service design for the cloud of diverse devices open desktop summit 2011 @clurr Monday, 8 August 2011 i’m claire. i’m here to talk about designing for an ecosystem of multiple devices. i’m afraid i’m a slave to the proprietary empire right now. i’m not a hacker, and most of what i do what tends to be within the bounds of what steve thinks is acceptable. but that doesn’t mean i don’t believe we need alternatives.

here i showed an introductory video on interusability and the smarcos project see it at http://vimeo.com/24511479 Monday, 8 August 2011 i’m not a technologist. i used to be a UI designer but i don’t really even do that anymore. i’m a researcher, which means i specialise in understanding how people use computers, and how to make them more user friendly. these days, my particular interest is in working out how to design for the ever increasing range of devices around us. here’s a video we produced for a project we’ve been working on recently that i think provides an introduction to this...

credit photo: johan nilson fun? Monday, 8 August 2011 Most of you are probably the kind of people who are quite happy fiddling around with technology to get it to work. You like looking under the hood, and enjoy plugging bits into other bits. That’s a tough thing to do - we need people like you. But most people, including me, are not like that.

fun :) Monday, 8 August 2011 Most people may enjoy watching pretend people fiddle around with pretend technology, but that’s about as far as it goes. They’re getting comfortable with increasing levels of technology around them, but they don’t want to have to become technical experts in order to get it to work. --- images from: http://www.pcmag.com/slideshow_viewer/0,3253,l%253D255511%2526a%253D255510%2526po%253D10,00.asp?p=n dr who image from : http://pauseliveaction.files.wordpress.com/2011/05/tardis-int.jpg

6 devices per person by 2020? Monday, 8 August 2011 Trouble is, the amount of technical stuff we have to get to work is increasing., as you saw in the video. Smartphones now outsell PCs / TVs are web and app platforms / Cars are becoming connected If we get to 50 billion devices, that’s 6 devices for every person on earth. Most of those are not going to be cars and mobiles, or even things with screens they’re going to be all kinds of random embedded devices: household appliances, environmental sensors and controls, machines in factories, pill bottles, dog tags, toilets... you name it. They might not all be things we interact with directly, but the reality is that there’s going to be a lot more of the complicated plugging things into other things type stuff that we have to do. That’s a lot of interconnectivity. But, most people don’t care about interconnectivity. They just want stuff to work. The more devices we have, the more the interplay between those devices becomes a key part of the user experience. That’s what I’d like to talk about today: designing experiences across devices. http://www.readwriteweb.com/archives/smartphones_outsell_pcs.php http://blog.chron.com/newswatch/2011/06/world-to-have-1-networked-device-per-person-by-2012/ http://mashable.com/2011/02/26/connected-car/

service design photos: amanda vincent-rous, adam lang, giving not@rocketmail.com, li’l wolf Monday, 8 August 2011 I call what I do service design. Services are hard to describe easily. They are intangible things that provide a benefit to someone: like a product, but not concrete. The postal service is a good example. Delivering a service requires the coordination of lots of components and people to ensure an effective experience. We interact with it across multiple touchpoints, but we have a perception of the postal service as being a coherent entity. All the touchpoints are easily recognisable to us as part of the service, and we have expectations as to how they will work together, and what benefits we will get as a result. End users usually perceive services to be one thing: when people’s mail goes missing, they get angry at the guy in the post office, even though it was probably nothing to do with him.

this is not an app Monday, 8 August 2011 Services can just as well be digital. This is the Withings bathroom scale, and iPhone app. By a service designer’s definition, the app is not an end in itself. it is a touchpoint for a service: in this case, one that tracks how much you weigh. Of course, there is an app... but the value is in the service as a whole: how the parts work together. In this case, it works well: data is shared seamlessly between devices, and setup is really, really easy. That’s a service experience. It doesn’t matter how good the app is in isolation if the service as a whole doesn’t work well.

we need a new design metaphor photo: buffetoblog Monday, 8 August 2011 Digital service design becomes important when you’re dealing with multiple devices. It’s a way of thinking about the glue in between. But much of the nuts and bolts of user experience are based on the old idea of one user, sitting at a single computer, doing one thing at a time. HCI experts, like Don Norman, have been talking for years about the rise of the information appliance, but we still haven’t figured out how to design for that well.

old skool usability • one core device, with a screen • work-centric design • context-independent • static, passive: waits for the user • application centric • one user at a time • fat client Monday, 8 August 2011

interusability • multiple devices, some without screens • not just work anymore • context-aware: has potential to be proactive • becoming content and activity centric • potential for multiple users • highly cloud dependent old skool usability • one core device, with a screen • work-centric design • context-independent • static, passive: waits for the user • application centric • one user at a time • fat client Monday, 8 August 2011

service UX interusability touchpoint UX cloud UX Monday, 8 August 2011 Interusability is UX across multiple levels: the service touchpoints on the devices, the cloud, and the service experience itself. this is a new area for design, but we’ve been getting some insights into what it might mean to create a good touchpoint, a good service, and how the cloud affects user experience.

