Hacker News new | comments | show | ask | jobs | submit login
The Physical Web (github.com)
250 points by gelliott on Oct 2, 2014 | hide | past | web | favorite | 82 comments

" . . . a vending machine, a poster, a toy, a bus stop, a rental car, and not have to download an app first in order to use it."

I can, and do, do that now. I don't need or use an app to interact with any of those things, and when I have been offered apps, they've been worse than useless. Seriously, worse than nothing, they've been annoying, harassing, clunky junk.

I just want to be able to approach any of these things and pull up all of the documentation regarding it. On my phone, tablet, laptop, etc. No "advertising tracking" nonsense. How to use it or fix it. If the owner wants to turn this feature off, they should be able to, but it should be on by default. If the owner thinks its a security loophole, he is both mistaken and lacking in security.

This includes stores. Why can I find a shelf stocker in Target and he can use his li'l belt computer to tell me where anything is in the store, down to the shelf, and how many they have, but I can't have access to that information before I enter the store? Big box stores were never so annoying until I got used to Amazon.

> I don't need or use an app to interact with any of those things, and when I have been offered apps, they've been worse than useless.

A store clerk recently invited me to download their app, and when I showed her that it requested every single permission, she was still baffled as to why I would want to cancel the download.

> Why can I find a shelf stocker in Target and he can use his li'l belt computer to tell me where anything is in the store, down to the shelf, and how many they have, but I can't have access to that information before I enter the store?

Home Depot's mobile app will tell you exactly where an item is on the shelf, including aisle, section, and level. They even have a map so you can find items that aren't on numbered aisles. It's fantastic.

Home Depot is terrible at actually tracking their inventory. There app is useless when it tells you where the item is, only to find out that they actually don't have any in stock despite what their tracking system says.

I order easily 20x as many items from Amazon, and never had an email saying "oops, we actually don't have any in stock".

And this is why the stores don't give out their actual inventory numbers to the public: because they can't accurately track those numbers. Things get stolen, lost in storage, and moved off their shelf to promotional displays.

And then people complain about how "useless" the app is. This is much less of a problem for Amazon where there aren't any thieving customers in the store.

It is useless if the information provided isn't correct. In some cases, it's actually worse than useless.

The Lowes app has a similar feature - it really is useful.

> but I can't have access to that information before I enter the store?

Because why would they want you to have that information? They are making money on you not knowing where stuff is - they're hoping that you'll pick up additional items you didn't plan to buy as you wander around the shop looking for the things you need.

Ditto for "worse than useless" apps - they are not meant to help you, they are meant to take money from you.

So pg recently wrote again, that the best way to get money is to make something people want. But it's not true. Time and again, businesses from Comcast to Target, to Uber, to browser toolbar makers, to the guys that give away pendrives that secretly install as keyboard and open their website, to your random SaaS "we'll kill e-mail and your puppy" toilet-paper startup - all of them keep proving that the best practical way to make money is to lie, cheat and bully people into paying.

We need a really strong, aggressive pushback against dishonest business practices, and that includes pretty much the entire advertising industry. We're already drowning in manure, having to use stuff built for the reverse of stated purpose.

A lot of brick-and-mortar store chains do that in France, actually (Darty, FNAC, etc.).

They have an online store, and on each product page there is a field "stock in your preferred store". That's very handy; you can even order it online and pick it up in the store, so you don't have to pay or wait for delivery.

Several times I was planning to buy online, but since I found out on their site that their price is competitive and I can get it today, I ended up taking my car to a brick-and-mortar shop.

> Several times I was planning to buy online, but since I found out on their site that their price is competitive and I can get it today, I ended up taking my car to a brick-and-mortar shop.

This is strong competitive advantage of brick&mortar stores. I actually buy more locally than on-line for the sole reason that I'm impatient and when I need something, I need it ASAP, so I prefer to just go and buy it instead of waiting few days (same-day delivery is not common in Poland yet).

Would it work out for you if you just had to install one app for all of those uses - e.g. one generic "physical web" app?

From Roy Fielding's thesis[0]:


REST provides a set of architectural constraints that, when applied as a whole, emphasizes scalability of component interactions, generality of interfaces, independent deployment of components, and intermediary components to reduce interaction latency, enforce security, and encapsulate legacy systems. I describe the software engineering principles guiding REST and the interaction constraints chosen to retain those principles, contrasting them to the constraints of other architectural styles.


