When the Implication Is Not to Design - Discussion

34 views
Skip to first unread message

Eric P. S. Baumer

unread,
Jun 29, 2011, 11:48:35 AM6/29/11
to sustain...@googlegroups.com, six
Last month at CHI, there was a paper by myself and Six Silberman titled When the
Implication Is Not to Design (Technology). The basic premise is that there are
some situations where a technological intervention may not be the most
appropriate. The paper provides specific ways of articulating when this may be
the case, as well as practical recommendations for applying this perspective.
Copies are available at http://doi.acm.org/10.1145/1978942.1979275 or
http://ericbaumer.com/publications/impl9-rev.pdf.

We would like to take this opportunity to solicit comments and critiques. This
paper was intended first and foremost to be part of a conversation, and we
believe that some of that conversation should happen here on the sustainable-chi
list. We hope that members of the general HCI community will join us in this
discussion.

~Eric and Six

Eric P. S. Baumer

unread,
Jun 29, 2011, 12:36:58 PM6/29/11
to sustain...@googlegroups.com, six
Apologies for a minor typo. The second link to the paper should be
http://ericbaumer.com/publications/impl10-rev.pdf.
~Eric

Bill Tomlinson

unread,
Jun 29, 2011, 6:55:09 PM6/29/11
to sustain...@googlegroups.com, Eric Baumer, six Silberman
Eric and Six,

Thank you for initiating this conversation. I agree that there are many circumstances where new technology is not the solution, and where the framing of "problems" and "solutions" isn't the most productive way forward.

I'd be interested to hear more of your thoughts on the framing of "design as an intervention in a complex situation." If metrics "cannot
fully capture the complexities of environmental sustainability" (which I agree with), how might one get a sense for whether an intervention is making the situation more or less sustainable? You point to some examples of simplifications proving invalid, but there are many simplifications that do provide viable means of understanding the world of which we are a part.

The topic of indirect effects, while you allude to in your paper, is particularly relevant to this discussion. Six and I, along with Jim White, recently published a column in IEEE Computer about how indirect effects call into question the often-assumed correlation between efficiency and sustainability in IT systems.
http://wtf.tw/text/more_efficiently_unsustainable.pdf

Understanding these indirect effects, at least to some degree, is critical to assessing the value of an intervention. And, if we do not have the cognitive apparatus to comprehend these indirect effects to a sufficient degree of accuracy (and I argued in my book Greening through IT that we do not), then this may well be a place for IT to be deployed effectively. These IT interventions, however, are likely to have applications that run strongly counter to sustainability as well, so it ultimately returns to being a question about the human aspects of these situations. This matches the point Toyama makes about ICT4D in his article in The Atlantic - "Our successes were due more to effective partners, and less to our technology."
http://www.theatlantic.com/technology/archive/2011/03/technology-is-not-the-answer/73065/
(Disclaimer: I first learned about this article from Six, so it's perhaps unsurprising that it matches with this paper. :) )

There are also some interesting implications of this paper for the boundaries between academic disciplines. Right now, it seems like a better strategy for success in academia to have depth in one area, rather than breadth across many. A person with 20 papers in the top conference or journal in a field will likely have a whole community at his/her back when it comes time for tenure, ready to write letters that say "wow, fantastic scholar, yes, tenure." A person with one paper in equivalent venues in 20 different fields, though, won't have the same coherent community to speak on his/her behalf. You quote, in your paper, that “It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail”. I might argue that, if you're in the Department of Hammers, that's not necessarily a bad strategy for personal success. So then perhaps we need to think about how to support holistic approaches in academia (if, in fact, academia is where we expect prosocial interventions in complex situations to arise). (But this, I know, gets into other thorny terrain that is often debated at CHI and elsewhere.) Thoughts on these topics would be welcome as well.

Best,
Bill

P.S. The idea of "technological extravention" reminds me of Amory Lovins' (2003) concept of "negatechnologies". He offered that "we tend to discuss only deploying better new technologies. Equally important is getting bad existing technologies out of use.”

> --
> You received this message because you are subscribed to the Google Groups "Sustainable CHI" group.
> To post to this group, send email to sustain...@googlegroups.com.
> To unsubscribe from this group, send email to sustainable-c...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/sustainable-chi?hl=en.

---
Bill Tomlinson
Associate Professor of Informatics
University of California, Irvine
http://www.ics.uci.edu/~wmt
w...@uci.edu


Eric P. S. Baumer

unread,
Jul 27, 2011, 1:33:46 PM7/27/11
to Bill Tomlinson, sustain...@googlegroups.com, six Silberman
Bill,

Thanks for your thorough and thoughtful comments. I want to take a moment to
comment on the question of metrics.

