The copy highlights something that I don't understand about Twitter. It reads:
Maybe you want to get a tweet when your laundry's done [...]
Who would actually do this? Why would I use Twitter to send a private notification like that? It's pretty much the worst system imaginable. Twitter works great for broadcasting to people, but it's nonsense for one-to-one messages.
Is this really how I want to set up my washing machine:
1) Create a Twitter account for the machine. Let's call it @pavlovs_2011_electrolux.
2) Subscribe my public Twitter account to @pavlovs_2011_electrolux. (Alternatively, maybe the washing machine can be configured with my public Twitter account so it can send messages to me.)
3) Keep an eye for my washing machine's notifications amongst a sea of disorganized tweets about public topics.
Lots of good points for and against the twitter integration. There are ways of setting up Twitter such that you'd have your privacy alright, wich I'm sure many people would still not care about—after all, what could possibly go wrong? [1]
So yes, it lets people shoot themselves in the foot; yes, people will most likely shoot themselves in the foot; but no, that's not the main problem and in fact you can hypothetically use twine + twitter in a more or less private manner.
The problem is that broadcasting through Twitter makes you depend on a service by a company, where you could just as easily have come up with a solution that works with a publicly-known protocol which can be deployed by many parties (the closest match here would be status.net).
Then again, their point is precisely to make using the bloody thing easy. It's increadible how the easy way to make things usable always implies lots of trade-offs in security, privacy, openness and whatnot... No wonder people who solve usability problems the right way are so valuable.
[1] DEFCON 18, "Web services we just don't need": http://youtu.be/mbZccSwgbIU (it runs almost 1h, and is equal parts hilarious and pointless).
the point isn't to get a tweet, it's that twitter is a great free IP-to-sms bridge. back in the day, in the small window of time when twitter was popular but not everybody had smartphones with push email, the twitter api was by far the easiest way to get push notifications. and now it's just cool.
Sure, good rant, but what are the alternatives for real-time notifications that do not require a "nerd degree" to configure? There are one.
Also, practically speaking - filtering out specific tweets and alerting on them is an issue with a Twitter client, and not the Twitter platform. Similarly, the privacy setup can be readily handled either via direct messages or by protecting tweets. In other words, Twitter is really not a bad choice for their purpose. Not ideal, but still a good functional fit.
Do you really need a reminder for when your washing machine is done? Is this a problem that society really needs to solve? My washing machine has this sweet feature called a buzzer. And that feature can be turned off in favor of wait an hour and go downstairs.
If you guys like this project, Green Goose is another sensor based system coming out: http://greengoose.com/
Overall sensors + connectivity will be getting very cheap soon and it is cool to see a project like this see beyond this to include ifttt type functionality integrating other existing systems and actions.
The move to become the platform for all of data being generated by devices like this is getting heated as well (not sure if the Twine guys are thinking this route or not with the Spool web app) With Pachube, Thingworx, and a whole host of others attempting to become this centralized hub. It will be interesting to see if we end up with specialized platforms for different segments of our lives (i.e: one for health, one for your home, etc) or if a 'Facebook' for your things will emerge.
Self Plug: We are tracking this "Internet of Things" on http://postscapes.com if you are interested in the topic at all.
What's most interesting to me is that with respect to hardware, all of the parts have existed for years, but we're just beginning to see products that target tech-savvy consumers who are want to build something that suits their needs.
I've been working on a more modular, open source system like this for the last year or so: http://rascalmicro.com. The Rascal has powerful hardware and a real web interface, but it works with Arduino shields.
Twine looks pretty good, and I'm heartened to see that the web interface has some polish to it.
(Supermechanical, if you're reading this, let's get together some time. I'm one town over in Somerville-- email is in my profile. I don't think we're competitors. My customers use soldering irons; I bet yours don't.)
Greengoose looks interesting, but I've just read every page I could find on their site (3 mins?) and still don't understand exactly what the 'sensors' are and what they can 'sense'.