Amazon is one of the few companies that has actually committed to this. There's no reason the Internet of Things can't follow this pattern too.

[0] http://www.ics.uci.edu/~fielding/pubs/dissertation/top.htm

Huh, this sounds like it could dovetail beautifully with FirefoxOS's attempt to blur the lines between "native apps" and web content.

It'd probably be pretty easy to have these things show up in an "Ambient" smart collection on the homescreen, or in the universal search pane...

Edit: Looks like you can get this running on an RFDuino (https://github.com/google/physical-web/blob/master/documenta...), which are currently $20 at MicroCenter, and a USB shield is $25. Still 2 in stock in the Twin Cities, if anyone else is interested!

Edit 2: ...Though, it looks like FxOS doesn't currently support BLE (https://wiki.mozilla.org/B2G/Bluetooth#Unsupported_Bluetooth...), so this may not be the most useful path at the moment. :)

I think this is a great interface. This is cast as a discovery system for public devices but I think it is equally well suited for remote control of private devices.

The writeup indicates the beacons run in broadcast mode. I think it would be good if there were also a discovery mode where the client could ping devices and they could respond. This would be good for a few use cases, including “private” devices like things in your home. These devices could be configured to only respond to certain clients.

For home devices, another feature that would be nice is if the device could use http-over-BLE to service the web request itself. The device would run a small web server and the web page on the client would be the user interface for the device. For home automation there are these complicated specifications for different device types, but HTML is the perfect interface to run any device.

The company I work for (IoT Design Shop - http://iotdesignshop.com/) is actually releasing beacons and software that does exactly this. We have a BLE beacon called Signul that couples with an app that allows you to do all kinds of hyper-local interactions.

Right now you can support our Indiegogo campaign (https://www.indiegogo.com/projects/signul-the-world-s-first-...) and get a Signul beacon. I have a prototype version and it's really quite good.

If the Indiegogo campaign hits $50k (it's currently sitting at $31k) we're going to release an SDK for Signul allowing anybody to write an app that interacts with the Signul beacon for either iOS or Android devices.

I kinda hoped the devices would actually serve the apps themselves. I guess a URL makes more sense though. But I wonder if it will lead to physical devices that give 404s when they are obsoleted.

Anyway, this seems pretty exciting.

There's no reason it couldn't mean that anyway - if a device is open to the Internet then the URL it serves via Physical Web could easily be its own. But it allows for less-smart devices too.

The biggest problem I can see, as with all proximity devices, is spoofing. How do I know the item listed as "Bus Stop" is actually the bus stop, and not someone's malicious Raspberry Pi hidden in the bush next to it?

>How do I know the item listed as "Bus Stop" is actually the bus stop, and not someone's malicious Raspberry Pi hidden in the bush next to it?

The same way you know that top search results are accurate for a given word/phrase. As stated multiple times throughout the introduction[1] the solution for filtering out spam will probably have an implementation similar to internet search engines.

[1]: https://github.com/google/physical-web/blob/master/documenta...

How do I know the top 100 items listed as "Bus Stop" aren't someone's malicious raspberry pi hidden in the bushes, spitting out hundreds of URLs and varying the signal strength so they appear to be coming from different locations and...wait a minute. I think my phone just got DOS'ed.

The addition/counter to the Physical Web proposal I have been discussing with Google prevents these kinds of security issues in many locations/places: meet Geo-Origins --> https://github.com/csuwildcat/geo-origins/blob/master/explai...

You will never stop DOSs. It's trivial (I imagine, can't think why it wouldn't be) to DOS someone's cell signal. Allowing users to see which items are "certified real" and not raspberry pis in bushes is an easy engineering problem we've solved in a number of different contexts.

That's a lot of money to DDOS people's phones at bus stops.

I don't think so. I think the proper analogy, that authors of the proposal have missed, is not search & spam, it's ATM and skimmers. How do you know you're not inserting card into a scanner?...

... well, everything proves that most likely, you have no clue.

