Thursday, September 3, 2009

Logjects: objects that record their own use

When Mack posted about his Kindle last week, one of the questions I asked him was whether it was a logject – that is, it recorded its own use with the possibility that the information gathered about his reading habits was transmitted back to Amazon. Amazon already gathers data concerning the purchases we make and if Kindles are networked logjects then they can marry that data up with what we actually do with those purchases (including our reading habits and what information we add to them, such as making notes). I’ve no idea if Kindle is a logject, and if it is what kind of logject it is, but I said I’d write a post explaining that neologism (a term that Martin Dodge and myself coined through a refinement and expansion of Bleeker’s notion of a blogject). I'm sorry if what follows is a bit academic, but it's culled from a paper we had published in Environment and Planning A earlier in the year (working paper here) but hopefully it might prove interesting because it concerns how our everyday objects are being remade to do different kinds of work in the world, including recording and communicating their use (and I'm sure they will increasingly turn up as plot devices in novels in the coming years).

The easiest way to introduce the notion of a logject is to consider the ways in which the embedding of software is changing the nature of everyday objects. Many objects now have software physically embedded into their material form. On the one hand software is used to enhance the functional capacity of what were previously ‘dumb’ objects enabling them to sense something of their environments and to perform different tasks, or the same tasks more efficiently, or to be plugged into new distributed networks that afford some value-added dimension such as data exchange on how they are used. On the other hand, software is used to underpin the design and deployment of new classes of objects, particularly mobile devices (such as PDAs, MP3 players, Satnav), that in some cases replace analogue equivalents (diaries and filofaxes, personal tape and CD stereos, paper maps and gazetteers) or undertake entirely new tasks. In some cases, software is not essential to an object’s use (such as digital clock in a cooker), but in others software has become vital to the performance of the object. The latter Martin and myself call codejects, of which there are three types differing with respect to their programmability, interactivity, capacity for remembering, their ability for anticipatory action in the future based on previous use, and their relational capacities. In summary these classes are:

Hard codejects that rely on firmware to function but are not programmable and therefore have low levels of interactivity (e.g. a USB memory stick).

Unitary codejects are programmable, exhibit some level of interactivity, although this is typically limited and highly scripted, and they do not record their work in the world. They can be subdivided into two groups:
(a) closed codejects that work independently of the world around them (e.g., alarm clock, CD player)
(b) sensory objects that sense and react to the world around them (e.g., a heating control unit using a digital thermostat).

Logjects are objects that have an ‘awareness’ of themselves and their relations with the world and which, by default, automatically record aspects of those relations in logs that are stored and re-used in the future. More specifically, a logject is (1) uniquely identifiable, (2) has awareness of its environment and is able to respond to changes in that environment that are meaningful within its functional context, (3) traces and tracks its own usage in time and/or space, (4) records that history, (5) can communicate that history across a network for analysis and use by other agents (objects and people), (6) can use the data it produces to make automated, automatic and autonomous decisions and actions in the world without human oversight, (7) is programmable and thus mutable to some degree (that is, it is possible to adjust settings, update parameters and to download new firmware). They can be subdivided into two groups based on their capacity to work independently of wider networks:

(a) permeable logjects - relatively self-contained units such as a MP3 player, a PDA, and a satnav (all of which have the potential to be connected to wider networks). Such devices trace and track their usage by default, recording this data as an embedded history; are programmable in terms of configurable settings and creating lists (e.g. play lists of songs, diary entries and route itineraries); perform operations in automated, automatic and autonomous ways; and engender socially meaningful acts such as entertaining, remember an important meeting and helping an individual to travel between locations.

(b) networked logjects - these logjects do not function without continuous access to other technologies and networks. Such logjects track, trace and record their usage locally but their full histories are also recorded externally to its immediate material form. Some networked logjects are relatively fixed in the environment (e.g. satellite/cable television control boxes, home security monitoring systems) and others are inherently mobile (e.g., mobile telephones, telematically monitored vehicles) that use a range of communication technologies such as GSM, Wifi, Bluetooth to maintain a network connection.

The other way objects are changing is that they are becoming uniquely identifiable through new tagging, so it becomes possible to know the logistics and purchase history of every item, but that's for another post.

