On the other hand, Android phones only support it if they are Japan-region phones.
Source: I had an iPhone 7, and was friends with one of the engineers who added FeliCa support to the secure enclave. The Japanese 7 was a one-off until the 8 made it ubiquitous.
Patents hold it back on Android. Apple just got themselves a licensing deal that apparently lets them just have it on all phones.
They could have a store value card that works worldwide. Effectively bringing Octopus or Suica card to the world.
I even remember there were proposal for future version of Felica that works under 10ms. Although I cant seem to find any reference of it.
There is also a new NFC proposal ( again my google fu is not helping ) that allows multiple NFC card / tags to be read at the same time. So we can do Membership card, discount card and payment card all in one. And hopefully someday we could have electronics receipt right on our phone alongside the payment record as well. Something I wished Apple have done for the past 10 years but they still haven't gotten Apple Card or Apple Cash to work outside of US.
https://en.wikipedia.org/wiki/Nationwide_Mutual_Usage_Servic...
https://www.justice.gov/archives/opa/pr/justice-department-s...
Japan has competition in payment systems. Paypay, D-Pay, Meri-pay, Line Pay, Rakuten Pay, etc... Each tries to entice both customers and retailers by offering discounts and bonuses.
Also I'm happy to pay cash as it's private.
Even here cash is coming back into vogue as costs are pushing small businesses to evade taxes.
I was in Tokyo last week and it was similar businesses i.e. mostly smaller and in lower margin industries.
Needless to say, I prefer to use cash in Japan.
When shopping, I prefer to use the Suica app in my iPhone as it’s just a quick touch, but some stores won’t accept it so I have to use the Nanaco app—which requires a face recognition step—or pay in cash. I haven’t bothered to set up a QR code app yet. Twice, when I tried to install PayPay, the most common one, I got stuck on an authentication step and gave up.
Even shops within the same department store accept different combinations of payment systems. In my local Takashimaya, I can use Suica to buy food in the basement but not in the restaurants on the upper floors. Shops in the nearby Sogo Department Store do not accept Suica, only Nanaco, except for the Starbucks on the third floor, which does accept Suica.
Convenience stores seem to accept nearly everything, as you can guess from the number of logos on this sign:
https://news.mynavi.jp/article/osusumecredit-107/images/003l...
Some relatives are arriving in Japan next Tuesday for a three-week visit, and they asked me what they should do about credit cards, digital money, cash, etc. I realized that, despite living here, I barely understood the situation myself, so I had Gemini Deep Research prepare a report for them:
https://docs.google.com/document/d/1WY4AM0mJS94uwPMK8XjIQMLf...
Another point to add is that in the 1980s and 1990s big security problems emerged with the magnetic cards that were used widely then for transportation and telephone calls. Here is what Perplexity has to say about that:
https://www.perplexity.ai/search/i-want-some-information-in-...
I well remember the “open street markets in urban areas like Shibuya ... known for selling counterfeit cards.”
I've only been to Japan very briefly, but between an international credit card (Visa or Mastercard) and a Suica in Apple Pay, I never had any issues.
Maybe this is due to having been only to very touristy areas, but almost everywhere I had to pay took at least one of the two, and thanks to being able to top up the Suica directly in Apple Wallet from my credit card, running out of a balance there was never an issue.
Of course you'll need some cash for some smaller food stalls and shops, but getting that wasn't any issue either: ATM fees didn't seem particularly high as far as I remember (and my bank reimburses them).
I was fascinated by the sometimes dozens of different contactless (stored value and otherwise) and QR code based payment methods accepted at some stores, but with the exception of a few vending machines (that also took cash), I don't think I've seen any place taking only these but not also either Suica or Visa/Mastercard.
Quite the opposite, actually – ubiquitous ice cream vending machines on train platforms that accept the same card for payment people need to get into the station in the first place seem like a real health hazard :)
This is a fun thing to keep in mind when people tell you Japan is a "homogenous society".
(It's not high-trust either.)
I really, really don't see how that connects to... payment systems. Surely you're not trying to claim that what people actually mean typically is that Japan is homogeneous in every arbitrary way possible, right? That's a bit too much even for a strawman, even if I factor in the kind of forum this is.
At least in some cases it is sufficient to change the phone SKU id (which requires temporary rooting) to the Japan SKU id to unlock the Osaifu-Keitai functionality on a non-Japan phone. I'm not sure if this means that the secure element had the necessary keys provisioned all along, or just that the Osaifu-Keitai app then provisions it on first use.