Some of the points you make were also raised during a question after the paper
presentation at CHI. We are not trying to say that metrics are completely
useless. However, any metric (for sustainability or most anything else)
necessarily abstracts away from certain details of the situation to focus on
others. The point is to attend to what's lost in that process of abstraction. As
the citation in the paper to Seeing Like a State suggests, this is not an
entirely new argument, but it seems highly applicable in the context of
sustainable HCI.

I'm also intrigued by your question of how to determine "whether an intervention
is making the situation more or less sustainable." The notion that a situation
can be more or less sustainable perhaps implies, or is predicated on, the
assumption that sustainability is something that can in fact be quantified as
such (hence in part the interest in metrics). This draws attention to two
important points.

First, what is sustainable? Or, perhaps, what is sustainability? As Six has
previously argued, sustainable HCI on the whole seems to be doing a significant
amount work to help, or to make, users be more sustainable without having in
depth discussions about what it actually means to be sustainable, or about what
exactly is to be sustained. Perhaps a conversation about metrics for
sustainability would help foreground this gap, but I'm still not convinced that
better metrics are what we need.

This brings me to my second point. We should not be focused solely on metrics
for sustainability because sustainability is not solely a technical or a
scientific category that needs measuring. It is also a social and a cultural
category. Woodruff et al.'s CHI 2008 paper provides some illustrative examples
of how sustainability as a cultural category is made manifest. There was also
some intriguing work discussed at the workshop on Everyday Practice and
Sustainable HCI this year at CHI that does a good job of foregrounding the
sociocultural dimensions of sustainability, and how people create and enact
sustainability as a cultural category. An emphasis on metrics and simplification
draws attention away from these complex social and cultural dynamics. Attending
to such social and cultural dynamics may provide an alternative,
non-metric-based approach to thinking about the ways in which a particular
situation may be more or less (un)sustainable.

~Eric

six silberman

unread,
Jul 31, 2011, 2:17:00 AM7/31/11
to Eric P. S. Baumer, Bill Tomlinson, sustain...@googlegroups.com

> what is sustainability?

I'll take the bait. I've been sitting on drafts of CHI papers about this literally for years, and I doubt I'll ever submit them, so I might as well post excerpts here.

I've tried to make my terms as precise as reasonably possible, but there's only so much precision you can get out of language. At some point you have to admit that it is, after all, turtles all the way down. I hope it's at least fun to chew on.

Four axioms for an interpretation of sustainability relevant to sustainable HCI research:

1. Sustainability is a property of a process in a context.

2. An unsustainable process is one whose effects preclude its preconditions. [1]

3. No process is indefinitely sustainable.

4. Sustainability is a means, not an end.

Explanations, examples, and implications:

Sustainability is a property of a process in a context

Explanation. The adjectives "sustainable" and "unsustainable" can be used to describe things like processes, systems, or practices. Things which cannot be described meaningfully as processes, systems, or practices cannot be described meaningfully as (un)sustainable. The (un)sustainability of a process cannot be inferred from knowledge of the process alone. To determine if a process is (un)sustainable, we must consider the process in its context.

"X is a property of Y" here means "X is a variable associated with things of type Y." The values x1, x2, etc. for the variable X associated with particular things y1, y2, etc. of type Y can be used to distinguish between those things.

"Process" here denotes any series of events which recurs. Being a series of events, a process has a first event and a last event, and the ordering of events is a property of the process. Being a series of events, each with preconditions and effects (which themselves may be preconditions of other events [2], and which vary depending on the context in which they occur), a process has preconditions and effects.

"Context" is a difficult term in the analysis of designed systems and the interactions they occasion. Dourish [3] writes that in the phenomenological view context, rather than being representable information, is a relational property which arises from activity. Whether something is 'in the context' of an activity at a particular time is not fixed but depends on the details of the activity. Activity and context are not distinct but mutually constitutive.

This nondualist conception of activity and context recalls a similar difficulty in distinguishing between "system" and "environment" in systems analysis. While 'objectively' (i.e., to a truly uninterested observer) there is no distinction between process and context or system and environment, in any particular analysis there always is, even if only implicitly. That is, the distinction is not a property of the world but of our way of seeing it. [4] "System" denotes that which is analyzed and represented in detail, while "environment" denotes that which is represented only schematically or selectively. Similarly, "process" denotes that which may be reconfigured by design intervention, "context" that which is assumed fixed. The boundary cannot be inferred from first principles; it is drawn according to the details of a particular analysis. The boundary between context and 'non-context' — that which lies beyond context and is irrelevant to the present analysis — is similarly arbitrary. But like the process/context boundary, it must be drawn — it is drawn, if only implicitly — for the analysis to proceed.

Implication. There is no such thing as a sustainable or unsustainable object.

If only processes or process-like things can be (un)sustainable, then any claim that an object is inherently (un)sustainable is meaningless. Such a claim is either a category error, if made seriously, or a technological determinism, if shorthand for a claim that the object will induce more or less sustainable practices regardless of user or context.

