My two previous blog posts were about the two workshops I attended before and after the recently held 4th International Conference on ICT for Sustainability (ICT4S 2016) in Amsterdam last week. This blog post is about the conference itself.
I want to start by pointing out a meeting I had at the workshop Elina and me organised the day before the conference. I was really very inspired by meeting Norberto Patrignani. I had become aware of his (and Diane Whitehouse's) paper about "Slow Tech" earlier this year and became all the more interested in that concept (and their paper) after talking to Patrignani personally. Patrignani and Whitehouse are apparently working on a book about slow tech right now and I look forward to it (perhaps next year?). With inspiration from the Slow food movement, Slow Tech is here defined as good, clean and fair technology (and there are no less than 22 wikipedia pages in the "Slow movement" category including Slow reading, Slow education, Slow design and Slow science!). Good = used for good purposes, to solve real human problems (instead of developing new surveillance or weapons systems or just to increase profits). Clean and fair = ecologically and socially sustainable. I haven't read the paper yet but I feel that this is a very interesting "lens" that harbours the promise of uniting ideals of User-Centered system design with sustainability. I suspect slow tech could be used as a "filter" or a "sieve" to sort out "bogus" uses of technology and "bogus" projects.
Just take the fact that he again brought a "Catchbox" microphone ("The throwable microphone for audience engagement"). The pictures below say it all and I want one (but they are unfortunately really expensive). It's so much less intimidating to ask a question when you can throw a microphone around in the audience instead of having to bother your neighbours and squeeze past them to stand in line for a shot at a microphone in a stand.
I understand Peter has also been involved in planning parts of the format of the conference which again utilised "ConverStations" instead of more traditional paper presentations (see below). The new invention this time around was the "WorkStations" (I attended one, see further below).
The conference venues were great. The workshops were held at the Amsterdam Business School and the conference itself was held at KIT. It was truly beautiful and luxurious - the main lecture theatre had plush comfortable movie theatre seats. I also love conferences where lunch is provided and you don't have to waste time looking for a restaurant nearby (in which direction?) that hosts 10 or 14 persons. It's just so much better when you can eat together and work the room by circulating cocktail-partywise. I've become a lot better at circulating during the last couple of years too and enjoy it a lot nowadays. It's either sink (stay put) or swim (circulate). It's always nice to talk to people you already know, but isn't the purpose of going to a conference to meet new people? The only fly in the beaker was the fact that the conference venue itself, KIT/"The Royal Tropical Institute" (previously called "the Colonial Institute") was some kind of shrine to the benefits of colonialism and it (for example) had beautiful woodwork of the industrious Dutch bringing trade, culture and progress to all the savages of the world.
This is the 4th ICT4S conference and I have been to all the three previous ones (Zürich 2013, Stockholm 2014, Copenhagen 2015). I have written blog posts about the 2013 and the 2014 conferences but did not write a blog post last year. In my 2014 text I was for the most part critical, or at least hesitant about the "ConverStation" format, but while I by now had attended ConverStation presentation in 2014 and 2015, I had not myself presented a paper using the ConverStation format before this conference.
The main idea is that instead of presenting your paper in a plenary session - where half the people will sleep or read e-mail - you present it at a small table that has seats for a limited number of persons (say six or so). I talk for 10 minutes using printouts, slides or perhaps no technical aids at all and then we have a conversation (Q-and-A session) together. There are several advantages to this format as you get a very intimate ambiance as well as for the most part very deep discussions. Those who listen have chosen to go to my ConverStation in competition with 15 others presentations that happen simultaneously. The twist is that I give my presentation three times in a row and so each attendee has the chance to listen to 3 out of the 15 presentations (basing their decision on the paper titles and abstracts). There is a limited number of "tickets" (post-its) for each ConverStation so that attendees are distributed more or less evenly among the tables.
