SMARCOS CNR Paper Workshop Distributed UI
0 likes273 views
This document discusses an approach to specifying distributed user interfaces using the MARIA language. The approach allows designers to specify distribution at various levels of granularity, from individual interface elements to entire presentations. It uses CARE properties - complementarity, assignment, redundancy, equivalence - to indicate how elements are distributed across devices. Examples show how interactors can be distributed according to CARE properties both within and across devices to create flexible, distributed user interfaces. The goal is to support multi-device environments and allow generation of implementations for different platforms from a single specification.
![Distributing User Interfaces with MARIA
Marco Manca, Fabio Paternò
CNR-ISTI, HIIS Laboratory
Via Moruzzi 1, 56124 Pisa, Italy
fabio.paterno@isti.cnr.it
ABSTRACT since each device has specific interaction resources and
The increasing interest for distributed user interfaces needs implementation languages to execute such user interfaces.
languages that better support their specification in multi-
device environments. In this paper we discuss an approach An interesting approach has been proposed [6] through the
to specifying this type of user interfaces, and how it is use of a XML-based language called HLUID (High Level
applied to an extension of the MARIA language aiming to UI Description), in which the user interface has a
provide flexible support for the description of distributed hierarchical structure and the leaves of the tree are Abstract
user interfaces. Interactor Object (AIOs) describing high-level interactors.
During the rendering process the AIOs are mapped onto
Author Keywords Concrete Interaction Object (CIOs) associated with the
Distributed User Interfaces, Model-based User Interface current platform. In addition, they introduce a split concept
Design, User Interface Description Languages. for the groupings through an attribute that, when it is set to
true, allows the distribution of the user interface elements
ACM Classification Keywords without losing its logical structure.
H5.m. Information interfaces and presentation (e.g., HCI): In our case we propose a different solution, still using a
Miscellaneous. model-based approach. One difference is that we support
the specification at the concrete level because at this level it
General Terms is easier to generate the corresponding implementations and
Design there is a better understanding of the actual effects that can
be obtained.
INTRODUCTION
One of the main technological trends is the steadily In the next section we describe the design concepts that
increasing number of devices per person. This has an characterise the proposed solution for specifying user
impact on the user interface languages and technologies interfaces distributed across various devices with MARIA.
because it will be more and more common to interact with Our approach is not limited to graphical user interfaces but
an application through multiple devices. A typical can address also devices supporting other modalities, such
applications scenario is using a graphical or a vocal mobile as voice. Then, we also provide some application examples.
device to interact with a large screen Lastly, some conclusions are drawn along with indications
for future work.
This trend has stimulated increasing interest in Distributed
User Interfaces (DUIs), which have been defined [2] as THE PROPOSED SOLUTION
those interfaces whose different parts can be distributed in The proposed solution has been introduced for the MARIA
time and space on different monitors, screens, and language [5], which provides both an abstract and a set of
computing platform, depending on several parameters concrete languages. The concrete languages refer to
expressing the context of use, such as the user, the different platforms, characterised by their interaction
computing platform, and the physical environment in which modalities (graphical desktop, graphical mobile with touch,
the user is carrying out her interactive task. vocal, graphical and vocal combination). Thus, a distributed
Model-based approaches [3] have often been considered in concrete user interface should be able to indicate how the
order to manage the increasing complexity derived from user interface elements are distributed among devices,
managing user interfaces in multi-device environments, which can belong to such various platforms.
One important point to consider is at what level of
granularity the distribution can occur. We have identified
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
four possible levels:
not made or distributed for profit or commercial advantage and that copies
Presentation, which means the set of elements that
bear this notice and the full citation on the first page. To copy otherwise,
or republish, to post on servers or to redistribute to lists, requires prior characterise the possible perceivable information
specific permission and/or a fee. at a given time (e.g. a page in a Web application);
CHI 2011, May 7–12, 2011, Vancouver, BC, Canada.