Jon Nordby demoed this idea with http://flowhub.io/ a couple of weeks ago: http://youtu.be/wCdKukf-dHI

1. Tap tablet to Raspberry Pi with NFC tag

2. Pi serves Flowhub IDE (noflo-ui) with the currently-running graph, ready for editing

This demo drops you directly into the source, but could easily have an interface layer first. It's just a link.

Isn't this what QR codes on signs were supposed to do?

You don't need an "app" just to talk to some retail thing. Many retail apps are a dodge to get inside your phone and snoop around, acquire your contacts list, and track where you are.

Yep, this is just QR codes made better. It's similar to pet microchip implants too. The only app you need is one that knows how to listen for the signals and point you in the right direction when you choose to visit one.

How is this similar to pet microchip implants? If there were a way to robustly put a QR code on my pet instead of using a microchip, I would - you'd be able to identify easily that the pet is tagged, and the readout devices would be more widespread.

I also don't see at all how this is better than a QR code on a physical object, other than that it's invisible - which is itself not great, since unless you allow these things to be intrusive and pop up something on your phone when you get too near to one (no thanks), there's no built-in call to action. You'd basically need a sticker on everything supporting this that says, "I support Physical Web!"... in which case you might as well have a QR code.

How is this similar to pet microchip implants

In that it's a thing that's hidden in an object that you can scan for that identifies it and gives information about it.

there's no built-in call to action.

It's trivially and plainly better than a QR code on a physical object. Even if you did have to have a sticker on everything saying "I support Physical Web", the fact that I don't have to scan the sticker -- instead I can just click it in my phone -- makes it better. But I don't think you're using your imagination if you think you'd have to have a sticker on every single thing that broadcast itself. Let me use my imagination for a second.

When I enter a garage it would be nice to not have to look around for a QR sticker, reach my hand out of the car with my phone, and scan it. I would just know that when I enter a garage my top hit when I bring up the Physical Web (for lack of a better name) will almost certainly be the garage's interface to allow me to pay.

When I'm walking around a city it might be nice if I, casually browsing the Physical Web, could notice that someone near me is willing to sell bitcoins for cash-in-hand in a virtually untraceable transaction, without having to see and scan a QR code on that person's forehead.

It would be nice if I could set my phone to automatically tell me over headphones that the next bus is expected in N minutes whenever I walked up to a bus stop, all without requiring me to pull my phone out or take my gloves off in the middle of a Chicago winter.

You're just not thinking outside the box.

I don't know that it's plainly better than a QR code, but your point is presumably well taken that if a broad area were saturated in ID signals that would be better than a single fixed QR code, once the user knows that the physical web is supported. I was thinking mainly of this in the context in which it was presented, which is being embedded in discrete objects like vending machines or the like, in which case the search space for a QR code is fairly small, and advertising that the physical web is supported could be replaced by existing technology.

I also don't see that it's "plainly better" to click on it through a phone. It depends on the different interfaces. How much do each of these things cost? If we're allowed to spend infinite money on any minor improvement, why not give everyone HUDs with high resolution cameras that can scan any QR code within line of sight? How much battery life are these scans taking? How much complexity do they add to the phone UI? What will the security issues be? I'm just not sold on the big benefits here.

Technologically, they are fairly dissimilar, but on use cases, they're more closely aligned to the layperson than their tech distinctions would indicate.

Both require relative proximity -- in the case of the pet sensor, physical propinquity is a constraint imposed by the technology, while a QR code, that constraint is imposed by the camera attempting to read it. Either way, you have to be close.

Both technologies act as a bridge for data. In the case of the pet sensor, it basically encodes a unique ID, which is then used to query against a database housed elsewhere, which holds the real information. In the case of the QR code, the same is basically true -- the real payload is after the redirect.

Either way, the mechanics aren't too different -- get close to something, read (or scan) it, and be taken to wherever the magic really happens.

I was thinking along these lines as well. I suppose the difference is that these are QR codes that "broadcast" themselves to whoever wants to listen?

I have been discussing this proposal with the folks at Google, and generated an additional proposal, Geo-Origins, to address some of the issues Physical Web introduces: https://github.com/csuwildcat/geo-origins/blob/master/explai...

Geo-Origins brings the trust model of the web to the physical world, and allows for more frictionless interactions in many locations.