I had some really good discussions at my three presentations (due to excellent, critical and insightful comments from for example Daniel Schien, Patricia Lago and Steve Easterbrook). Much better than the questions and the discussion you usually have after a paper presentation. But, the ratio of presenters/audience was skewed this time around and I only had three attendees listening during each of my presentations and I also missed the opportunity to listen to the other 14 presentations that happened in parallel to my own. I think the ConverStation idea is a good one, but the implementation has to be thought through a little more. I also in hindsight realise that the local organisers could probably have done a better job of working up interest and drawing local people from Amsterdam, The Netherlands and neighbouring countries to the conference. Is it really worthwhile to give a 10-minute informal talk to an audience of only three (no matter how qualified these people are)?
Dawn Walker, presenter (CA), Daniel Pargman (SE), Lorraine Hudson (UK) and Roy Bendor (NL)
The big advantage of the ICT4S conference is that it draws people from various parts of computer science (and beyond) that you'd hardly meet otherwise. I know that the chair, Patricia Lago, loves the fact that she can talk to people she would never meet at the "ordinary" (more specialised) conferences she attends. I like that too. I have also come to love the fact that I get to meet a lot of people I nowadays know but seldom meet as well as the fact that we have managed to rope in a fair number of people from Sustainable HCI (S-HCI) to the ICT4S conference. A list of attendees would have been even better at this point as memory fails me after only a week. I even asked for such a list but it did alas not materialise.
I think all the keynote speakers were ok but my absolute favourite was Frits Verheij (Executive board member at the Universal Smart Energy Framework (USEF) Foundation) who gave a talk about "Why ICT is key in Sustainable urban energy". He had a long track record, he had lots to say and was not shy about saying what was on his mind in terms of what works (or could work) and where things go wrong. I wish I had a copy of his information-rich slides instead of just a bunch of photos! I also love the new term I learned, "prosumerification":
I know that the four parallell "WorkStations" (1.5 hour long workshops) were supposed to connect to each of the four keynote speakers, but I don't think that connection was made very clear. I only knew about it due to the fact that I lived on a houseboat together with my colleagues Mattias Höjer and Elina Eriksson and Mattias was co-responsible for organising and leading one of the WorkStation. His WorkStation was called "The role of the individual" but I instead chose to attend another WorkStation, "The Internet of Things and Big Data for sustainability". That was an... interesting experience. The WorkStation was led by an older tech guy who worked with datacenters and cloud computing (Chief Technology Officer at EMC Netherlands) and a younger Italian post-doctoral researchers who worked in Amsterdam. I don't know if they knew each other before the conferece or how much time they had spent preparing the workshop but what was interesting to the point of bizarreness was their combined enthusiasm about the benefits of using the Internet of Things and Big Data for sustainability. They for sure didn't convince me nor did they convince a polite but critical Roy Bendor and (at least) several others workshop attendees.
We all agree that both people (social media) and things (sensors) nowadays create an avalanche of data. The workshop organisers started out by stating - stating! - that more data leads to more knowledge leads to more wisdom leads to more intelligence. I would want to put intelligence in quotations marks ("intelligence") but they for sure didn't and they defined intelligence as the ability to understand implications and make correct predictions (and by extension correct - not "correct" - decisions). Without at this point asking if anyone had any thoughts or opinions this they instead forged ahead and started off the workshop by asking if we, as an effect of these developments would be able to "predict everything" in the future - a statement that was sorely rejected by those in the audience who spoke up. Bendor ruefully asked: "if we could predict everything, what would say about us?". I guess the predict-everything hypothesis was one of the two organisers' pet idea but it is just too weird to discuss here, a little like Elon Musk saying that the chances are a billion-to-one that we actually live inside a computer simulation).