what makes a good touchpoint? Monday, 8 August 2011 fundamentally, it’s about context. who’s the user, what’s the activity, what devices are available, where are they being used, under what circumstances...

appropriate to the device Monday, 8 August 2011 I’m not going to talk about device usability in detail, that’s a whole other world. What service designers are interested in is doing the right things on the right devices. Each type of device offers its own capabilities and limitations. Not every device in your service needs to do everything, even if it can...

demands only as much attention as it needs Monday, 8 August 2011 if you’re surrounded by devices, there’s a lot going on. we tend to design interfaces as if they need to have our full attention but that doesn’t scale. sometimes, they need to get out of the way. this is an NFL stats app for the ipad, designed by my fjord colleague brian mclaughlin. it’s designed to be glanceable, so you can watch it during a live game. (this is why interactive football tv failed: it gave people too much choice)

does exactly what’s needed (and no more) Monday, 8 August 2011 this is a beta of starling, a social service for live tv watching. it looks a bit rough right now but it’s got some clever designers behind it, so let’s assume they’re going to fix the visuals. what’s interesting is that the designers limited messages to 50 chars - they realised that much of what people wanted to say around tv programmes was OMG and WTF??? and even !!! in the right context, that’s highly meaningful. the constraint supports watching live tv and you don’t want to switch too much attention away from the screen. designed to enhance, not detract, from the live tv experience. . http://www.fastcompany.com/1656531/social-tv-apps-starling-google-tv-miso-tunerfish (twitter 140)

Monday, 8 August 2011 we need to get better at prioritising the features users need there, and then. Take smartphones. Right now, most of them have a screen that looks more or less like this. It’s a mini desktop, or, as I call it, a bucket of apps. Solutions like multiple homescreens, like Launcher Pro here, just make the bucket a bit bigger. you can organise them, but they’re all just sitting there passively, waiting for you to pick one and tell it what you want to do. That’s not always inappropriate, but it is a bit clunky. Mobiles have the potential to know stuff about you. They have sensors, for location, movement... and they often hold your calendar and contacts.

anticipates needs Monday, 8 August 2011 But maybe we can do better. Can we use some of the mobile’s capabilities, to understand the environment, anticipate what users need, and be proactive in bringing it to them? Scott Jenson from Frog gives a good example... Say you’re at the bus stop. if you want to use your phone to find out when the next bus is coming, you have to open an app, find the bus line, find the route, and find the stop. it’s a faff. Right now that’s doable in a number of ways, but this kind of experience is not the norm.

surfaced at point of need Monday, 8 August 2011 we can also embed triggers to initiate the interaction in the environment: linking the physical and digital world in a much more natural way. this is washtech... a system created by Nokia Research Centre and Deep Springs International to ensure people in haiti who do not have access to clean water are correctly chlorinating their drinking water. fieldworkers used to use paper forms, but they were slow and mistakes were easily made. now each bucket has an RFID tag. passing an NFC enabled phone over this identifies the bucket, brings up a list of relevant questions the fieldworker needs to ask, before responses are sent over SMS to headquarters. http://washtech.wordpress.com/2011/05/01/battling-cholera-with-nfc-rfid-tracked-drinking-water-in-haiti/ Sure, we’ve had the tech to trigger interactions for a while with QR codes and Bluetooth but these are awkward. NFC is a much lower friction enabler, and we’ll soon start to see it used more in the next year, not just for payment but as a way of embedding interactions in the environment (not just paying for things).

what makes a service feel like a service? Monday, 8 August 2011 Informed in part by: Minna Wäljas, Katarina Segerståhl, Kaisa Väänänen-Vainio-Mattila, Harri Oinas-Kukkonen: Cross-platform service user experience: a field study and an initial framework. Mobile HCI 2010: 219-228

clear mental model Monday, 8 August 2011 every service needs a clear mental model. that’s understanding how the service is made up, what does the service do, what devices do you need? (nike+ example). what device does what (which we call composition)? this isn’t just about communicating it. it’s about designing around a clear metaphor. dropbox has arguably been more successful with consumers that other syncing services not through better technology but a clearer mental model - “it’s a shared folder”. the proposition is not “a service to back up and sync files” - these can seem like scary, complex technical things to non technical people.

continuity... Monday, 8 August 2011 You need to be able to switch from one device to another without losing where you are: maintaining the state of the interaction so you don’t need to repeat stuff. There aren’t many good examples of this yet. Simple example is Echofon: registers which tweets have been read on one device, doesn’t download them to other devices. In UX terms, that’s like having one experience of your Twitter stream instead of several, very repetitive ones. But that’s really just identifying what you’ve read. In UX terms, it’s no cleverer than email. A really clever cross-device system would work out what the user is doing and enable them to switch devices at any point, without losing anything. Or even to switch modalities.

...continuity of tasks and data Monday, 8 August 2011 There’s a team at CNR ISTI in Pisa working on migratory interfaces: tools that enable users to switch devices on web apps without losing anything. The examples here are demos and not the prettiest but this is pretty cool: here the user’s been viewing their bank account on the mobile browser, but the phone is running low on battery. So it offers the option to push everything to a nearby PC. why can’t we use something like this to ping a video you want to share from your phone to the nearest tv?