Oh, now iBeacons(https://developer.apple.com/ibeacon/) will make more sense. Let's implement iBeacons to vending machines and also to other machines, then use Pushmote(https://pushmote.com) to manage interactions from cloud and import Pushmote SDK into operating systems or Chrome natively. Sounds great. :)

I like the idea, but the first version seems quite limited. It is more of a GPS based webapp discovery tool. You click on the button, the app sends your GPS coordinate to the main server, which returns all useful webapps in the neighborhood.

I think that using NFC chips, which are extremely cheap, is very appropriate here. Use case: You point your phone to the vending machine/renting car and the respective webapp is shown automatically on your phone.

NFC is tough because you have to be within about an inch of it to get a reliable connection. For the use cases of a parking meter or vending machine it works, but for a bus stop or store it becomes yet another thing to stand in line for.

I haven't even had good luck with the parking meters in Seattle :(

I thought this works by each beacon broadcasting a URL with BLE, with the app sorting the received URLs by signal strength and not how you just described it. Correct me if I'm wrong, though!

> "The number of smart devices is going to explode (...)"

But why? Why are you/they so sure? Why do we assume that humans will just appropriate any kind of technology thrown at them?

I assume this comes from educated tech guys, who are thorough with their work, tools and respective usage. How can they be so lighthearted with the premises which relate to human usage of technology?

This is a rather unusual comment and this link is probably the wrong place to have this conversation, as hacker news is probably the wrong place in general.

Anyway, it's rather simple. It's just taking what we have seen in the past (people appropriating any kind of technology thrown at them [1]) and extending it into the future.

Is this the correct approach? Probably not.

I like to remind myself that there was a time when smoking was the thing to do. Little knowledge about potential dangers was available. Projections about future adaption of the habit might have looked similar.

[1]: Strictly speaking this is not true, since there is a huge cemetery of rejected technology that remains mostly invisible. Smartphones just turn out to be one of the technologies that have become widely accepted, creating a wrong impression overall.

I'm not sure what you mean by "appropriating technology". The number of smart devices is going to explode because we want to automate all the things, and collect lots of data. It'd be great if I could ask my stove what food I've eaten this month, how healthy it was, how expensive it was, maybe even tell it to start preparing dinner at a certain time. The way to do that is to put a computer in everything.

> I'm not sure what you mean by "appropriating technology".

People accepting, wanting, using and integrating said technology in everyday life.

> The number of smart devices is going to explode because we want to automate all the things

This is precisely what I'm questioning. "We want"? How can you know that? I believe you want that, I believe that maybe people in your circles do, too. Still, that is no evidence at all to extrapolate that "we want".

There's an easy way to check: let's see what sells.

Last time I checked, mildly smart home "robots", like automatic cookers, dishwashers, washing machines, even vacuum cleaners sold rather well, and were widely adopted. The smart devices, paradoxically, liberate you from thinking about dull details: you just put some clothes into a washing machine, press a button explaining approximately what type of clothes are there, and walk away. The machine figures the rest.

This is a type of liberation people seem to actually like.

Also, people usually don't like uncertainty puzzling new experiences. Letting people on a stop see where the bus is does make people happier. (Myself I use a mobile web site that shows real-time position of NYC buses, and can attest to that.) If you walk into an unfamiliar store, especially abroad, it could take some time to find simplest items like a bottle of water. If there was a more-or-less unified 'online' interface for finding things in this particular store on your phone, it would make many tourists happier.

> Last time I checked, mildly smart home "robots", like automatic cookers, dishwashers, washing machines, even vacuum cleaners sold rather well, and were widely adopted. The smart devices, paradoxically, liberate you from thinking about dull details: you just put some clothes into a washing machine, press a button explaining approximately what type of clothes are there, and walk away. The machine figures the rest.

I'm in the UK, and am rather middle-class, sort of. At the least, we don't have to worry about money much, and if I want to travel half way across the UK, I don't have to think much about it. I've just booked a film a day for the next week at a film festival.

On the other hand... I've literally never seen smart versions of these machines. I've barely ever seen a dishwasher outside a bar or restaurant, I've never seen an "automatic cooker", washing machines are generally much the same as they were 20 years ago, and robot vacuum cleaners are decidedly still a futuristic thing - most vacuum cleaners are, at their core, the same sort of thing we had 20 years ago with a new shell.