Copyright 2011 ACM 978-1-4503-0267-8/11/05....$10.00.](https://image.slidesharecdn.com/819cnrchi2011paperworkshopdistributedui-120119170449-phpapp01/85/SMARCOS-CNR-Paper-Workshop-Distributed-UI-1-320.jpg)
![ Composition of elements, groups of elements In the application of the CARE properties, we consider a
inside a presentation that are logically related user interface specification as a hierarchical structure: the
among them; user interface is composed of presentations, presentations
contain single interactor elements, and compositions of
Interactor element, single user interface elements; such elements. In particular MARIA has various operators
Interaction phase, we have identified three general for composition of interactors:
phases that characterise interactions: prompt, The different types of interactor compositions are:
input, and feedback, and it is possible to distribute
the user interface even across such sub- Grouping: a generic group of interactor elements.
components of one interactor. For example,
Relation: when two or more interactors or
entering a text through a mobile device and
compositions are related to each other.
receiving feedback both on the mobile and a large
screen device. Composite Description: represents a group aimed
to present contents through a mixture of
The next aspects that needs to be addressed is to define how
Description and Navigator elements.
distribution can occur. For this purpose we have considered
the CARE properties [1], which were introduced for Repeater, which is used to repeat the content
describing multimodal user interfaces, and have already according to data retrieved from a generic data
been applied for this purpose to the MARIA language [4]. source
They are a simple but expressive set of properties:
When the CARE properties are applied to a high element in
Complementarity, indicates that the user interface the structure then then they are automatically inherited by
elements that we are considering are distributed all its children. Thus, for example if we indicate that a
across multiple devices; presentation is redundant across two devices (one desktop
and one mobile), it means that there will be two versions of
Assignment, indicates that the user interface
such presentation, one for each device. Consequently,
elements considered are assigned to one given
when from one point of the user interface there is an access
device;
to such presentation the two versions are activated at the
Redundancy, indicates that the user interface same time. Another example is a grouping associated with a
elements considered are replicated in multiple complementarity property. This means that part of the
devices; interactors in it are assigned to one device and part to
another one. The corresponding device of each included
Equivalence, indicates that the user interface interactor is indicated at the grouping level, and does not
elements considered can be supported by either need to be specified again inside the interactor definitions
one device or another. also in this case.
The basic idea is to associate CARE properties to the This mechanism is an efficient way to associate flexibly
various granularity levels considered. There are a few cases devices to various user interface components without
in which one property is not meaningful. In particular, the having to duplicate such device allocation in all the
equivalence property mainly refers to any user input, and interface element definition.
allows the designer to indicate that the user can choose
among multiple devices for entering such input. However, SOME EXAMPLE APPLICATION
the equivalence property is not meaningful for prompt or If we consider a composition operator, such as the
feedback phases of interaction, since such parts are grouping, then the specification of the involved interactors
controlled by the system according to its implementation varies depending on the CARE property. In the case of
and there is no possible choice for the user. redundancy, it is sufficient to list the devices in which all
In this way we can flexibly specify distributions, even at a the grouped interactor should be supported. For example:
very detailed level. For example we can take one interactor <grouping>
to enter some text and specify that it should be distributed <output care_value="redudancy">
across three devices: a mobile, a vocal, and a graphical PC <bind>
with a large screen in such a way that the prompt is <device id="paterno" platform="desktop"/>
complementary, the input is equivalent, and the feedback is <device id="sisti" platform="vocal"/>
redundant. This means that at the beginning the three <device id="iphone_lab" platform="mobile"/>
<device id="manca" platform="multimodal"/>
devices will provide different prompts, which can address
</bind>
various aspects of the text to enter, then the user can freely </output>
choose the device to enter the requested text, and then all <grouping>
the three devices will present the feedback of the actual text
entered by the user.](https://image.slidesharecdn.com/819cnrchi2011paperworkshopdistributedui-120119170449-phpapp01/85/SMARCOS-CNR-Paper-Workshop-Distributed-UI-2-320.jpg)
![The case of complementarity is more complex. There are An even more complex case of distribution is the case in
various possibilities. In one case we can indicate for each which the interactors in a grouping composition are
interactor to what device is associated, and such interactors distributed according to the complementarity property, and
are not distributed in terms of their sub-components then some of such interactors are further internally
(prompt, input and feedback). Thus, in this case at the distributed according to the interaction phase. In this case
grouping level we indicate the interactor/device allocation part of the distribution is specified at the grouping level and
and nothing is specified at the interactor level regarding part at the level of the single interactors. Below an example
CARE properties. This also happens in the case that there is of this case, which distributes one interactor
a complementarity (thus the interactors are distributed single_choice_id to both a desktop and a multimodal
across various devices) in which there is some redundancy device and one interactor vocal_text to one vocal device.