...continuity across interaction modes Monday, 8 August 2011 As devices have different capabilities, they can even migrate across interaction modes: e.g. this is a half completed cinema booking on a web page, which can be switched to a voice interaction program without losing state. The system has built a model of the user task which is independent from how it is presented. Technically these may look like very different types of challenge, but from a UX perspective they are more or less the same: maintain my interaction with the service, no matter which device I’m using, or it’s capabilities. Key here is the same data, same interaction logic, and clearly signposting links to other devices - e.g. an alert on the TV says ‘check your phone’, and an alert on the phone says ‘I’m taking over’.

consistency Monday, 8 August 2011 consistency means using the same design elements across devices. that could be the same functionality, layout, visual styling, words, labels, and logical structure to tasks. this is a challenge when a service works across different devices with different UIs and capabilities: for example, a PC, a touchscreen phone, and a household device. certain elements need to be the same, or it won’t feel like the same service and you’d have to learn how to do everything from scratch on every device. but certain elements won’t make sense, or feel right on different devices. consistency for consistency’s sake isn’t good. a big part of the challenge in cross platform design is working out what’s appropriate, and the more different types of device you have, perhaps some without screens, the harder that gets. we’re still investigating this but my feeling so far is that the most important thing is to be appropriate to the device. your phone does not have to pretend to be a washing machine in order to control a washing machine. you don’t have to replicate the entire washing machine UI on the phone, push buttons, dials and all. the phone doesn’t have to make the same beeps as the machine - it should make phone like noises (perhaps a sound that’s associated with other home automation services you have). but you should at least label the wash programs the same, and maintain the same logic of tasks: don’t turn w machine on by entering program then go on device but other way round on phone). and of course, doing the same thing has to have the same result. [music software - represent a mixing desk, or use ableton live - completely different metaphor] ---- studies show there is some tolerance for inconsistency. it’s secondary: it helps the service feel like a coherent experience. According to a study by Wäljas et al [1], users do not comment on consistency; it may be a secondary factor contributing to the overall coherence of user experience.

what’s good cloud UX? Monday, 8 August 2011

graceful degradation of connectivity photo: ben sutherland Monday, 8 August 2011 connectivity will only get better, but will never be perfect. if a service relies on connectivity, and that’s not available, then from the end user’s perspective it doesn’t work at all. we haven’t figured out how to do this well. spotify provides an offline mode, but if you lose connectivity, it cuts the music and then forces you to log in again. but the more we shift data into the cloud, and the more we expect to share, and collaborate, around that data, the worse our experience is hit by loss of connectivity. it’s never going to be a good experience but we need to find ways to mitigate it. those may depend on the nature of the service. for example, can you support planning ahead, by downloading data before it is needed? Ovi maps for Symbian stores map data offline, whereas google maps does not. can we even try to anticipate poor connectivity?

control of data Monday, 8 August 2011 who can access my stuff? we understand access through the physical control of having it on a device, but that means nothing anymore. getting normal people to understand privacy, or maintaining control of their data, is hard. most of them can’t even manage their facebook privacy settings correctly. how do they cope when all their data is in the cloud, with different access privileges? we have to help them. it has to be really clear what’s shared, and with who, and what can be done with it. and above all, we have to try to set sensible defaults because most people don’t change them! recent example: fitbit.

control of data Monday, 8 August 2011 poor jeff probably didn’t expect this was going to end up on the internet. being indexed by google. if we do nothing else, we must set sensible defaults that minimise harm!

what’s the desktop got to do with all of this? Monday, 8 August 2011

can it be part of a free ecosystem? Monday, 8 August 2011 our interactions with technology are distributed across many more devices. we still need multi-function devices like PCs and smartphones, and they still need OSs & UIs. but they are parts of an ecosystem, and we need to learn how to design for the connections between the things in that ecosystem. it’s easier to design good services if you control the whole ecosystem, hardware and software. for that reason, i have more apple devices in my living room than i care to admit to you and they work quite well together. but frankly even i don’t want my entire life run by apple, or microsoft, or samsung. but if you believe in alternatives to the proprietary desktop, you really need to believe in alternatives to the proprietary ecosystem taking over your home, and work, and whatever else. it’s much harder to create a great service UX when you are having to integrate hardware and software from different sources but that’s what we have to learn to do better, and that’s a great opportunity for the open source community. i was really excited to see the plasma active talk earlier. an increasing number of embedded devices run linux. are there conduits for PCs, tablets, phones to talk to them? how do we design those conduits so they make sense to consumers? the technical challenges here are outside my knowledge, that’s your area. but i think these are some of the most interesting challenges in UX right now.

thank you @clurr claire@clairerowland.com big thanks to: helen le voi, martin charlier, daniel soltis, charlie gower, adam crickett, giles turnbull and the SMARCOS project partners: www.smarcos-project.eu Monday, 8 August 2011