smart goods - the story so far · 2006-03-26 18:13

This post has come from several directions – seeing Bruce Sterling talk and reading his latest book, Shaping Things, getting a nabaztag rabbit, and reading the current frenzy about the Internet of Things. This has ended up being part review, part history, part design recommendations.

We’re seeing a new breed of devices – not white or brown, nor gizmos or gadgets. They barely do one thing, at most two. Most of the time they sit there, seemingly dormant, maybe dead. They’re smart goods, for want of a better classification.

I don’t think of these things as spimes – Bruce Sterling’s purposeful neologism (I disagree, too, with the semantics of the word). Spimes have 3 or 4 different concepts rolled into them, including a self-referential idea of how and who constructed, that smart goods don’t have. Spimes are transparent goods.

Smart objects, however, are translucent goods, and not just for their current wave of sub-iMac styling. They convey information from other places, people and things, but often not in the original form. They are a lens for network information: distorting, focussing and signaling.

Quite simply, a smart good is a very expensive doorstop until it’s plugged into a network.

The clock is perhaps the object closest to a smart object without being one (only by virtue of quartz precision, and those controlled by atomic clocks via radio broadcasts mean they are truer smart goods). It conveys a useful piece of information, and is an object placed into many environments. Without telling it what information to display (the current time as a convenient handle on timezones), it is useless.

It’s an exciting time, because they are coming out of research labs and into consumer products. The first was the Ambient Devices orb – the company now has a larger product line of different displays. The orb, and indeed the whole company, was probably slightly too early for the market. They built their own radio network in the US, which unfortunately means the products don’t work elsewhere. It does mean they’re easier to use out of the box (including some crude location detection), but you are tied to them as an information provider, and all the products are completely one-way (receiving information). Newer companies are piggybacking on the growing pervasiveness of home and office wifi, and using the Internet as information delivery mechanism.

Something Ambient Devices does get right is the ambientness of these objects. They shouldn’t be flashy or call attention unless necessary. A smart good should only just about be seen or heard. They’re also glanceable – information is quickly conveyed even with one passing look at the object. This still holds true for their more complex devices, such as the 5 day Weather Forecaster, but appears to have disappeared in their future-looking concepts, such as the e-ink weather forecaster (PDF).

Microsoft’s SPOT watches (now MSN Direct) fall into many of the same traps as Ambient Devices, as well as being even more banal, with simple services based on mass media. Expensive subscription, too.

So, onto wifi rabbits – the Nabaztag. I will state that when I first saw the rabbit I was intrigued but dismissed it as “rubbish”, mainly because at the time, you were tied into the services operated by the rabbits’ maker, Violet. I had another look when I saw that Timo had bought some, and told me an API was available. This, well, changed things. I saw Nabaztag as a relatively cheap ambient information display with 5 colour-programmable LEDs, and 2 motors. So, I ordered one, and I’ve been living with it for a few weeks.

There are some free services available direct from Violet, such as a speaking clock and weather that are fine enough, but have issues: the clock is 5-10 minutes late speaking the hour change, and the weather service, which uses the three belly lights, is in no way ambient, constantly flashing. I don’t see much point subscribing for any other services, especially if they’re as ‘noisy’ as the free ones.

Setup is pretty easy, if it’s an open wifi network – plug in and go, then register on the website, using the MAC ID from the rabbit’s bottom. This was a bit flaky, and I got two connection messages through the rabbit in the wrong order (others have had bigger problems with registration flakiness). I soon realised that as well as the LEDs and ears/motors, it also has a quite loud speaker, and you can throw it MP3s to play.

The API (PDF) is moderately easy to understand – as an owner, you have to get a token and turn on reception of external events. It is, however, lacking certain elements that would make it far more useful. You can send it a variety of different events – rotating ears, messages from the stock library or that you have created, text to be spoken, and ear/LED choreography. Unfortunately, most of these come through as ‘messages’ which means the rabbit plays a tune before and after, and are archived and replayable (with the Nabaztag flashing its nose until you clear them). Not exactly ambient. The other limitation is that ‘choreographies’ are non-repeatable, so it is currently impossible to replicate the constant information displays that Violet offer themselves, such as the weather. I suspect you could hack it with an ultra-long choreography corresponding with a server that repeated the command every 5 or 10 minutes, but it will suffer from the ‘message arrived’ song and dance each time. This basically means only the ears can act as constant information displays.