Source - I’m sitting in Kyoto right now having travelled all over Tokyo and then on to Kyoto using only my phone to interact with Japan Rail. Verified with two 16Es and a 12. In fact we were able to add the Suica cards to our phones and charge them fromApplePay while still stateside. That let us skip the Welcome Suica line at Haneda and go straight to the monorail. Highly recommended
Many tourists already get intimidated by language barriers and inscrutable timetables and transit routes; add tariff complexity (and the chance of getting charged with fare evasion!), and many just end up taking a taxi.
With these stored-value systems, you pretty much don't have to ever worry about doing anything wrong as long as you properly tapped in (and have the necessary cash to top up the card at the exit turnstile if you ended up "overdrafting" your card).
Now the remaining complexity is learning about and getting the card in the first place, and Apple Wallet really does an amazing job there in Japan. Not even having to install an app or create any account is absolutely amazing.
Now it's a fiddle with an app, then try to get the right angle on a smudgy reader. Getting onboard takes much longer and it feels like technology sent us 2 steps back instead of forward.
IC cards are better, and if they could be integrated in the phone then it would be even better and faster for everyone.
Forcing use of an app and QR codes does seem like a significant step back, although I guess it makes paper tickets much easier to implement with the same scanner.
Actually other western transit system cards are similar to Japan's. For example in Paris the transit passes (Navigo cards) are stored-value systems and hold a record of the last few scans (three I think). You can also read them with any NFC smartphone and see what's stored inside. The tap at the entrance of a transit vehicle is near instant as the reader doesn't need to interact with a backend since everything is stored on the card itself.
And a good chunk of vending machines in Japan accept the IC card. Sometimes even food shops a step above vending machines.
Magnetic tickets are already slower than IC cards, and are both more expensive to produce and harder to recycle than QR codes printed on regular paper.
>It will continue to accept national IC cards such as Suica and Icoca
I'd be sad if they do go or get relegated to some app, I love the little mascots.
I feel like they could combine the IC and contactless reader into one bit of hardware with some engineering.
Some are cutting back to just Suica and Icoca. Some are switching to, or using from the start, tap-and-pay (Visa, EMV, etc.).
so there's that, I mean if we can optimize QR code system. the winner obviously would be QR because no need to have an dedicated hardware for this
Yes, IC card would be faster but at what point the difference is matters???
Practically, many people however only start thinking about possibly needing to open some app to display it while they’re already blocking everybody else’s way at the transit gate…
The biggest practical benefit of IC cards is that they are by nature always “armed”, unlike QR codes in apps, and are readable from both sides.
On top of that, due to being able to run mutual authentication and being able to store a trusted balance, they are much more resilient to outages of any backend system. QR code tickets invariably need networking and a central backend.
This really comes down to adoption. In China, where QR is ubiquitous, almost everyone has the QR ready to scan well before they reach the scanner.
Technically the way to think about latency is that a process has N serial steps and you can (a) reduce N, (b) run some of those serial steps in parallel, and (c) speed up the steps.
For one thing, different aspects of the organization own the N steps. You might have one step that is difficult to improve because of organizational issues and then the excuses come in... Step 3 takes 2.0 sec, so why bother reducing Step 5 from 0.5 sec to 0.1 sec? On top of that we valorize "slow food" [1] have sayings like "all good things come to those who wait" and tend to think people are morally superior for waiting as opposed to "get you ass out of the line so we can serve other customers quickly" (e.g. truly empathetic, compassionate, etc.)
Maybe the ultimate expression of the American bad attitude is how you have to wait 20 minutes to board a plane because they have a complicated procedure with 9 priority levels and they have to pay somebody to explain that if you are a veteran you are in zone 3 and if you have this credit card from an another airline that this airline acquired you are in zone 5, etc... meanwhile they are paying the flight crew to wait, paying the ground crew to wait, etc. Southwest Airlines used to have a reasonable and optimized boarding scheme but they gave up on it, I guess the revenue from those credit cards is worth too much.
[1] it's a running gag when I go to a McDonalds in a distant city that it takes forever compared to, I dunno, Sweetgreens, even "fast" food isn't fast anymore. When I worked at a BK circa 1988, we cooked burgers ahead of time and stored them in a steam tray for up to ten minutes and then put condiments on them and put them in a box on a heat chute for up to another ten minutes. Whether you ordered a standard or customized burger you'd usually get it quickly, whereas burger restaurants today all cook the beef to order which just plain takes a while, longer than it takes to assemble a burrito at Chipotle.
