[dcchairs2016] Review of UbiComp/ISWC 2016 Doctoral School submission 111

[dcchairs2016 at ubicomp.org]

Paper  111 - Robotic Displays for Eco-feedback
Reviewer 2 - Max Muhlhauser

Overall rating:  3  (scale is 1..6; 6 is best)

Confidence

   Very confident - I am knowledgeable in the area 

Contribution to UbiComp

   marginal (robotic displays are taken as a given, not
   designed or put in a "closed" user-device-task loop as common for
   UbiComp research
 

Overall Rating

   3  (Maybe reject: I would agree with rejecting this paper.)

R&R Suitability (hidden from author)

   It would be possible to improve within 5 weeks, but difficult	 

The Review

   The contribution is entitled "Robotic Displays for Eco-feedback” and,
   in the beginning, claims to address the question if robotic displays are
   better than ambient displays for making building occupants adapt
   (improve) energy saving behavior. 
   Three major issues make the paper appear to be below the threshold of
   acceptance for the UbiComp DC:
   1.	Lacking consistency and lack of coherent PhD plan
   2.	Doubtful claims and conclusions
   3.	Lacking pertinence for Ubicomp
   This critique will be further detailed below.

   1.	Lacking consistency and lack of coherent PhD plan: in the beginning,
   quite some space is lost with a verbose description of the need for
   energy saving etc. – the space lost seems to be lacking when it comes
   to providing the information requested in the call for papers (PhD plan,
   scientific methods, related work etc.); 
   lacking consistency further relates to: (a) "Focus-hopping" (e.g., for a
   large portion, eco feedback is not the issue addressed) (b) inadequate
   focus of related work (most of the related work is used to [pseudo-] back
   all sorts of *non-coherent* research questions, but not for showing the
   state of the art that the authors plans to exceed: what was achieved in
   the past that shall be outperformed?
   2.	Doubtful claims and conclusions: since many claims and issues are not
   related to concrete *effective* eco-feedback, they have little value; to
   cite two examples: 
   a.	Robotic displays are claimed to achieve better attention than ambient
   displays; this is ridiculously banal since ambient displays are
   deliberately meant to NOT raise much first-level attention; a much more
   valid question would be if (to put it a bit bluntly) the performance AND
   acceptance of a noisy, obtrusive, over time nerve-racking “waiving and
   gesticulating” robot leads to better *long-term* eco-behavior of
   building occupants than a non-obtrusive, gentle yet subconsciously active
   ambient display.
   b.	The “trust” associated with robots (in comparison to non-moving
   outputs) is hypothesized several times as a key (for eco-feedback and
   much more); all the literature cited for backing this point is centered
   around short-term interaction (such as following the robot in a rescue
   situation after probably a couple of minutes of acquaintance with it);
   what is worse, the hypothesis was also tested by the author with a rather
   short experiment (finding the center of a maze); however, there is vast
   evidence reported in the literature that humans trust mobile robots more
   than normal Computers because they tend to anthropomorphize them much
   more – often leading to way exceeding expectations that will sooner or
   later be disappointed, which again leads to dramatic loss of trust; of
   course, this process can be mitigated, but then the proposed research
   becomes an aweful lot more difficult, and “trust” is no more a simple
   anchor for major claims
   3.	Lacking pertinence for UbiComp: while the subject of robot displays is
   well slanted towards UbiComp research, the contribution evaluated here
   does not look at an ensemble of robot display, task-at-hand, and user
   with the aim to design and design and evaluate an eco-feedback system
   proper; rather, the research “picks” particular empirical issues
   w.r.t. how *users* interact with robotic displays; while UI guidelines
   that may emerge from such research may inform a UbiComp project, they are
   not at the very core of UbiComp; given the quite competitive
   submission:acceptance ratio possible, this third aspect must also be
   taken into account.

   Overall, the paper will probably have to be rejected since Ubicomp
   pertinence, clear Overall PhD plan with all major elements sufficiently
   explaned, and quality of the research altogether are required to be
   excellent in order to cross the high threshold of acceptance.
 

Confidential Comments (Optional) (hidden from author)

 


To see the review, go to https://precisionconference.com/~ubicomp?goto=ubicomp16c