I suspect that a Kindle is at the very least permeable logject, recording its own usage. The key question with respect to surveillance and privacy is whether it’s a network logject communicating back to Amazon details of our reading patterns and our notes? If anyone has the answer I’d be interested to know.


Maxine Clarke said...

Fascinating. Is this the same mechanism as semantic matching? Amazon uses this type of matching, I believe (but am happy to stand corrected) for showing you things you might be interested in buying at the website or via is this Kindle logject you discuss an extension of that, or does it operate by some other mechanism?

Of course the semantic web is the next generation of web as envisaged by Tim Berners Lee, in which (aided by xml tagging as you mention as one category in your post) the web itself searches out connections and relationships, eg within and between huge datasets - using "learning" technologies. This way, connections and directions emerge that are beyond the capacity of individual analysis by a human (eg metagenomics).

At the moment, we seem at a rudimentary stage of all this - given what little xml tagging is actually done by content providers. (I was only reading today an article in the Chronicle of Education about how Google book search is a disaster for academics because it will swamp all competition and itself has such poor metadata - I am not so pessimistic as I believe that if a niche is there, eg by Google book search not being based on good metadata - then someone else will fill that niche - but it is a good point.)
I don't quite understand the relationship between deliberate xml tagging and the semantic web (and logjets?) but it is all fascinating.

Bernadette said...

My inner conspiracy theorist is a wee bit concerned with such things. In my work (IT in the Health sector) there has been much talk of the the possibilities of logjects - to an extent they are used already (e.g. a fitted pacemaker can tell us much about its human when it is serviced or otherwise 'read') but there is always talk of doing much more, especially with people who have chronic diseases or conditions. Because at the moment I work in the public health environment there is much discussion of "us" having a "right" to knowledge about the clients because they're using public funds. I can see a slippery slope from us knowing when their pacemaker is working hard to giving someone some kind of electronic nudge if they eat something we don't think they should eat. I've been very wary about introducing potential logjects into my personal life. Just in case :)

Rob Kitchin said...

Maxine, the semantic web is about how information is processed and used, sometimes for the benefit of the user by giving recommendations etc or for the benefit of the provider by supplying useful information for directed marketing etc. What logjects are about is collecting and recording 'useful' information. There are a number of objects that presently collect information and communicate that information back to a service provider - mobile phones are the classic example, but as Bernadette says it also includes things such as pacemakers. What that means is that much more of our lives are being opened up to surveillance often without us really knowing about it. And rather than it just being states who do this, often the data is being harvested by businesses. As Bernadette notes this raises all kinds of questions about privacy and the 'right' to intervention and policing. In our paper, what we do is undertake an audit of three different households to assess the extent to which software has been embedded in different objects and what that software does. You should try it yourself, you'll be amazed at how many things in your home now have code either core to or augmenting their use. The question is, what is that code used for or could potentially be used for? It's these kinds of questions that I'm interested in, as well has how the information generated is organised, processed and utilised.

Unknown said...

nice blog....

Bernadette said...

I am always gobsmacked by the level of privacy people willingly give up these days by introducing these items into their lives. It's mostly the early adopters at present but I'd have thought they'd be the most likely to see the risks inherent in living so publicly but they seem to just like the shiny new toys and forget about the risks.

Maxine Clarke said...

Thanks, Rob and Bernadette - and agreed. As well as the "surveillance" concerns there is also the continual issue of whether the system actually works correctly (eg is providing correct info), what happens when it breaks down, etc. (Breaking down is OK in some circumstances - the feedback just stops - but not in others eg as you say Bernadette a pacemakder that might be part of a treatment-based electronic feedback loop.)

Rob Kitchin said...

There are now telemedicine packages that literally monitor all aspects of an older person's home including room movements, bed occupancy, appliance use, food consumption, and so on. The idea is allow independent living for as long as possible, but to be monitor for signs of trouble (such poor eating or little movement) and able to react to any problems. Looking at the solutions available one can't help wondering if it is the precursor to something else. Everything the person does is being tracked from a monitoring centre. As an example have a look at some of the products at