Airline boarding is not the only class system in play. At every level of government, even within transit agencies, transit and its customers are seen as and treated as second class citizens. The idea of investing money, time or energy to shave even scores of minutes off the commute of someone who uses a bus, often seems as if it’s an unthinkable thought in these organizations.
Most flight attendant and pilot union contracts only pay them based on the hours with the door closed or in flight. (This is changing, but it's how it's been for a long time.) This reduces the incentives for quick boarding, as most of the flight crew is not being paid for that time.
The purpose of the many boarding groups is IMHO, to make those in groups > 1 feel as though they're missing out on some perk that they could get if they paid more. It's an intentional class system where some are encouraged to look down on those who paid less, and vice versa. It's good for revenue, bad for people.
Japan has its own versions of these things so I doubt it's this. The whole culture is, in general, not built for efficiency either.
Here's an example video to show the gates in action: https://www.youtube.com/watch?v=YffjxN3KsD4
Paris adopted an IC system as well and it is pretty usable, but not pushed to the extremes of the JR system because they weren't ready to invest as much, manage the money aspect and really make it a full blown business.
[0] some lines and areas are definitely money pits. That's where some companies bail out of the IC system altogether for instance, or go with a different, incompatible but cheaper implementation.
Japanese IC transaction: Less than 100 ms.
EMV IC transaction: Hundreds of milliseconds.
The person in front of you in the New York subway only realizing that they might have to look for their phone or card in their bag as they're already blocking one of very few turnstiles, and your train is arriving: Timeless.
Governments who think this set speed limits on roads to 45 mph, since that's the speed where most cars pass per second on a busy road.
Those same governments then act all surprised when it turns out their whole population is depressed and overworked when they work a 9 hour workday, commute 1.5 hours each way, and have no free time for life, relationships, hobbies, etc.
Japan's transportation infrastructure is great!
Unless things have changed, the ICOCA uses Felica as well. And old Offline Credit Card transactions used to be 200-300ms. Compared to Felica that is sub 100ms. May be offline transactions for credit card have gotten faster? Or may be not latency sensitive enough to notice the lag.
> Since there's no point in generating keys for a device which will not be used in Japan, non-Japan SKUs don't have Osaifu-Keitai functionality.
AFAIK, all iPhones from all regions since like iPhone 10 (internet says iPhone 7) support Osaifu-Kaitai unless I don't understand what that means. My USA iPhones works everywhere Suika, Pasmo, Icoca, etc work. Every station, bus, vending machine, convenience store, super market, restaurant, and retail store that accept these forms of payment, they all just work.
Given that all of this works, what is it I'm missing?
I will write a correction for this section to clear up the confusion.
I rooted my US model Pixel 9 Pro on my Japan trip last year to enable it. :D Literally a boolean in a config file.
https://github.com/kormax/osaifu-keitai-google-pixel
(The author's write up has more theories on why Google blocks it on non-Japan SKUs)
Since these cards actually "store money offline", the symmetric keys involved are really not something the issuers ever want to see leaked, so I assume it's not only a question of money (i.e. licensing fees), but also trust in the security posture of the secure element of the phone (if it even has one; many Android phones don't).
Contactless cards manage to boot up a JVM and run bytecode with the energy provided by the field, and many also run RSA with fairly large keys. I doubt that toggling some bits in an EEPROM factors in too much.
> I am not sure if single write is enough for making this robust/transactional.
Transactions are usually implemented via some form of two-phase commit, I believe, to support what's called "tear resistance". If the transaction is incomplete, the reader tells you and you just tap the same card at the same gate again.
Not sure about the implementation, but the feature is supported by all contactless stored-value cards I'm aware of.
That must be one of the things that ISO 14443 has improved on then, compared to FeliCa:
At least 14443-A (and I believe also B) allows addressing each card in the field individually by its serial number, and then selectively "halt" or resume it. Practically, this means that the reader can talk to cards one by one until it finds the one it's interested in (if any).
Unfortunately I don't know any practical system making use of that (it could get pretty confusing with payment cards, for example – charging a random card of several possible ones sounds like an anti-feature), but I still find it very neat.
PS: For anyone in doubt https://atadistance.net/2017/09/12/iphone-x-keynote-global-f...