I voiced my objecttions and referred to the paper I co-authored earlier this year about limits to policy-modeling ("In this paper we are particularly interested in the myth of increased quality, objectivity and truth that emerges from the introduction of [Big Open Linked Data] BOLD within policy modeling"). It's hard to raise complex rather than simple arguments at a fast-paced workshop, text is better for pondering most issues and so I here choose to refute the predict-everything hypothesis by quoting Gregory Bateson on the differences between engineering and physics on the one hand and the social sciences on the other hand. The first sentence below comes from my ph.d. thesis the rest is from Bateson (1972, p.229), "Steps to an ecology of mind":
Traditional engineering disciplines are based on intimate knowledge of how different physical materials react in a variety of conditions (temperature, vibration, torsion, stress, age etc.). Canine behavior – not to mention human behavior – is less predictable, less calculable:
”If I kick a stone, the movement of the stone is energized by the act, but if I kick a dog, the behavior of the dog may indeed be partly conservative – he may travel along a Newtonian trajectory if kicked hard enough, but this is mere physics. What is important is that he may exhibit responses which are energized not by the kick but by his metabolism; he may turn and bite.
This, I think, is what people mean by magic. The realm of phenomena in which we are interested is always characterized by the fact that “ideas” may influence events. To the physicist, this is a grossly magical hypothesis. It is one which cannot be tested by asking questions about the conservation of energy”
Perhaps the most succinct critique of the whole idea of Big Data (by necessity) leading to better decisions comes from Evgeny Morozow who slaughters the very idea of forging a useful link between Quantified Self (Big Data) and causality (prediction):
"Peter Austin data-mined the health records of 10 million Ontario patients to draw some fascination conclusions about them. One ... finding was that "Virgos vomit more, Libras fracture pelvises." ... Austin notes that you only need to "replace astrological signs with another characteristic such as gender or age, and immediately your mind starts to form explanations for the observed associations. Then we leap to conclusions, constructing reasons for why we saw the results we did." However, he argues, "the more we look for patterns, the more likely we are to find them, particularly when we don't begin with a particular question." In other world, what Austin takes to be the mark of bad research has somehow become a defining, beloved feature of the Quantified Self movement."
Another early, controversial and weird position the organisers took was that more data will (causally, automatically) lead to more intelligence and more sustainability. End of the discussion. Except it wasn't. I (and others) for sure were not willing to go along with such a simplified view of the connection between data and "intelligence" and sustainability. Trying to "convince" the organisers to give up or at least reconsider that position became one of the issues we spent the most time on in the workshop (it was hard work and I doubt we were particularly successful). I found it curious that the organisers assumed that increased "intelligence" would always be used in ways that would be beneficial to us as end users, consumers and citizens. I honestly think they had never considered that it could be any other way. I told them I felt as if I might as well have attended a workshop a hundred years ago on the topic "The internal combustion engine for sustainability!".
Another example that was raised (by Dawn Walker from CA) was the case of John Deere tractors that primarily seem to have the interests of the John Deere corporation - rather than their owners (users?) - in their (software) minds. From Wikipedia: "John Deere locks tractors digitally ... to prevent the DIY repairing by the owning farmers" and here's Wired's take on this interesting issue (and here is NPR's):
"In a particularly spectacular display of corporate delusion, John Deere—the world’s largest agricultural machinery maker —told the Copyright Office that farmers don’t own their tractors. Because computer code snakes through the DNA of modern tractors, farmers receive “an implied license for the life of the vehicle to operate the vehicle." It’s John Deere’s tractor, folks. You’re just driving it. ... The pièce de résistance in John Deere’s argument: permitting owners to root around in a tractor’s programming might lead to pirating music through a vehicle’s entertainment system. Because copyright-marauding farmers are very busy and need to multitask by simultaneously copying Taylor Swift’s 1989 and harvesting corn? (I’m guessing, because John Deere’s lawyers never explained why anyone would pirate music on a tractor, only that it could happen.)”
To sum things up, more data and "intelligence" could lead to more sustainability or it could lead to less sustainability. My bet is on "less" if companies can earn more money that way. The workshop organisers resisted even this simple and pretty neutral statement both when the workshop was summed up and (I believe) when they later reported on the results in a plenary session. I'm not very impressed.