Perhaps these things are a lot more common in the US, and possibly in more upmarket parts of the UK, but I'd quite happily suggest that the vast majority of people are not aching to buy a smarter washing machine.

I wonder if it's simply the case that they are more popular in the US. Every house I have lived in has had a dishwasher (some are smart enough to figure out when dishes are clean and stop, instead of just using a timer), and it's generally expected that apartments will have one also.

Washing machines typically have settings for the type of clothes that are inside. And while I've never used one, the robotic Roomba vacuums are cheap and seem to be pretty popular.

If it is indeed the case that uptake of these items is much higher in the US, it would be very interesting to understand why. I doubt very much that it's cost since they're available at multiple price points. Maybe it's something like the reason why essentially no US homes have an electric kettle while essentially no UK homes don't have one. IOW, it's cultural!

> There's an easy way to check: let's see what sells. > > Last time I checked, mildly smart home "robots", like automatic cookers, dishwashers, washing machines, even vacuum cleaners sold rather well, and were widely adopted.

Do you have proper data to claim this? Good sales and wide adoption? And how can you be sure that the buyers are "everyday people", not techno-enthusiasts and early adopters?

I mean: it is rather "easy" to see a big increase in sales for a given technology, when it is "young". The big question is how sustained that growth will be, when the early adopters are served and the company needs to target "regular" people.

I thought it was pretty obvious, but here are some numbers for you. According to http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9N... over 300,000 people had installed the Crock-Pot app for iOS by 2010. It costs $3.99 so that's $399,000 in app sales, plus whatever happened in the last 4 years.

Around 75% of homes in the US have a dishwasher. Nearly 20 million dishwashers are sold in the USA each year. http://qz.com/29147/death-of-a-dishwasher-families-around-th... For comparison, only 5 million homes are sold in the US each year, so it's not just that they come with houses and are never used.

>And how can you be sure that the buyers are "everyday people", not techno-enthusiasts and early adopters?

I don't think it matters. The future of smart-devices is looking good of a number of reasons and this is why more than a few people think that smart-devices are going to see widespread adoption.

"Everyday people" don't need to be the ones adopting this widely just yet. There's a trend in the evolution of technology where people don't know that they want/need a specific technology until they've been shown the power of it. "Everyday people" never asked for Home PCs; they were pushed to them and now it's hard to imagine life without a home computer. "Everyday people" never asked for smartphones; they were pushed by technology companies and now they're almost a mandatory device for navigating modern culture.

I think smart-devices will see the same fate. There is a lot of talk and development into IoT and technologists see the power of such networks even if "everyday people" do not. Smart-devices are a natural progression of IoT and the other devices everyone already has in their pockets. I think it will be adopted in the same way other technologies have: It will see small, and then large-scale adoption in a niche area, people will see the power of IoT and smart-devices, and then people will begin to want it everywhere. This will be driven by technologist support and marketing.

The article may have presented the outcome a little optimistic but I don't then they're far off the mark.

This was certainly true for the past, but I'm not sure you can extrapolate for the future. The events of the recent years have given the public not just a sense of the power of technology but also its negative side.

At least here in germany, the Snowden relevations have done a good deal to make people realize just much much data about them is on their net and how hard it is to control access to it.

Revelations like Apple using iPhones for large scale profiling of location histories and (much later) Samsung having its TVs report home with a history of viewed programs didn't put smart devices in a good light either.

In parallel, numerous hacker and data leak scandals taught the public that maybe data entrusted to the cloud isn't as secure as everyone promised.

Finally, many people only now start to realize just how much a paradigm shift the internet really caused and what some of the psychological and sociological implications might be: There is the (still vaguely defined) "internet" or "mobile addiction", there is the growing trend of viewing phone usage (or usage of other devices, i.e. Google Glass) during social gatherings as impolite, there is the whole discussion about what role privacy should play in the future, etc, etc.

All of this doesn't stop people from buying new phones, TVs, fitness bands, etc. But depending on who you ask, they do it with a growing bad conscience. If you try to introduce new technologies that have a higher cost and less obvious benefits (like smart homes), it might have an effect.