They used to (and still do?) work faster with the Oyster fare card. Virtually instant. But paying with EMV cards/devices does add a noticable transaction latency before the gate opens. A few hundred milliseconds, I’d say.
EMV also uses faster symmetric cryptography between the card and the issuer, but the latter is not in the picture in a transit gate transaction – too much latency – so asymmetric cryptography it is.
That's also the reason some (mostly older, mostly international) cards don't work with TfL: Some don't even have the coprocessor (or just additional processing power these days, I suspect) required for RSA to cut cost, and only work for online payment transactions.
Google Pay emulates the card on the application processor, so theoretically it could be faster, but I wouldn't be surprised if anything won in terms of more performant RSA cryptography is lost to higher command processing latency between the NFC interface and application processor.
It would be interesting for somebody to do a latency comparison between Apple Pay, Google Pay, and a physical card!
But the London Underground gates are fast enough, with enough space from the reader to the gate, that if you’re ready (and the gate isn’t congested) there’s no need to slow down even from a very brisk walk to pass through.
In addition, the annoyance of these gates comes from having to fiddle with the wallet, etc. in order to find the card or the phone, or the fact that multiples tries may be required for the reader to actually read it; not the 200ms it takes for the reader to do so. I'm going to bet that faster NFC bandwidth makes the entire thing even more finicky, not less.
If you really want to speed how people go through the gates, then _remove the damn gates_. It's not rocket science, and there are some European cities that have _no_ gates in their underground systems. München comes to mind, but even in London less central stations have no gates. Beat that.
In addition , the article bashes NXP for using obscurity as a defense, then goes to praise Felica, whose apparently main barrier of defense is:
> the crypto is proprietary, and probably buried underneath a mountain of NDAs, so the public can't audit it independently.
This is literally the definition of security by obscurity. When I read the two examples set by the author, my only possible conclusion is that security by obscurity actually works, but only when you can keep it actually obscure. The only problem is that NXP failed to keep their algorithms obscure while apparently FeliCa did. It is basically the opposite conclusion to what the author is trying to convince me to believe. I find it totally unjustified to blame NXP for trying to keep the algorithm obscure by the courts when apparently FeliCa also does it -- just much more successfully.
Note: I personally consider MIFARE classic being _almost_ trivially clonable a requirement.
(1) it means people pay for usage. Someone going 20km pays more than someone going 1km since they have proof where you entered.
(2) it prevents more accidents/fines? - I had this happen to me in Paris, maybe because I was used to Japan. I was going to Paris Disneyland. I wrongly assumed I couldn't get there without going through a gate. I got on the subway, followed the signs in the tunnels to some train in Châtelet les Halles, was on the train, showed up at Paris Disneyland, was told I had to pay a 3x fine for not having the correct ticket. Or I could get on the train, go back to Paris, get the correct ticket, then come back (2hrs?) In Japan, for the most part, I couldn't have gotten on the train without a ticket and if I did go further than I paid can just adjust my fare at the end. No fines needed or imposed.
For the record, I am French. I used to be proud that nothing physical prevented you from boarding a train you had no ticket for. But, IMHO sadly, people like me have lost, because now trains also have ticket gates in France, which means that I:
A) No longer can accompany my ailing relatives to their train seats if I don't have a ticket myself (/as I could twenty years ago)
B) No longer can board the train when all my hands are full with luggage (since I need a free hand to search for the ticket in my wallet/bags to go through the damn machine).
The NFC readers on the gates in Japan will read cards from several centimeters away. My phone, which has Osaifu-Keitai setup, can be left in my bag and I just wave my bag over the reader as I walk by. It is incredibly rare for a misread to occur. They just work.
That comes with other problems.
Transit gates have the big advantage of providing you with instant feedback on whether your ticket or mode of payment is valid or not.
I've seen tourists get in trouble with ticket inspectors at least once in a very unpleasant way for simply having pressed the wrong button at an inscrutable ticket vending machine.
> I don't understand -- we are talking about 100ms or so of latency? which is almost completely dwarfed by any mechanical action such as the gates actually opening??
As others have already said, Japanese gates are open by default, and are more of a user interface indicator (to you and potentially station attendants nearby) than an actual physical barrier. You could step over or between the "barrier" that extends when the gate rejects your payment pretty easily in most, but it would be very obvious that you're skipping the fare.
So do the validators that are put in waiting areas, inside the trains, etc. in cities with no gates. That you can literally use at the time you want to use them (waiting for the train, inside the train, etc. ) , rather than forcing a bottleneck to everyone.
> As others have already said, Japanese gates are open by default
And this by itself already makes more of a latency difference than the entire IC card system does. Imagine what removing the gates altogether does.
Sure, you could spread the readers out a bit better across the platform etc., but that significantly weakens the "impossible to accidentally evade the fare" UX, as it still allows people to forget to tap when rushing for a train.
If you already live in an area where there are no gates, would you make the argument that "I need gates so that I don't forget to tap"? Put yourself in my shoes: you would laugh at the idea.
And it doesn't need mentioning that people who want to intentionally skip fare can do so, gates or not.
> would you make the argument that "I need gates so that I don't forget to tap"? Put yourself in my shoes: you would laugh at the idea.
Yes, I'm making that argument. I've lived in places that don't have transit gates for the majority of my life, and I absolutely forgot buying a ticket a few times (since I usually have a monthly pass).
Being reminded about a monthly pass having run out by the gate, automatically charging for a single ride (if I have enough balance) so I can solve the problem later, is great UX.
The way it's supposed to work is that if you pass your card while the gate is opened, it keeps the gate open longer.
In practice most people wait for the gate to close in front of them before reopening it.
I'm not sure why people do fgaf. I think the reason is that this makes you look like a fraudster going right behind a valid passenger.
Sometimes people will be low on balance and get rejected. People reroute quickly in normal conditions, but in rush hour that'll be a huge mess and everyone will be pissed at the culprit.
In most stations, every available inch of width is used for these ‘gates’, and people move at a walking pace through them except for when people screw up. It’s a remarkably effective system.
On busses they are de-facto soft gates, assuming the bus driver yells at you if you don’t use it - which often they do.
And why would they give up that sweet sweet rush hour revenue?
But I’m with you, with an express travel card set in Apple Pay (so you can just plop your phone down on the reader without opening anything) there is a beat for it to read on the Underground barriers, but it’s basically the same length of time as it takes to do the step and a half from the reader to the gate. At a normal walking speed I don’t feel as though I need to interrupt my stride.
Yes and yes
> What happens if a card is wrong/doesn’t scan/is invalid/etc at the higher speed?
Then the fare gates close in time to stop that person and the next 3 or so people behind them get annoyed and go around.
Stored value cards are optimized for speed and cost, and as a result have shared symmetric keys available in both card and reader. That enables extremely fast transactions.
EMV cards are (also) optimized (at least in the offline case, which transit gates use – the latency to talk to the bank backend would be too high) for security. The security model doesn't trust readers to keep keys secret, which means you need asymmetric cryptography for card authentication. The specs are old, which means RSA – which is very slow to run in cheap ICs embedded in these cards produced at scale.
Also^2, Japanese train stations have ads for B2B services, whereas almost every ad in London stations is for a musical. I'm not sure what this means.
(Also fondly remember the surprisingly numerous signs at Kings Cross about how you shouldn't assault any train employees, and how teenagers weren't allowed to buy matcha drinks because they have too much caffeine.)
(And my hometown Warsaw subway is even younger - 50 years, and we don’t have AC whilst temperatures are at a perfect level).
What London underground might need is not AC, but a process to cool down rocks - importing coolness during winters. To maintain equilibrum you’d need to pump out around 1TWh heat every year. To bring it down to normal levels in say 20-30 years you’d need to pump out 2-5TWh a year.
Osaka's Hankyu trains are full of ads for musicals (it owns the Takarazuka Revue), I think that all this shows is that London has a far more vibrant cultural scene, which is apparent at all levels of society. I'd rather see ads for musicals than the ads for male hair removal clinics.
And faster throughput would just increase that.
If the protocol is designed well, high speed doesn’t mean high error rate either.
The only time I ever had to wait in line to get into the station was after a stadium event, which is understandable. Rest of the time, you can pass the gates almost without stopping.
Did you mean it reminded you of others you'd seen on HN or just generally?
I don't know enough about these technologies to speak to the authoritativeness or veracity of gpt output, but I appreciate the gesture.
While some people have a reflexive dislike for AI output, I've done maybe a hundred o3 Deep Research queries now and found the reports to be generally high quality as well sourced as most human generated ones. I shared this one since I think it was a particularly interesting review the various systems around the world (I've personally used all those transit turnstiles personally and am generally familiar w/ RFID/NFC/EMV systems and didn't spot anything egregious).
(I find Deep Research reports to on average be high signal to noise than most of the human tokens being output on sites like HN for example.)