How does a sensors tell if I've fed the dog today?
Pretty sure the sensor detections motion above a threshold. So basically "the food bag was moved 45 degrees" is the trigger. That should kick off a 915Mhz analog pulse.
It's not a digital stack IIRC, which is why the battery is so small and circuit is like a dollar total.
Green Goose looks interesting. But the TC article says that they were going to start selling kits on 02/28/2011, but I don't see any on their site.
To me, it's basically a matter of cost. If the "washer is done" sensor costs $50, I'd much rather just get up and check it myself. If it costs $5, I'd be more than happy to buy it. A lot of these sensor companies seem to be targeting the $20-$30 price point; to many, that is too expensive!
Make a base station that is about $20, and a sensor that costs less than $5. You'll be selling millions of those things in no time.
On the topic of sensors, I was making pecan pie yesterday (with corn syrup) and thought to myself how awesome it would be to have the jar/bottle of whatever substance be able to measure the output/input of the opening. This would enable making recipes without measuring cups/spoons, automatically order new stock when running low, and even track the consumption of calories for a family.
I love the idea, but conceptual and economically, I find the idea of using one sensor in a multisensor device on a long term basis problematic. I'd prefer a modular system where I can provide a wifi base module with specific sensors as required through plugin parts.
It feels wrong, I agree, but the individual sensors are dirt cheap. From poking around on farnell.com, it seems likely that the total bulk value of the sensors is probably south of a few dollars per unit, which is much, much cheaper than introducing the complexity and brittleness of modularity.
As it was said in a discussion here in HN a couple of says ago ("Welcome to the new web"), RSS and Atom feeds are becoming less and less supported. Which is a sad and bad thing with regard to the internet of things that is to come with Twine and GreenGoose and Arduino/xbee/sensors.
The appropriate protocols are there. Please use them.
pubsubhubbub doesn't have privacy/authentication/authorization mechanisms as part of the protocol, so creating near real-time push APIs isn't just as simple as using some pubsubhubbub library. I think its the complexity of coming up with one's own auth system that stops a lot of people from using pubsubhubbub. However, it is pretty awesome for things that are published to everyone.
I've seen a lot of people talking about Twine today (HN, Twitter, TC), but I haven't yet heard any use cases that are actually useful. I am certain there are plenty of great uses for this, but I haven't seen any yet.
I don't want a tweet when my laundry is done. It's pretty unlikely my basement will flood. How could this add value to my life, say once a week?
I don't know you, but I imagine lots of people who could use a bunch of these for a summer home or winter cottage. Temperature, did someone enter the door, is it dry... Good things to be notified about.
So it hopes to become an alarm system for everything, not just robbery and fire. The smartphones are replacing specialized devices like cameras and music players, so maybe there's an opportunity for a generalized alarm system. People want to listen their babies, prevent kids from going to the pool alone, and this anti-flooding alarm seems quite unique too. Certainly the pods should alert us when the batteries are empty, and it should be an install-and-forget type of thing.
We'll see if it's the iPhone or the Newton of alarms systems.
Great idea. This is obviously not the final step for this product. It is a great concept with a great deal of potential (basically any domotics application) and it will probably have a large potential user base. Think of all the DIYers, techies, semi-techies and others who want to create their own way of domotics.
Maybe you want to get a tweet when your laundry's done [...]
Who would actually do this? Why would I use Twitter to send a private notification like that? It's pretty much the worst system imaginable. Twitter works great for broadcasting to people, but it's nonsense for one-to-one messages.
Is this really how I want to set up my washing machine:
1) Create a Twitter account for the machine. Let's call it @pavlovs_2011_electrolux.
2) Subscribe my public Twitter account to @pavlovs_2011_electrolux. (Alternatively, maybe the washing machine can be configured with my public Twitter account so it can send messages to me.)
3) Keep an eye for my washing machine's notifications amongst a sea of disorganized tweets about public topics.