(which means that in this distribution it can occur that some When we move to the specification of the single_choice_id
interactor is duplicated on multiple devices). Below you can interactor, we can see that it is distributed in such a way
see an example in which there is complementarity at the that the input is assigned to the desktop system while the
grouping level. Thus, one interactor (description_video) is prompt and the feedback are complementary across the
associated with one desktop device and another interactor desktop and the multimodal device. Then, the specification
(description_text) is associated with three devices (one of the multimodal interactor must be consistent and using
vocal, one mobile and one multimodal). Then, the the CARE properties to distribute across modalities only
corresponding concrete descriptions are provided. It is those aspects that have been assigned to the multimodal
possible to note that in the case of the concrete multimodal device in the specification of the distribution across
description, CARE properties are used to indicate the devices. Thus, in this example only the prompt and
allocation of the sub-components to the various modalities feedback can be distributed across the modalities of the
(in this case graphical and vocal) available on the device under consideration since the input part was
multimodal device. Since in the example we consider an distributed only to the desktop device.
only-output object, only this aspect is assigned to one of the <grouping>
CARE properties (since in this case the distinction in <output care_value="complementarity">
prompt, input, and feedback does not apply given that it is <bind>
only for interactive elements). <interactor interactor_id="single_choice_id"/>
<device id="paterno" platform="desktop"/>
<grouping> <devices id="manca" platform="multimodal"/>
<output care_value="complementarity"> </bind>
<bind> <bind>
<interactor interactor_id="description_video"/> <interactor interactor_id="vocal_text"/>
<device id="paterno" platform="desktop"/> <device id="sisti" platform="vocal"/>
</bind> </bind>
<bind> </output>
<interactor interactor_id="description_text"/>
<device id="sisti" platform="vocal"/> <single_choice id="single_choice_id">
<device id="iphone_lab" platform="mobile"/> <input care_value="assignment">
<device id="manca" platform="multimodal"/> <devices id="paterno" platform="desktop"/>
</bind> </input>
</output> <prompt care_value="complementarity">
<devices id="paterno" platform="desktop"/>
<description id="description_video"> <devices id="manca" platform="multimodal"/>
<description_desktop> </prompt>
<video src="video.flv" alt="alternative_text"/> <feedback care_value="complementarity">
</description_desktop> <devices id="paterno" platform="desktop"/>
</description> <devices id="manca" platform="multimodal"/>
</feedback>
<description id="description_text"> <choice_element value="ch_1"/>
<!—REDUNDANT DISTRIBUTION --> <choice_element value="ch_2"/>
<description_mobile> <desktop_single>
<text><string>Text </string></text> <drop_down_list>
</description_mobile> <label label_position="left">
<description_vocal> <text_width="200px"><string>Label</string></text>
<speech><content>Text </content></speech> </label>
</description_vocal> </drop_down_list>
<description_multimodal output="redudancy"> </desktop_single>
<!-- [graphical part] --> <multimodal_single
<text><string>Text</string></text> prompt="complementarity" feedback="redundancy">
<!-- [vocal part] --> <drop_down_list>
<speech><content>Text</content></speech> <label label_position="left"><text_width="200px">
</description_multimodal> <string>Label</string></text></label></drop_down_list>
</description> <single_vocal_selection>
</grouping> <!-- PROMPT -->
<question><content>Choose...</content></question>](https://image.slidesharecdn.com/819cnrchi2011paperworkshopdistributedui-120119170449-phpapp01/85/SMARCOS-CNR-Paper-Workshop-Distributed-UI-3-320.jpg)
SMARCOS CNR Paper Workshop Distributed UI
- 1. Distributing User Interfaces with MARIA Marco Manca, Fabio Paternò CNR-ISTI, HIIS Laboratory Via Moruzzi 1, 56124 Pisa, Italy fabio.paterno@isti.cnr.it ABSTRACT since each device has specific interaction resources and The increasing interest for distributed user interfaces needs implementation languages to execute such user interfaces. languages that better support their specification in multi- device environments. In this paper we discuss an approach An interesting approach has been proposed [6] through the to specifying this type of user interfaces, and how it is use of a XML-based language called HLUID (High Level applied to an extension of the MARIA language aiming to UI Description), in which the user interface has a provide flexible support for the description of distributed hierarchical structure and the leaves of the tree are Abstract user interfaces. Interactor Object (AIOs) describing high-level interactors. During the rendering process the AIOs are mapped onto Author Keywords Concrete Interaction Object (CIOs) associated with the Distributed User Interfaces, Model-based User Interface current platform. In addition, they introduce a split concept Design, User Interface Description Languages. for the groupings through an attribute that, when it is set to true, allows the distribution of the user interface elements ACM Classification Keywords without losing its logical structure. H5.m. Information interfaces and presentation (e.g., HCI): In our case