One possible salvation is an open source proxy server that is being written to control the rabbit directly, or filter and change commands coming from the Nabaztag server. This is made possible because of the wifi settings you can make manually to the rabbit, to connect it through some firewalls (such as at work). Note that if your work wifi network needs VPN, however, there’s no way to make the rabbit work (without running a local server and opening up many horrendous security holes).

Despite these quibbles, what Violet get right, though, is the thingness of the thing. The choice of rabbit is right, the industrial design is right, and some of the actions of the rabbit (such as the breathing LED at the rabbit’s base) is right. It’s a boundary object, a conversation piece, and a statement of intent. When sitting dead at work, Nabaztag still attracted attention. I wish I could have taken pictures of people’s faces when they asked what it was – “It’s a wifi rabbit”... “Errrr, oh, umm”. Putting a few pictures on Flickr has so far mobilized at least 7 different friends to buy rabbits too. I had intended to take the cover off as soon as I got it, to turn it into other information displays, but it’s keeping its rabbit overcoat for now.

This is one other slight failing with Violet’s vision of what the Nabaztag is and how it works. There’s the usual Warranty Void if Opened, No User-Serviceable Parts Inside, and devious triangle headed screws that you might imagine from typical consumer goods. I want to hack, I want to personalize (more than the ears), I want to reconfigure. The API is a start towards openness, but it’s clear that the hardware is meant to be locked. Also, the ear motors don’t appear to be that strong, so putting a Nabaztag in as the motor of some more complex ambient display may not be possible.

That said, it’s time to get hacking. Friends, Coders, Nabaztags – lend me your ears! And your LEDs too! Yoz has started hacking a Ning app to allow even easier interface to your bunny, and I’ll be hacking around that in the next few weeks. I noticed Phillip Torrone has started playing with one too. What I would hope is that Nabaztag hackers can crystallize into one community – most of the current API discussion is in French, which is fine, but my schoolboy language skill don’t quite get me far enough. I propose using the English part of one of the Nabaztag owner boards – let’s meet there and try to bend the bunny into doing new things.

Click here to make my bunny’s ears wiggle. Go on. You know you want to. (note to self/Yoz – it will also wiggle its ears whenever a search engine tries to hit the URL. Oh Well.)

Was it good for you?

So, the current Nabaztag isn’t the be-all and end-all of smart goods. Distilling the above reviews into design and experience principles, ideas and areas currently missed:

Sell goods not services
I want to buy an object, and for it to work. Completely.
This also means not having to setup software and servers on computers to work.

Inputs as well as outputs
Everyone is thinking about outputs, but inputs are just as important. This ties into work such as that for NFC by Timo and Janne.

Ambient, most of the time
The joy of these devices is in the alerting – not on or off, like computers and phones tend to alert, but somewhere in between.

User-servicable parts inside
Let us fiddle and change, both hardware, software and servers. Be as open as possible, and you might establish yourself as the target platform.

Be good or get commoditized
With mobile phones, we often think about running out of pockets (which is why incorporating a camera, for example, into a phone makes so much sense). With smart goods, we’re fighting shelf space – inhabited by TVs, clocks, 3 games consoles, VCR, DVD… any and all of these could incorporate smart object functionality, especially information display.

History, and real time
Don’t ignore time. Most current devices show current state or future predictions (especially for weather). Displays of history and change are just as important and powerful.

Some links:
hat tip to Mike Kuniavsky, who has been following this area for longer than most
Adam’s doing sterling work currently about everyware
Anne remembering links about the Internet of Things
I’m collecting Nabaztag hacking links on delicious

comments

<< previously home furthermore >>

© Chris Heathcote, 2002-2010. I write for me, not for them.
CSS styling from plasticbag.org