What sales figures won't tell you though is how well people are doing who choose not to buy or can't afford those things.

Also, your argument does not account for tomorrows sales figures, which might look very different, as executives from any industry that software has replaced can tell you. Of course that could never happen to us.

Because time and again history proves a point that a surprising amount of smart people keep missing - that technology works on Blub paradox. People will use everything you throw at them, so you can't assume that "no one wants to use feature X". If you build it and ship it and make people buy it (hello advertising), they will use it, for things you haven't even thought of.

> I believe you want that, I believe that maybe people in your circles do, too. Still, that is no evidence at all to extrapolate that "we want".

There are three kinds of people. A small group that wants to automate stuff, a smaller group that doesn't want to automate stuff, and the majority who just mindlessly buys whatever is in stores today. You think people wanted smartphones? The tech crowd, maybe. But the rest? They just started buying them as mobile networks rolled them in to their offers.

People and businesses are paying money - sometimes a great deal of money - for smart devices. That's a pretty good indication that they want them.

Who is "we"?

I guess one of the biggest questions here would be authentication and reputation. You want to make sure the device is what you think it is. So they will need to have cryptographically signed messages, ie MAC, and you need to obtain a certificate as well from their server on the web in order to interact with them.

But once you have those, you WON'T need the internet in order to communicate. The vending machine could be in a basement and bluetooth would be enough.

I'm a big proponent of re-decentralizing the internet again, and the web (and physical web) is a big part of that. There's no reason why people in an african village MUST have all their cell signals travel up to a balloon and to Facebook's servers just to communicate with each other. Mesh networking and distributed power generation will bring about a revolution.

Most normal people I see carrying a smartphone actually never use the browser. Never. I mean NEVER.

People use the browser very regularly and don't realize it because it's an embedded webkit. A classic example is people sharing articles on social networks.

Native apps are a bad experience for this kind of application because they are not ephemeral. If everything in your environment has an app associated with it, then your phone will be clogged with hundreds or thousands of apps, 99% of which you never open. You want a transaction-cost free model for envirnonmental interaction, the same as 'surfing' the web. No persistent cost for experimenting or trying something out.

The whole idea of "installing" stuff has really regressed computing back to the 90s. The web introduced the idea of emphemeral applications which are cached, so that you never need to manage the memory of your device.

I'd like to see mobile native apps adopt Web-style model. You can pin the cache and permanently install apps that are important to you or have giant resources (like huge games), but most of your apps should backup-and-delete themselves on the fly when coming into disuse and space is needed.

I'm sick of being asked to rate apps, and sick of having to delete apps to make space.

Steve Jobs once said "if you see a task manager, they blew it" Because people shouldn't have to manage the kernel resources of their phone. To that I would add, "if you see an install button, they blew it", because people should not have to manage their flash memory.

I thought the nightmare of cleaning up your computing devices went out with utilities to clean Windows drives and registry in the 90s.

I agree with you. The problem as I see it is iOS-induced obsession with hardware I accelerated special effects. It's easy to make a web app that's silky smooth and has highly optimized paths for interaction, but iOS worshipping has caused us all (web and native developers both) to abandon interaction design for special effects design.

Sad to think that 2000-2010 was actually the golden age for genuine focus on interaction design. Ironic too that Apple was the company who killed it.

I should add: Apple generally does great interaction design. But they bundle it with expensive special effects and visual design. Everyone else tries to copy all three, but they can't, because it's insanely hard to do all three. Thats Apple's moat. And when mortal dev shops bump up against the "interaction/effects/visual, pick two" bargain, interaction is often the first one to go because it's the one that doesn't come through in a demo. And they're relying on the parallax effects to wow their client/boss. Plus they want to feel like they're going for the moonshot "Apple-quality" bar so their ego can get a boost. And honestly most people stop doing interaction design the moment they get a picture in their head of what their app could be. We become emotionally wedded to interactions the minute we invent them. God forbid we actually pay attention to the friction in our users' lives and put those things at the top of our prioritized lists.

I think we need a return to boring, native-to-the-web software with superb interaction design. And sacrifice pretty fonts and sacrifice 3d parllax effects and sacrifice animation, except where it is truly impacting comprehensibility (and not just feel).

