=Paper=
{{Paper
|id=Vol-1679/paper7
|storemode=property
|title=DiRec: A Distributed User Interface Video Recommender
|pdfUrl=https://ceur-ws.org/Vol-1679/paper7.pdf
|volume=Vol-1679
|authors=Wessam Abdrabo,Wolfgang Wörndl
|dblpUrl=https://dblp.org/rec/conf/recsys/AbdraboW16
}}
==DiRec: A Distributed User Interface Video Recommender==
https://ceur-ws.org/Vol-1679/paper7.pdf
DiRec: A Distributed User Interface Video Recommender
Wessam Abdrabo Wolfgang Wörndl
Technical University of Munich Technical University of Munich
Boltzmannstrasse 3 Boltzmannstrasse 3
85748 Garching Bei München, Germany 85748 Garching Bei München, Germany
wessam.abdrabo@in.tum.de woerndl@in.tum.de
ABSTRACT platforms. A typical situation is a user carrying out tasks
Distributed User Interfaces (DUIs) are graphical interfaces in a multi-device environment that presents itself effectively
whose components are distributed in one or many of the UI to the user as a single UI, but which is actually distributed
distribution dimensions: Time, space, platforms, displays, or along these platforms. Such situations represent typical cases
users. In this work, we have investigated the impact of the of Distributed User Interfaces (DUIs). Hence, DUIs represent
application of DUIs, with respect to the different DUI dimen- an attempt to overcome the limitations of user interfaces
sions, on the experience of users of recommender systems. that are manipulated by a single user, on a single platform,
We developed two prototype video recommendation mobile in a fixed environment, providing few or no variations along
applications: Monolithic Interface Recommender (MiRec), these distribution dimensions.
and Distributed Interface Recommender (DiRec). Sharing To our best knowledge, surveyed studies for the applications
mostly the same interface, DiRec additionally offers the pos- of DUIs do not include any which tackle single-user recom-
sibility of migrating parts of the UI between the mobile mender systems; the fact that provided the main motivation
application and a larger display (LD). A user study was con- for this research. We hypothesize that the distribution of
ducted in which participants used and evaluated both MiRec recommender systems’ UIs leads to an enhanced user ex-
and DiRec. Our results show a significant difference between perience. To verify our hypothesis, we developed two high
DiRec and MiRec in attractiveness (general impression and fidelity prototypes for video recommendation: Monolithic
likability), stimulation, and novelty measures, which posits Interface Recommender (MiRec), which is a conventional
the existence of a strong interest in DUI recommender sys- mobile video recommendation application, and Distributed
tems. Nonetheless, MiRec was found more easy-to-learn and Interface Recommender (DiRec), which is a distributed ver-
easier to understand than DiRec which gives room for further sion of the mobile video recommender where the interface is
investigation to pinpoint the reasons of DiRec’s relatively distributed among a mobile device (SD) and a large-display
lower perspicuity measures. screen (LD).
The proceeding sections describe this research’s main contri-
butions: A proposal for a generic model for UI distribution
CCS Concepts for recommendation applications, the design of DiRec which
•Human-centered computing → User interface de- is considered as an instance of this generic model, as well as
sign; the results and conclusion of a user study that was conducted
to test the impact of our DUI recommender’s design on users’
Keywords experience.
Distributed User Interfaces; Recommender Systems; Mi-
gratable Interfaces; Mobility; User Study. 2. BACKGROUND AND RELATED WORK
Enhancing the experience of users of recommender systems
1. INTRODUCTION through developing more sophisticated recommendation al-
With the advancement of ubiquitous computing and the gorithms, taking in consideration aspects such as the novelty,
trend of the ever-increasing number of devices per user, users diversity, and accuracy of recommendations, has become
of interactive systems no longer perform tasks that reside the focus of many recent studies. However, fewer studies
mainly on a single device, but are rather confronted with investigate the possibility of enhancing the user’s experi-
situations where they need to complete tasks across several ence through providing novel UI solutions for recommenders.
