demonstrably don't have the knowledge and experience to take care of those risks, so we push them to the payment networks to handle. And that's why I'm focusing on the "why isn't there a retail platform suitable for unsophisticated users?" - because if you can't do that, you've not got a solution to the hard problems in payments, just to the easy ones.

UI automation is necessary for accessibility

Posted Jul 16, 2026 6:32 UTC (Thu) by raof (subscriber, #57409) [Link] (2 responses)

And, to amplify that point, the “regulatory friction” preventing this hypothetical universal payment system from becoming mainstream is the technical solution to the fraud and security risks that a mainstream system requires. I am, in fact, reasonably convinced that it is approximately the only feasible technical solution — that any solution to these problems will look a lot like the current regulatory environment, perhaps with some actors shuffled around.

I've certainly never seen a serious attempt to solve these problems which would not rapidly devolve into something looking like the current arrangement (but generally worse).

UI automation is necessary for accessibility

Posted Jul 16, 2026 8:46 UTC (Thu) by paulj (subscriber, #341) [Link] (1 responses)

> I've certainly never seen a serious attempt to solve these problems which would not rapidly devolve into something looking like the current arrangement (but generally worse).

The growing ecosystem of digital services using various kinds of decentralised, distributed Internet money would disagree with that.

UI automation is necessary for accessibility

Posted Jul 16, 2026 10:53 UTC (Thu) by pizza (subscriber, #46) [Link]

> > I've certainly never seen a serious attempt to solve these problems which would not rapidly devolve into something looking like the current arrangement (but generally worse).

> The growing ecosystem of digital services using various kinds of decentralised, distributed Internet money would disagree with that.

"that growing ecosystem of digital services" has _never _ seriously attempted to solve those problems. Quite the opposite, in fact.