we propose a different solution, still using a Miscellaneous. model-based approach. One difference is that we support the specification at the concrete level because at this level it General Terms is easier to generate the corresponding implementations and Design there is a better understanding of the actual effects that can be obtained. INTRODUCTION One of the main technological trends is the steadily In the next section we describe the design concepts that increasing number of devices per person. This has an characterise the proposed solution for specifying user impact on the user interface languages and technologies interfaces distributed across various devices with MARIA. because it will be more and more common to interact with Our approach is not limited to graphical user interfaces but an application through multiple devices. A typical can address also devices supporting other modalities, such applications scenario is using a graphical or a vocal mobile as voice. Then, we also provide some application examples. device to interact with a large screen Lastly, some conclusions are drawn along with indications for future work. This trend has stimulated increasing interest in Distributed User Interfaces (DUIs), which have been defined [2] as THE PROPOSED SOLUTION those interfaces whose different parts can be distributed in The proposed solution has been introduced for the MARIA time and space on different monitors, screens, and language [5], which provides both an abstract and a set of computing platform, depending on several parameters concrete languages. The concrete languages refer to expressing the context of use, such as the user, the different platforms, characterised by their interaction computing platform, and the physical environment in which modalities (graphical desktop, graphical mobile with touch, the user is carrying out her interactive task. vocal, graphical and vocal combination). Thus, a distributed Model-based approaches [3] have often been considered in concrete user interface should be able to indicate how the order to manage the increasing complexity derived from user interface elements are distributed among devices, managing user interfaces in multi-device environments, which can belong to such various platforms. One important point to consider is at what level of granularity the distribution can occur. We have identified 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 four possible levels: not made or distributed for profit or commercial advantage and that copies Presentation, which means the set of elements that bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior characterise the possible perceivable information specific permission and/or a fee. at a given time (e.g. a page in a Web application); CHI 2011, May 7–12, 2011, Vancouver, BC, Canada. Copyright 2011 ACM 978-1-4503-0267-8/11/05....$10.00.
- 2. Composition of elements, groups of elements In the application of the CARE properties, we consider a inside a presentation that are logically related user interface specification as a hierarchical structure: the among them; user interface is composed of presentations, presentations contain single interactor elements, and compositions of Interactor element, single user interface elements; such elements. In particular MARIA has various operators Interaction phase, we have identified three general for composition of interactors: phases that characterise interactions: prompt, The different types of interactor compositions are: input, and feedback, and it is possible to distribute the user interface even across such sub- Grouping: a generic group of interactor elements. components of one interactor. For example, Relation: when two or more interactors or entering a text through a mobile device and compositions are related to each other. receiving feedback both on the mobile and a large screen device. Composite Description: represents a group aimed to present contents through a mixture of The next aspects that needs to be addressed is to define how Description and Navigator elements. distribution can occur. For this purpose we have considered the CARE properties [1], which were introduced for Repeater, which is used to repeat the content describing multimodal user interfaces, and have already according to data retrieved from a generic data been applied for this purpose to the MARIA language [4]. source They are a simple but expressive set of properties: When the CARE properties are applied to a high element in Complementarity, indicates that the user interface the structure then then they are automatically inherited by elements that we are considering are distributed all its children. Thus, for example if we indicate that a across multiple devices; presentation is redundant across two devices (one desktop and one mobile), it means that there will be two versions of Assignment, indicates that the user interface such presentation, one for each device. Consequently, elements considered are assigned to one given when from one point of the user interface there is an access device; to such presentation the two versions are activated at the Redundancy, indicates that the user interface same time. Another example is a grouping associated with a elements considered are replicated in multiple complementarity property. This means that part of the devices; interactors in it are assigned to one device and part to another one. The corresponding device of each included