That's the basket my eggs are in.

This is exactly how I see it.

I would disagree. I know a lot of "normal people," and most use their browser to google or search using their preferred engine or a particular website, as well as load links from emails, or visit websites on the go.

Personally, I'm not sure that using a particular company or product's app really saves me any time. Half the time, the app takes forever to load or is so non-properly story-boarded, that I resort to google anyway. Maybe I'm the outlier.

I'd like to see some statistics on this. I think you're right, by the way, but I hope you're wrong, and either way, I'd like to know the real answer.

I don't see how this is relevant. The proposal is for ble to broadcast a URL. Native apps can open URLs too.

I would imagine that this holds true for social media apps, but, in my experience, the browser is still king for commerce.

So basically, its QRv2 without cameras and instead nfc/ble or some other physical layer.

Meh, we need AI to understand API interfaces automatically and can communicate together seamlessly, without any hard coding. That's when device to device integration gets REALLY interesting. Wait, isn't this just bluetooth + web?

We don't necessarily need AI, we just need APIs to be standardized so everything knows how to interface with them. This is IMO the killer app for the Semantic Web (I just wrote an article on this: https://medium.com/@mcriddy/semantic-web-design-92ef35f66c9f).

How does that eliminate the need for AI? Having a "standardized" format doesn't fix much of anything, except a bit of search engine display logic. It does nothing of the sort of eliminating the need for render code per app/site. Maybe some actual, solid, examples of exactly what a "standard API" would accomplish would help explain.

It is, but I think the really interesting part is to making the connection seamless and as direct as feasible. e.g. how do you do it on one end of the spectrum if neither the embedded device or the users device has access to the internet, but they can see one another via bluetooth and wifi?

I think an app makes a lot sense to initially bridge embedded devices (that may only support simple web interfaces or even lower level network messages) to smartphone browsers (which are currently oriented to handling fairly rich web interfaces from full servers).

But where it should get really interesting is if you make the app support a generic enough protocol that you don't need an app-per-device type, but the app/protocol can handle a wide range of generic embedded device interfaces without requiring a full featured / high performance web server on every embedded device. After that you could basically get the protocol supported at the browser/os level and the app disappears.

I agree that the AI component is important to understand. Essentially we're talking about putting everything on a computer system, and installing the same OS on every "machine." One of the problems that I foresee mirrors that of Microsoft, in terms of monopoly, as well as the ongoing debate between PC vs Mac usage. Inevitably, the shared OS maker will have the monopoly, though others may try to dethrone it, or there will be a battle between two to three different systems each competing for market share, and proving almost incompatible with one another… until some Hackers develop a patch.

we need AI to understand API interfaces automatically and can communicate together seamlessly

I agree with that statement in general, but I don't see what it has to do with Google's Physical Web proposal/idea. You'd want what you describe even in a world with only traditional networked computers.

I think having the AI to accomplish that would be the really interesting bit. That makes a lot more technological advancements possible. And yes, this is just finding new applications for the web and expanding its value/reach.

The underlying tech is bluetooth + web, but the user interface for this does not currently exist.

The way I see it, this is better implemented with AR and CV using specific items as "markers" than with a "nearby lookup" table. The thing you want may be across a highway and still be near making the listing somewhat imprecise.

The range of interactions is also greater with AR, as it gives the ability to be triggered (eg. highlighted) just by looking at it with the AR camera.

Reading the page, the comments here and all... I can't stop thinking Firefox OS... just... can't...

I really don't understand why people aren't agreeing that "native" apps are a thing of the pre-internet era.

The project's ethos is very similar to Rob Pike's 'dream setup' that he describes in his usesthis interview, http://rob.pike.usesthis.com/

Well, as long as I don't have to use HTML and JavaScript.

Looks like yet another vector mostly for local advertising.

So many typos in the open PRs right now.

Obligatory http://xkcd.com/1367/ I'm not saying we can't do better than web apps, but even web apps are better at this than native apps.

Just put QR codes everywhere.

It is funny, but I would argue that the reason people don't use QR codes is that they haven't yet gotten enough value out of them. When there will be an incentive to scan those codes suddenly everyone will be using them.

^^^ LOL. I hit the back button thinking "WTF?"

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact