Hacker News Reader

Comment by ccmoberg | Hacker News Reader

Yikes - working on getting it back up now. I guess load balancing for our www site needs to move up in priority.

In the meantime, you can see much of the same info on our AngelList page: https://angel.co/snowshoe-stamp

Nice concept, is the dot pattern secret? Which is to say is this something you 'train' between the app and the stamp or is it programmed? And does it do a 'multi-touch' type stamp or are you doing sequences? And if the latter I'm guessing you can do a sequence that fingers can't and thus avoid 'faking' the stamp manually.

All in all very clever. Not sure what the use case is though.

Great questions. Our basic stamps have dot patterns that are not obscured and are static on each stamp. They are precision milled out of the aluminum stamp body, and, thus, they are still VERY hard to spoof because the location of the observed dots needs to be precise down to the pixel resolution of the phone. You can't spoof them with you fingers. The nice thing about these stamps is that they are nothing more than a solid block of aluminum - no circuitry whatsoever. You can run them through a dishwasher or an autoclave - great for food service and medical applications.

We also have more secure "dynamic" stamps wherein the capacitive touch points are modulated in sequence amongst an array of potential contact points. These are still in the prototype stage, but we get a lot of requests for them, so may be looking to launch this new product line in the spring.

Finally, as for use cases - we can do many of the things people are currently trying to use NFC, QR Codes and/or GPS gating to authenticate. We already have apps that do retail loyalty, coupons, and micro-payments launched on the platform. We also have developers working on apps to authenticate EMRs (electronic medical records) and industrial systems maintenance. Finally, we won the TechCrunch DisruptSF Hackathon back in September by using the stamp as the authentication step for a webservice actuated door lock (article: http://snow.sh/ack)

Very much like the idea, and am definitely interested. My pre-question qualifier: most of the use cases that immediately spring to mind for have the user holding an authenticating block, and not the touchscreen device.

Security question for the use cases where that would matter. What is to prevent someone from making an impression mold and then using a technique like lost wax casting to quickly generate a copy?

As a materials scientist, this is the first thing I thought of.

Cheers and Kudos,

-p

Thanks for your feedback! Responses:

"My pre-question qualifier: most of the use cases that immediately spring to mind for have the user holding an authenticating block, and not the touchscreen device." -Could you say a little more about the use cases you are thinking of? Who do you consider to be the "user"? Most of our targeted use cases involve transactions where many smartphone users need to interact with a single, centralized system, person or installation. For example, our loyalty apps are downloaded by a retailer's customers, and then credits for purchases are added to their apps via a stamp placed at the retailer's point of sale. We consider both the customers and the retailer to be "users", though.

As for the security question, the capacitive touch points are embedded in a low capacitance elastomer, so physical duplication (e.g. casting) without destroying the stamp isn't an option. We also have other optional authentication layers (GPS-gating transactions to a stamp's known lat & lon, time-gating transactions to a stores hours of service, etc.) that would make even a duplicated stamp much harder to use.

Remember, its not hard to make a functional copy of the mag stripe on your credit card. I would argue it is much easier to do that than it is to spoof one of our stamps. Further, we don't print the "secure key" (the stamp point coordinates) on the front of our stamps, but your credit card number is plastered across the front of your plastic . . .

its not hard to make a functional copy of the mag stripe on your credit card

They aren't expected to be secure, so the consequences for being caught doing so are immense, and the industry around them has been dealing with this issue since its inception. We have newer technology that supersedes magnetic stripes for secure transactions.

Silly, but actionable use case example: I run a llama shearing business, and I attach an aluminum ID block to each llama's collar as means to track each llama's shearing schedule. That's the type of example, where there are many blocks and few readers. Replace llama with "something else in inventory" for more practical examples. Clearly all is golden for the POS retail coupon verification case, with few blocks and many readers.

---

Embedded in elastomer makes sense. I had several ways to skin that cat pop into mind, but I figured there was just one piece of the extant ID block I was missing. Thanks for clarifying.

---

All other security points well taken. Wasn't trying to get down on the idea at all. Indeed have numerous use cases in mind that would work well in a reversed use-case of the consumer facing implementation. Not all situations need high grade security.

-phil

You can't spoof them with you fingers

Well, okay, but I could trivially present a custom application for stamping that would record them to make my own stamp. That's probably fine for loyalty programs, but is most certainly not fine for EMR. A dynamic QR code might be clunkier, but a static key will never be secure.

(or sniffing the real application from the background by jailbreaking)

You can't do the NFC use case, that is a user unique security token with qualifier exchange. But that certainly isn't to say this isn't a good idea.

So one thought, given the durability, is that you could build this into a kiosk type display add. So part of the frame has the stamp facing 'up'. For more info put your phone, face down, on the stamp briefly while running your app. Then you translate from kiosk lat/long to ad vendor.

Similarly in museums where you want to listen to various bits of the 'tour'.

Should be back up now. Sorry for the inconvenience.

Reply on news.ycombinator.com