Equivalence, indicates that the user interface interactor is indicated at the grouping level, and does not elements considered can be supported by either need to be specified again inside the interactor definitions one device or another. also in this case. The basic idea is to associate CARE properties to the This mechanism is an efficient way to associate flexibly various granularity levels considered. There are a few cases devices to various user interface components without in which one property is not meaningful. In particular, the having to duplicate such device allocation in all the equivalence property mainly refers to any user input, and interface element definition. allows the designer to indicate that the user can choose among multiple devices for entering such input. However, SOME EXAMPLE APPLICATION the equivalence property is not meaningful for prompt or If we consider a composition operator, such as the feedback phases of interaction, since such parts are grouping, then the specification of the involved interactors controlled by the system according to its implementation varies depending on the CARE property. In the case of and there is no possible choice for the user. redundancy, it is sufficient to list the devices in which all In this way we can flexibly specify distributions, even at a the grouped interactor should be supported. For example: very detailed level. For example we can take one interactor <grouping> to enter some text and specify that it should be distributed <output care_value="redudancy"> across three devices: a mobile, a vocal, and a graphical PC <bind> with a large screen in such a way that the prompt is <device id="paterno" platform="desktop"/> complementary, the input is equivalent, and the feedback is <device id="sisti" platform="vocal"/> redundant. This means that at the beginning the three <device id="iphone_lab" platform="mobile"/> <device id="manca" platform="multimodal"/> devices will provide different prompts, which can address </bind> various aspects of the text to enter, then the user can freely </output> choose the device to enter the requested text, and then all <grouping> the three devices will present the feedback of the actual text entered by the user.
- 3. The case of complementarity is more complex. There are An even more complex case of distribution is the case in various possibilities. In one case we can indicate for each which the interactors in a grouping composition are interactor to what device is associated, and such interactors distributed according to the complementarity property, and are not distributed in terms of their sub-components then some of such interactors are further internally (prompt, input and feedback). Thus, in this case at the distributed according to the interaction phase. In this case grouping level we indicate the interactor/device allocation part of the distribution is specified at the grouping level and and nothing is specified at the interactor level regarding part at the level of the single interactors. Below an example CARE properties. This also happens in the case that there is of this case, which distributes one interactor a complementarity (thus the interactors are distributed single_choice_id to both a desktop and a multimodal across various devices) in which there is some redundancy device and one interactor vocal_text to one vocal device. (which means that in this distribution it can occur that some When we move to the specification of the single_choice_id interactor is duplicated on multiple devices). Below you can interactor, we can see that it is distributed in such a way see an example in which there is complementarity at the that the input is assigned to the desktop system while the grouping level. Thus, one interactor (description_video) is prompt and the feedback are complementary across the associated with one desktop device and another interactor desktop and the multimodal device. Then, the specification (description_text) is associated with three devices (one of the multimodal interactor must be consistent and using vocal, one mobile and one multimodal). Then, the the CARE properties to distribute across modalities only corresponding concrete descriptions are provided. It is those aspects that have been assigned to the multimodal possible to note that in the case of the concrete multimodal device in the specification of the distribution across description, CARE properties are used to indicate the devices. Thus, in this example only the prompt and allocation of the sub-components to the various modalities feedback can be distributed across the modalities of the (in this case graphical and vocal) available on the device under consideration since the input part was multimodal device. Since in the example we consider an distributed only to the desktop device. only-output object, only this aspect is assigned to one of the <grouping> CARE properties (since in this case the distinction in <output care_value="complementarity"> prompt, input, and feedback does not apply given that it is <bind> only for interactive elements). <interactor interactor_id="single_choice_id"/> <device id="paterno" platform="desktop"/> <grouping> <devices id="manca" platform="multimodal"/> <output care_value="complementarity"> </bind> <bind> <bind> <interactor interactor_id="description_video"/> <interactor interactor_id="vocal_text"/> <device id="paterno" platform="desktop"/> <device id="sisti" platform="vocal"/> </bind> </bind> <bind> </output> <interactor