For example, "suburbia is unsustainable" and "this phone is more sustainable than that phone" are meaningless propositions, even if "suburbia" is well-defined. They are like the proposition "this rock is sustainable" in that suburbia, phones, and rocks are objects, not processes. They have neither preconditions nor effects; they are therefore neither sustainable nor unsustainable. Suburban driving, however, is a process, with preconditions and effects. It can therefore be evaluated for sustainability: the statement "suburban driving is unsustainable" is either true or false rather than meaningless. Whether it is true or false depends on our definitions and the scope (especially in time) of our particular analysis.

Implication. To plan or evaluate an intervention to make a process for sustainable, the whole process must be examined in its context.

An unsustainable process is one whose effects preclude its preconditions

Implication. If a process is unsustainable, it will stop.

For example, a rubber ball bouncing on a table will stop bouncing because with each bounce energy is lost (to the "context" or environment) as heat and sound. The ball's bouncing (a process) is unsustainable: its effects (dissipation of energy) preclude its preconditions (height above the table).

No process is indefinitely sustainable

For example, Earth's orbit is unsustainable: in a few billion years the sun will expand, destabilizing Earth's orbit and eventually engulfing it.

Implication. To say that a process is sustainable is meaningless without an indication of for how long.

Implication. To say that one process is 'more sustainable' than another can mean that it is sustainable for a longer time.

Sustainability is a means, not an end.

Implication. Unsustainability alone cannot justify design intervention into a process.

If a process whose effects are judged 'bad' is unsustainable, perhaps no intervention to make it more sustainable is called for. If a process with effects judged 'good' is unsustainable over a relevant period of time, intervention may be warranted. But as no rigorous moral argument can be constructed without recourse to arbitrary values, justification of such intervention always involves subjective judgment or preference.



[1] This articulation of this old idea is adapted from Eric Zencey's essay 'Ecology and guilt', collected in Virgin Forest (Univ. of Georgia Press, 2000).

[2] D. Adams. Mostly Harmless (Del Rey, 2000).

[3] P. Dourish. What we talk about when we talk about context. Pers. and Ubiq. Comp. 8(1): 19-30, 2004.

[4] Put another way: if we were truly uninterested, we would not bother to make the analysis at all. It is our interests that allow us to decide what to call "system" and what "environment" — that is, to decide what is 'of interest'. It is our interests that allow us to perform the analysis at all.

Marcela Musgrove

unread,
Jul 31, 2011, 2:31:34 AM7/31/11
to sustain...@googlegroups.com
A good read to support or supplement your argument is "Design is the
Problem: The Future of Design Must be Sustainable by Nathan Shedroff

Some notes I took from book:
Xxiii
"A sad truth is that almost every solution designed today, even the
most 'sustainable' one, has more of a negative impact on the planet
than a positive one. This means that the world would be better off if
most of what was designed was never produced. This is changing and it
doesn’t have to be the case in the future, but we have a long way to
go in order to change this pattern."

Carl DiSalvo

unread,
Jul 31, 2011, 4:05:23 PM7/31/11
to sustain...@googlegroups.com, Eric P. S. Baumer, Bill Tomlinson
Thanks all for initiating this important conversation.

Certainly, we should discuss and debate what constitutes sustainability. The discussion below by Six is exciting because it offers propositions we might use as the basis of discussion and debate (for example, in regards to whether or not objects have effects, I would not quickly agree to the claim that only processes have effects, but it's important to assert and explore such claims and to parse out ideas and their implications within sustainable HCI). In addition, we should also consider the boundaries of HCI as a field of research and practice. 

Many of the good points brought up in the paper "When the Implication is Not to Design (Technology)" are specific to HCI. I am not sure if they apply in the same way if the parenthetical (Technology) is removed. For me, this is part of the crux of the problem -  it seems much of the issue is with the attachment to technology that defines HCI. As Bill points out - this has professional/disciplinary significance (and I empathize with this point). For example, if we were addressing sustainability from within a broader design or design research community then technology wouldn't really be an issue - it would be fine to turn away from technology as an assumed solution or aspect to a solution. 

Within sustainable HCI research, perhaps the most important condition to work for is a condition of pluralism. Pluralism allows for both the operational metrics and also for more critical examination of sustainability as a social and cultural category. The operational metrics may indeed be important in the construction of systems (I'm thinking here of how AutoCad tries to support sustainable design through their software, I *assume* this is based in some part upon metrics within architecture). But we also need more than metrics, we need social / cultural / political analysis - not necessarily to support practice, but as strand of research within HCI broadly. 

Carl

Carl DiSalvo
Assistant Professor
School of Literature, Communication, and Culture
Georgia Institute of Technology

Reply all
Reply to author
Forward
0 new messages