Interaction design Project Research Ubicomp

The address book desk

Underneath the desk I have stuck a grid of RFID tags, and on the top surface, the same grid of post-it notes. With the standard Nokia Service Discovery application it is possible to call people, send pre-defined SMSes or load URLs by touching the phone to each post-it on the desk. On the post-its themselves I have hand-written the function, message and the recipient. This is somewhat like a cross between a phone-book, to-do list and temporary diary, with notes, scribbles and tea stains alongside names.

Initial ideas were to spraypaint or silkscreen some of the touch icons to the desk surface, and I may well do that at some point. But for quick prototyping it made sense to use address labels or post-it notes that can be stuck, re-positioned and layered with hand-written notes.

This is an initial step in thinking about the use of RFID and mobile phones, a way of thinking through making. In many ways it is proving to be more inconvenient than the small screen (particularly with the occasionally unreliable firmware on this particular cover, I can’t speak for the production version). But it has highlighted some really interesting issues.

First of all it has brought to the forefront the importance of implicit habits. Initially, it took a real effort to think about the action of using the table as an interface: I would reach for the phone and press names to make a call, instead of placing it on the desk. But for some functions, such as sending an SMS, it has become more habitual.

SMSes have become more like ‘pings’ when very little effort is made to send them. At the same time they are more physically tangible: I rest the phone in a certain position on the desk and wait for it to complete an action. The most useful functions have been “I’m here” or “I’m leaving” messages to close friends.

I have had to consider the ‘negative space’ where the mobile must rest without any action. This space has potential be used for context information; a corner of the table could make my phone silent, another corner could change my presence online. Here it would be interesting to refer to Jan Chipchase’s ideas around centres of gravity and points of reflection, it’s these points that could be most directly mapped to behaviour. I’m thinking about other objects and spaces that might be appropriate for this, and perhaps around the idea of thoughtless acts.

If this was a space without wireless internet, I could also imagine this working very well for URLs: quick access to google searches, local services or number lookups, which is usually very tricky on a small screen. Here it would be interesting to think about how the mobile is used in non-connected places, such as the traditional Norwegian Hytte [pdf].

This process also raised a larger issue around the move towards tangible or physical information, which also implies a move towards the social. As I was making the layout of my address book and associated functions, I realised that maybe these things shouldn’t be explicit, visible, social objects. The arrangement of people within the grid makes personal sense; the placement is a personal preference and maps in certain ways to frequency and type of contact. But I wonder how it appears to other people when this pattern is exposed. Will people be offended by my layout? What if I don’t include a rarely called contact? Are there numbers I want to keep secret, hidden behind acronyms in the ‘Names’ menu?

It will be interesting to see how this plays out and changes over time, particularly in the reaction of others. I’ll post more about the use of NFC in other personal contexts in the near future.

h3. The making of…

The desk is made from 20 mm birch ply, surfaced in Linoleum. I stuck a single RFID to the underside, in the place that felt most natural. A 10 cm grid was worked out from that point, and the RFIDs were stuck in that grid, and the same worked out on top. If I were to re-build the desk with this project in mind, the tags should probably be layered close to the surface, between the ply and Linoleum. This would make them slightly more responsive to touch by giving them a larger read/write distance.

p(caption). Rewriteable 512 bit, Philips MiFare UL stickers.

p(caption). 10 cm grid of tags on the underside of the desk.

p(caption). Blank post-it notes on the surface, with the same grid.

More photos at Flickr.

Interaction design Mobility Research Ubicomp Usability

Nokia 3220 with NFC

h3. First impressions

Overall the interaction between phone and RFID tags has been good. The reader/writer is on the base of the phone, at the bottom. This seems a little awkward to use at first, but slowly becomes natural. When I have given it to others, their immediate reaction is to point the top of the phone to the tag, and nothing happens. There follows a few moments of explaining as the intricacies of RFID and looking at the phone, with it’s Nokia ‘fingerprint’ icon. As phones increasingly become replacements for ‘contactless cards’, it seems likely that this interaction will become more habitual and natural.

Once the ‘service discovery’ application is running, the read time from tags is really quick. The sharp vibrations and flashing lights add to a solid feeling of interacting with something, both in the haptic and visual senses. This should turn out to be a great platform for embodied interaction with information and function.

The ability to read and write to tags makes it potentially adaptive as a platform wider than just advertising or ticketing. As an interaction designer I feel quite enabled by this technology: the three basic functions (making phonecalls, going to URLs, or sending SMSs) are enough to start thinking about tangible interactions without having to go and program any Java midlets or server-side applications.

I’m really happy that Nokia is putting this technology into a ‘low-end’ phone rather than pushing it out in a ‘smartphone’ range. This is where there is potential for wider usage and mass-market applications, especially around gaming and content discovery.

h3. Improvements

I had some problems launching the ‘service discovery’ application. Sometimes it works, sometimes it doesn’t and it’s difficult to tell why this is. It would be great to be able to place the phone on the table, knowing that it will respond to a tag, but it was just a little too unreliable to do that without checking to see that it had responded. The version I have still says it’s a prototype, so this may well be sorted out by the released version.

Overall there is a lack of integration between the service discovery application and the rest of the system: Contacts, SMS archive and service bookmarks etc. At the moment we need to enter the application to write and manage tags, or to give a ‘shortcut’ to another phone, but it seems that, as with bluetooth and IR, this should be part of the contextual menus that appear under ‘Options’ within each area of the phone. There are also some infuriating prompts that appear when interacting with URL, more details below.

h3. Details

p(caption). The phone opens the ‘service discovery’ application whenever it detects a compatible RFID tag near the base of the phone (when the keypad lock is off). This part is a bit obscure: sometimes it doesn’t ‘wake up’ for a tag, and the application needs to be loaded before it will read properly. Once the application is open (about 2-3 seconds) the read time of the tags seems instantaneous.

p(caption). The shortcuts menu gives access to shortcuts. Confusingly, this is different from ‘bookmarks’ and the ‘names’ list on the phone, although names can be searched from within the application. I think tighter integration with the OS is called for.

p(caption). Shortcuts can be added, edited, deleted, etc. in the same way as contacts. They can be ‘Given’ to another phone or ‘Written’ to a tag.

p(caption). There are three kinds of shortcuts: Call, URL or SMS. ‘Call’ will create a call to a pre-defined number, ‘URL’ will load a pre-defined URL, and ‘SMS’ will send a pre-defined SMS to a particular number. This part of the application has the most room for innovative extensions: we should be able to set the state of the phone, change profiles, change themes, download graphics, etc. This can be achieved by loading URLs, but URLs and mobiles don’t mix, so why should we be presented with them, when there could be a more usable layer inbetween? There could also be preferences for prompts: at the moment each action has to be confirmed with a yes or a no, but in some secure environments it would be nice to be able to have a function launched without the extra button push.

p(caption). If a tag contains no data, then we are notified and placed back on the main screen (as happened when I tried to write to my Oyster card).

p(caption). If the tag is writeable we are asked which shortcut to write to the tag.

p(caption). When we touch a tag with a shortcut on it, a prompt appears asking for confirmation. This is a level of UI to prevent mistakes, and a certain level of security, but it also reduces the overall usability of the system. With URL launching, there are two stages of confirmation, which is infuriating. There needs to be some other mode of confirmation, and the ‘service discovery’ app needs to somehow be deeper in the system to avoid these double button presses.

p(caption). Lastly, there is a log of actions. Useful to see if the application has been reading something in your bag or wallet, without you knowing…