One last trail of thoughts that were only partially explored at the workshop came from one of the workshop organisers' half-assed but intriguing idea that since we pay for various services we use with our data (think of Google and Facebook), then data perhaps "could be seen as a currency?". I've come across this intriguing idea before, here's what I wrote in April 2014:
"This made me remember a slogan I've heard; "data is the new oil" (or, "data is the oil of the 21st century"). This is an intriguing notion even though I personally think oil is the new oil (as well as the old oil) because without the oil (energy), nothing will run and data will loose much of its current allure)."
This time around I was at least a little more prepared for thinking about this particular issue and immediately started to ask piercing questions like "what do you mean?" (that for the most part went unanswered). Building on a great short text by Belgian economist and money theorist guru Bernard Lietaer called "Community Currencies: A New Tool for the 21st Century", I have come to understand that money can have several different parallell functions (ex: "a store of value" - think of gold and and other precious metals) of which Lietaer states that only two "are essential":
- A standard of measure. We compare the value of the proverbial apples and oranges by expressing each of them in dollars, for example.
- A medium of exchange that is more efficient than other forms--barter, for instance.
The point here is that saying that "since data has a value, it could (perhaps) be thought of as a currency" is just waaay to simplistic. The value of knowing which horse will win can be "invaluable" before the race but worthless afterwards. The same piece of data could simultaneously be immensely valuable for someone and of no value whatsoever for someone else; what is the value of figuring out that a pension fund is trying to offload millions of shares of blue-chip company X exactly one microsecond before anyone else realises it? The answer is "not a lot" - unless you happen to own a small data-center's worth of computing power and be deep into high-frequency trading... and so on... friendship also has value (we all agree friendship is valuable), but who would suggest we use "friendship" as a currency? I agree data has (or at least can have) value, but that is far from saying it can be used as a currency no matter how titillating the idea sounds at face value.
Roy Bendor again contributed with some sharp thinking. His first observation was that data not just has (potential) value but that it (from a sustainability point of view) also has costs; more data means more processing and more processing means more servers and more servers means more data centers and more data centers means higher electricity use and higher electricity use is more unsustainable. Bendor then suggested that we could perhaps fold the material aspects (the sustainability impact) of (using huge amounts of) data into "the price of data itself". That should at least be a lot easier than estimating the value of data since it's much easier to count on what the sustainability costs and the environmental impact of data of running a data center are. Data in more polluting (coal-powered) data centers should thus "cost" more than data in less polluting (water-powered) data centers even if the cost of the energy (say, 1 GWh) is he same in dollars. It's an intriguing though and I can't quite recall if I've heard that particular idea been suggested before.
- Ole Schultz (DK): All the Internet of Things sensors have a materials and energy footprint. Chances are they can have a materials/energy/carbon backpack that is larger than their prospective energy/sustainability savings due to their embodied energy, e.g. the energy expended to build them in the first place. It might not be wrong to build such sensors but a case should first be made that they will have a net-benefit effect.
- Roy Bendon (NL): can we really even use terms like "intelligent" and "smart" if a product isn't sustainable? How can non-sustainable or un-sustainable ever really by intelligent or smart?
- Me: I'm willing to go this far but no longer: these technologies are promising. And they also represent a threat. We can not know in advance how they will be used, but let's work towards, and hope for the best. Just assuming things will work out without any efforts on our behalf will on the other hand only increase the chance that things won't work out.
- Me again, on the organisers' high hopes for "intelligence" enabling the circular economy. Organisers: We will be able to develop products that can tell about how they feel, including if they are about to have a breakdown (which can then be fixed). (Grumpy) me: we might get data and intelligence and some of it might be used to enable the circular economy, but, the same data/intelligence could on the other hand be used to disable the circular economy (like Sofie's printer example above).