None of the surveyed research has considered the impact of
the distribution of the UI of recommenders on the user’s
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are not experience. This is where our study provides its main contri-
made or distributed for profit or commercial advantage and that copies bear bution.
this notice and the full citation on the first page. Copyrights for components During the course of our investigation, we surveyed many
of this work owned by others than ACM must be honored. Abstracting with studies that laid the foundation of the relatively new field
credit is permitted. To copy otherwise, or republish, to post on servers or to of DUIs. Mostly relevant to our study is Vanderdonckt et
redistribute to lists, requires prior specific permission and/or a fee. Request al. [9] ’s description of what constitutes a distributed UI
permissions from permissions@acm.org.
IntRS 2016, September 16, 2016, Boston, MA, USA. environment: “UI distribution concerns the repartition of
Copyright remains with the authors and/or original copyright holders. one or many elements from one or many user interfaces in
Figure 1: Recommended video consumption and rating as an instance of the generic DUI model.
order to support one or many users to carry out one or many 3. DESIGN OF A DUI SINGLE-USER REC-
tasks on one or many domains in one or many contexts of OMMENDER
use, each context of use consisting of users, platforms, and
environments.” To deepen our understanding of the various Scenarios of our DUI video recommender depict a multi-
dimensions of UI distribution, we surveyed several studies ([2], device environment, in which the flow of control (logic) and
[3], [5], [9]). However, one that has been especially relevant the application’s user interface are decoupled in a way that
to our study is the 4C model described by Demeure et al., allows for the distribution of UI components along the dif-
through which we could define the 4Cs of our proposed DUI ferent devices. In other words, the user of such a system is
recommender: Computation (what is distributed?), in other provided with a distributed solution, which enables him/her
to perform tasks on whichever device in this environment
words the element of distribution, which could be the task
or the platform, Communication (when is it distributed?) or (by for example migrating the UI components between the
time, Coordination (who initiates distribution?) which is a different devices) independently of where the application is
variation on the user dimension, and Configuration (from running, and of the constraints presented by the different
platforms running the application.
where and to where is the distribution operated? on the
physical pixel level, or the logical level) [2].
On the other hand, a number of studies have found DUI
techniques useful for their applications among which are IAM
3.1 Generic Model for UI Distribution
[1], Aura [7] and ConnecTables [8]. The following are generic scenarios for UI distribution
For implementation of our DUI recommender, we adopt a of interactive systems that are applicable to recommender
dual display (SD-LD) approach which is similar to Kaviani systems:
et al’s, who argue that the use of ubiquitous cell phones as
• Migration of Item Consumption: present the recom-
an SD component in a DUI not only offer a means to interact
mended content on one device while giving the user the
with LD displays, but increasingly offer a small, but high
ability to consume the content on another device.
quality screen to complement the LD [4].
Moreover, in our previous work [10], we investigated the • Performing Parallel Activities: user can perform tasks
application of DUIs in group recommender systems. We simultaneously and independently from each other.
developed a scenario of a movie recommender, where the
UI is distributed among two platforms: a PDA that works • Overview and Detail Presentations: show different ver-
as a small display (SD) and a table-top that works as a sions of the presented content at different levels of
large display (LD). Users get to view and rate recommended granularity on different nodes.
items on their PDAs individually, and as a group, they get to
reach a consensus by doing the voting on the table-top. This • Content Filtering: distribute the task to filter the user’s
DUI solution to the voting part of group recommendation choice of what to consume.
is proved by the study to improve the process of reaching
consensus among a group. This study takes a further step • Content Redirection: content could be transferred to
by investigating the benefits of using DUIs in single-user be presented on a different node.
recommender systems.
• Migration of Items Between Users: content redirec-
tion/migration of a list of recommended items (or an
item in this list) from one user of the system to one or
more other users.