interactor_id="description_text"/> <device id="sisti" platform="vocal"/> <single_choice id="single_choice_id"> <device id="iphone_lab" platform="mobile"/> <input care_value="assignment"> <device id="manca" platform="multimodal"/> <devices id="paterno" platform="desktop"/> </bind> </input> </output> <prompt care_value="complementarity"> <devices id="paterno" platform="desktop"/> <description id="description_video"> <devices id="manca" platform="multimodal"/> <description_desktop> </prompt> <video src="video.flv" alt="alternative_text"/> <feedback care_value="complementarity"> </description_desktop> <devices id="paterno" platform="desktop"/> </description> <devices id="manca" platform="multimodal"/> </feedback> <description id="description_text"> <choice_element value="ch_1"/> <!—REDUNDANT DISTRIBUTION --> <choice_element value="ch_2"/> <description_mobile> <desktop_single> <text><string>Text </string></text> <drop_down_list> </description_mobile> <label label_position="left"> <description_vocal> <text_width="200px"><string>Label</string></text> <speech><content>Text </content></speech> </label> </description_vocal> </drop_down_list> <description_multimodal output="redudancy"> </desktop_single> <!-- [graphical part] --> <multimodal_single <text><string>Text</string></text> prompt="complementarity" feedback="redundancy"> <!-- [vocal part] --> <drop_down_list> <speech><content>Text</content></speech> <label label_position="left"><text_width="200px"> </description_multimodal> <string>Label</string></text></label></drop_down_list> </description> <single_vocal_selection> </grouping> <!-- PROMPT --> <question><content>Choose...</content></question>
- 4. <!-- FEEDBACK --> Lastly, there is an activator interactor (validate button) <vocal_feedback><content>You_choose...</content> </vocal_feedback> whose input and prompt are assigned to the B1 device while </single_vocal_selection> the feedback (the message “Your answer was RIGHT !!!”) </multimodal_single> is assigned to the B2 device. </single_choice> CONCLUSIONS AND FUTURE WORK <description id="vocal_text"> <output care_value="assignment"> We have presented the method that we are using to develop <devices id="sisti" platform="vocal"/> the possibility of specifying distributed user interfaces with </output> MARIA. Our approach allows designers to specify <description_vocal> distribution in a flexible way, at various levels, not only at <speech><content>Vocal text to speech </content> the level of interaction elements as it happens in other </speech> approaches. This is based on a logical framework, which </description_vocal> distinguishes both the granularity levels for distribution and </description> the properties indicating how such distribution should be A similar approach can be applied also when the other applied to the considered elements. Such framework allows composition operators of the MARIA language are us to have a better understanding of the design space for considered (relation, repeater and composite description). In distributed user interfaces, and exploit it in the design of the addition, such approach can be considered also when new features for the MARIA language. distributing interactors or composition of interactors at the presentation level. Future work will be dedicated to supporting authoring of distributed user interfaces extending the MARIAE EXAMPLE APPLICATION environment for this purpose. In this section we show an example application considering a museum application in which a user plays an association ACKNOWLEDGMENTS game on a mobile device and at some point decides to This work is supported by the EU ICT STREP SERENOA migrate to a distributed user interface across a mobile and a Project and the EU ARTEMIS SMARCOS Project. large screen (see Figure 1). REFERENCES 1.Coutaz J., Nigay L., Salber D.,.Blandford A, May J., Young R., 1995. Four Easy Pieces for Assessing the Usability of Multimodal Interaction: the CARE Properties. Proceedings INTERACT 1995, pp.115-120. 2. Demeure, A., Sottet J.S., Calvary G., Coutaz J., Ganneau V., and Vanderdonckt J.. The 4C Reference Model for Distributed User Interfaces. Proceedings of the Fourth International Conference on Autonomic and Autonomous Systems 2008, pp. 61-69. 3.Fonseca J.M.C. (ed.), W3C Model-Based UI XG Final Report, May 2010, available at http://www.w3.org/2005/Incubator/model-based-ui/XGR- mbui-20100504/ 4.M.Manca, F.Paternò,. Supporting Multimodality in Service-oriented Model-based Development Environments, Proceedings HCSE 2010, 3rd Conference on Human-Centred Software Engineering, pp.135-148, LNCS 6409 Springer, Reykjavik, October 2010. 5. Paternò F., Santoro C., Spano L.D.: MARIA: A Figure 1: The example of distributed user interface. universal, declarative, multiple abstraction-level language for service-oriented applications in ubiquitous The original grouping of elements, (a) in the figure, is environments. ACM Trans. Comput.-Hum. Interact. distributed in a complementary way on the B1 (the mobile) 16(4): (2009). and B2 (the large screen) devices. The description (text) 6.Vandervelpen C., Conix K., Towards Model-Based interact (b) is distributed in a redundant manner, while the Design Support for Distributed User Interfaces, NordiCHI description (text) (c) is assigned only to the B2 device, 2004: 61-70, 2004. along with the images (d). On the other hand, there are four text edit interactors that are assigned only to the B1 device.