- There was no time (and I'm not knowledgeable enough) to initiate a sorely needed discussion about intermediate, appropriate, and convivial technologies as well as perhaps bringing up and discussing the intriguing concept of "minimal computing" which I would also like to look closer at. These ideas were not part of the workshop but I wish they would have.
This is a long blog post and my time and patience is running out. I want to wrap it up and here are some concluding machine-gun fire bullets:
- My paper "Designing for Sustainability: Breakthrough or suboptiminsation?" (pdf file) was one of six Best Paper Nominees. I got to present it from the center stage together with the other five nominees - of which two papers had been written by colleagues of mine at the Center for Sustainable Communications (CESC). Alas, none of our papers won because the winning paper was "Sustainable and Smart: Rethinking What a Smart Home is" by M. Salman, S. Easterbrook, S. Sabie and J. Abate from the University of Toronto.
- I had a cold and also had problems with my neck (tense) and with (at times severe) dry coughs for the duration of the whole conference. That sucked big time. After our workshop had been held (the day before the conference started), I went directly home (to the houseboat, see the picture below) and just crashed. I didn't even have the appetite to eat any dinner. I did go out and have dinner with other conference attendees the very last night but then went home to sleep. I certainly wasn't a party animal this time around.
- The next ICT4S conference will be held at the University of Toronto in mid-May 2018. Steve Easterbrook will be the General Chair and Birgit Penzenstadler will be the Program Chair. I hope to be able to attend and already look forward to that conference. If you've read to the bitter end, you should obviously consider going there too!
- There might be an ICT4S summer school next summer (2017) under the tutelage of Steve Easterbrook and supposedly at University of Toronto(?). I might be involved. More info to follow later (or not).
Also, remember that just because we can do it doesn't mean it's the right thing!
Addenum (Sept 15, 2016). It turns out I misunderstood a few things and Sofie Lambert has also conducted some more "Internet research" on her printer problem:
Interesting blog post, thanks for sharing it, this way I learned a bit more about how the rest of the conference went!
After seeing you refer to my printer story in the post, I retraced what it was exactly that I had to reset that one time to fix my old printer. I should clarify though that I never made any hardware modifications – all I did was reset a software counter (without having to take anything apart). The website that listed the method to fix the “error” on the display never explained why it was there, so I just assumed it was some made up issue that would be hard enough to fix so that it would coerce most customers into buying a new printer instead of googling for half an hour and then doing a manual reset.
But after retracing the issue I now found out that this error signals that the “waste ink tank” which is used to store the waste from cleaning cartridges is almost full:
Resetting the counter could result in overflow of that waste tank at some point (though probably long after the counter indication, safety margin and all that). So in fact, there seems to be a real “issue” that needed to be fixed here. Instead of resetting the counter, in theory I should have taken my printer to a repair center to get the waste tank cleaned. Which at first sight makes this a story of maintenance rather than planned obsolescence.
But on the other hand, I think this may just be a clever cover-up of planned obsolescence, where the story goes: Printer manufacturer X adds a component that needs to be replaced periodically, and which consumers can’t replace themselves (I’ve googled around for a while, so far no luck – and definitely no clear step-by-step guide from printer manufacturer X). Printer manufacturer X knows they have few service centers and that any repair is by default expensive (sending a printer to a repair center and back would already cost around 15 euros I guess). At the same time they make sure that the “periodic” replacement only has to take place when the printer has been used for a long time (outside warranty period), so they know consumers will doubt whether their old device is still worth the repair cost at the point when it refuses to print due to this “error”.
So basically, I’d still call this planned obsolescence, but printer manufacturer X has been clever enough to make it impossible to prove malicious intent… I have a feeling there are many examples out there of other brands and manufacturers using components that fail for “valid” reasons after using them for a while. As long as expectations of infinitely growing economies and profits push companies to increase their sales year after year I don’t see any reason this would change, either.
Long story short, I guess I’m with you on the “smarter is not necessarily more sustainable” front. As usual, ICT is a tool that can be used either way, depending on the user’s (or better, companies’) intentions.