We will describe more specific scenarios that can be consid- (SD), the user performs a pan gesture on the video image,
ered as an extension of this generic UI distribution model which then triggers the migration of the video consumption
(Figure 1) in a distributed video recommender application in from the mobile device to the LD.
the next subsection. The video player automatically starts on the LD, providing
the user with all controls for the video playback. After
3.2 DiRec: Distributed Interface Video Rec- the video playback starts automatically on the LD, the LD
ommender triggers the mobile device to display the rating page for the
We assume the users are working with a smaller (SD), e.g. user on the SD. Hence, the two tasks could be carried out
a smartphone or other mobile device, and a larger display simultaneously by the user (Figure 3).
(LD), e.g. a display screen.
3.2.5 Filtering Recommended Items
3.2.1 Pre-Configuring UI Distribution Options Filtering is done by performing a right swipe gesture on the
This scenario presents the initiation point of the system, in video item in the list on the SD which redirects the content
which the user is given an option to pre-configure the different of the video to the LD. The display of the content on the LD
options the system offers for UI distribution, and hence be is also done in an overview-detail coupling manner. After
the initiator of UI distribution. This offers the ability to the user is done filtering the LD will contain all the selected
delay the decision of which UI components to present on items displayed as an overview.
which platform, making the system distributed in time. This
is made possible by presenting the user with a Meta UI in 3.2.6 Redirecting Favorites Lists
which he/she is asked to drag and drop the components of Unlike previously described scenarios which involve a single
their choice to the target platform. user of the system, this scenario involves two or more users.
On the SD, the user selects a favorite-items list. On applying
a long-press on the list, the user is prompted with a list of
users from which he could select one or more users to transfer
this list to.
Figure 2: Redirecting recommended item consumption from
SD to LD.
3.2.2 Presentation of Recommendation Results Figure 3: Rating a recommended video on SD in parallel to
The presentation of recommended videos is shown in par- watching it on LD.
allel on the SD and LD, however, in different levels of gran-
ularity. The mobile device shows a detailed list of all the
recommended videos, together with detailed information 3.3 Prototype Implementation
about the video, in tabular form with different categoriza- A subset of the suggested distribution scenarios was se-
tions. On the LD, an overview presentation is shown for lected for implementation. MiRec is developed as the non-
the recommended items that scored the highest for the user distributed version of DiRec and is meant for comparison
without details, however shown in different sizes to indicate with DiRec’s interface through our comparative user study.
the recommendation scores. Both applications share mostly the same design, however,
thorough DiRec, the user could complete tasks in a dis-
3.2.3 Recommended Item Details Presentation tributed manner between a mobile application and a large
Moreover, in our proposed design, we offer the possibility display screen, while with MiRec, users could only complete
of distributing parts of the UI with a fine granularity. The tasks on the mobile device. MiRec is developed as an iOS
user selects a single table-cell in the videos list and could mobile application while DiRec is distributed along an iOS
move it to the LD by applying the gesture, as opposed to just application and an LD Python application with a communi-
mirroring or transferring the UI at a more coarse granularity. cation layer in between which mainly relies on light-weight
TCP-IP based message passing between both platforms (e.g.:
3.2.4 Recommended Item Consumption and Rating play: is passed from SD to LD in DiRec to play a
Starting a video on the LD is done as depicted in Figure 2 in video on LD).
our prototype. On the video details page on the mobile device
4. USER STUDY 5. CONCLUSIONS AND FUTURE WORK
To evaluate our approach, we have conducted a user study This work investigates the impact of using distributed user
in three phases. 24 participants were asked to use both interfaces on the experience of users of recommendation appli-
MiRec and DiRec and rate their experiences of the products cations. Our comparative user study’s UEQ results could be
using the User Experience Questionnaire (UEQ) method [6] interpreted as follows: The use of DUIs aids the stimulation
shortly after finishing the test. and novelty of recommendation applications, hence, enriches
the user’s experience, does not hinder the efficiency or limit
the span of the user’s control of recommendation applica-
tions, results in more attractive recommendation applications,
however, might affect the learnability and ease-of-use of rec-
ommendation applications. Notwithstanding the promising
results of our study, the study has fallen short in providing
an explanation of whether the relatively lower perspicuity
measures of DiRec is a result of insufficient explanation of
the study’s procedure, or if it was DiRec’s design that was
relatively less easy to understand and learn. A possible fu-
ture work would be to further investigate this aspect. Lastly,
we strongly believe that giving more span of control to the
Figure 4: Participant’s interaction with DiRec. user through allowing pre-configuration of UI distribution
schemes could further enhance the DUI experience.
4.1 Setup 6. REFERENCES
Each participant was first briefed about how to use MiRec [1] J. Coutaz, L. Balme, C. Lachenal, and N. Barralon.
and DiRec, then he/she was asked to complete a set of tasks Software infrastructure for distributed migratable user
on both applications including navigating recommendations’ interfaces. In Proc. of UbiHCISys Workshop on
lists, playing and rating of videos. Each participant was given UbiComp, volume 2003. Citeseer, 2003.
an iPhone with both DiRec and MiRec installed and was [2] A. Demeure, J.-S. Sottet, G. Calvary, J. Coutaz,
being asked to interact with the LD screen component during V. Ganneau, and J. Vanderdonckt. The 4c reference
the course of the experiment (Figure 4). During the last model for distributed user interfaces. In Autonomic and
phase of the experiment, participants were asked to give their Autonomous Systems, 2008. ICAS 2008. Fourth
direct impression of the application using the UEQ method International Conference on, pages 61–69. IEEE, 2008.
[6]. UEQ consists of 6 scales with 26 items which measure [3] N. Elmqvist. Distributed user interfaces: State of the
Attractiveness (overall impression or the likability), Perspicu- art. In Distributed User Interfaces, pages 1–12.
ity (learnability and ease-of-use), Efficiency (the ability to Springer, 2011.
perform tasks without exerting extra effort), Dependability
[4] N. Kaviani, M. Finke, R. Lea, and S. Fels. Dual
(user’s control over the experience), Stimulation (excitement
displays: towards an interaction model and associated
and motivation) and Novelty (innovation and creativity).
design guidelines. DUI 2011, page 69, 2011.
4.2 Results [5] J. Melchior. Distributed user interfaces in space and
time. In Proceedings of the 3rd ACM SIGCHI
Figure 5 shows the result of UEQ’s comparison of MiRec
symposium on Engineering interactive computing
(left side, blue) and DiRec (right side, red). With respect to
systems, pages 311–314. ACM, 2011.
attractiveness, stimulation, and novelty, DiRec scores higher
than MiRec. For efficiency and dependability, they measure [6] M. Schrepp. User experience questionnaire handbook.
almost similarly with MiRec scoring slightly better than ueq-online.org, 2015.
DiRec. MiRec, however, scores much higher than DiRec when [7] J. P. Sousa and D. Garlan. Aura: an architectural
it comes to the perspicuity scale. Conducted t-Tests showed framework for user mobility in ubiquitous computing
statistical significance with regard to perspicuity (α =0.0092), environments. In Software Architecture, pages 29–43.
stimulation (α=0.0007), and novelty (α =0.0000), but no Springer, 2002.
significance for attractiveness, efficiency and dependability [8] P. Tandler, T. Prante, C. Müller-Tomfelde, N. Streitz,
with an alpha level of 0.05. and R. Steinmetz. Connectables: dynamic coupling of
displays for the flexible creation of shared workspaces.
3.00 In Proceedings of the 14th annual ACM symposium on
2.25 User interface software and technology, pages 11–20.
1.50
ACM, 2001.
0.75
[9] J. Vanderdonckt et al. Distributed user interfaces: how
0.00
to distribute user interface elements across users,
-0.75
platforms, and environments. Proc. of XI Interacción,
20, 2010.
-1.50
Attractiveness Perspicuity Efficiency Dependability Stimulation Novelty [10] W. Wörndl and P. Saelim. Voting operations for a
group recommender system in a distributed user
Figure 5: Comparison of scale means in MiRec (left/blue) interface environment. In RecSys Posters. Citeseer,
and DiRec (right/red) 2014.