Proceedings Template WORD doc (DOC download) by liningnvp

VIEWS: 1 PAGES: 6

									              Pervasive Computing Devices for Education
                                                 Ralph Deters
                                           University of Saskatchewan
                                        Department of Computer Science
                                             deters@cs.usask.ca


ABSTRACT
This paper has two aims. First it is an attempt to articulate the possibilities and challenges of using mobile
devices to support learning. Second, it gives an overview of our current projects in developing/deploying tools
that support students in using available on-line resources and collaborating with each other.

1. Introduction
Every new trend in the IT-sector is presented as the most important development ever. During the nineties it was
the web, which promised to revolutionize the way we work and learn. The protagonists of the new media were
quick to promise that learning would become easier, more productive and of course more affordable. And while
many questions concerning costs and benefits of using web-based applications to support learning and
collaborations among learners are still unanswered, the mobile and ubiquitous computing trend has arrived. And
once again every major IT show (like Comdex or CeBit) presents the all too familiar prediction that the impact of
this new development will be tremendous and that we will all be working and learning in new better ways. In
this paper we will try to analyse some of the existing pervasive technologies and speculate about their potential
applications for education.
We start with a brief discussion of the potential of mobile and ubiquitous learning. This is followed by a
discussion of the currently available mobile devices. The client server model is presented as the most likely
software model to emerge in the area of mobile wireless devices. Adaptation and the use of agents is discussed in
the sections 5 and 6. Finally we report about a few of our current research projects and conclude with a summary.


2. Ubiquitous Learning
The IT industry is currently experiencing a paradigm shift. In retrospective the nineties were dominated by
discovering the potential of HTTP and the fairly simple markup language HTML as a means for creating uniform
clients. But the clients and servers were bound to the wired Internet, which represents a serious limitation for an
inherently mobile society. With notebooks that and mobile phones it became possible to overcome the limitations
of stationary Internet access. The initially high connection fees began gradually to drop and consequently mobile
Internet access became more common. With the drop of data transmission fees and the widespread use of mobile
computing devices mobile and ubiquitous computing is slowly becoming a reality.
While it seems obvious that mobile Internet access enables new services, it seems difficult to determine its
impact on supporting the needs of learners. Our research is focusing on the possibilities and difficulties in using
the mobile computing devices to support the learner in a ubiquitous way. Can we provide reliable help and
support for a learner anytime and anywhere?
Since the question of how to support a learner is domain specific, it would seem ambitious to give a general
answer. Consequently we focus our research on the specific problems of computer science undergraduate
students. How can we support them using mobile and wireless devices? Is it possible to reduce the dropout rate in
introductory computer science classes? Is it possible to increase the quality of learning for graduating students by
providing them with tools that enable ubiquitous access to already existing online resources? And maybe most
importantly is it possible to foster the formation of learner groups and teams?


3. Looking at the Devices
The market for mobile and wireless connected devices seems at first chaotic and volatile; new products are
announced nearly daily. But despite the ongoing changes in device-design and features there are currently only
three basic types of devices:
       Very Small Screen Devices (up to 4 inches)
       Small Screen Devices (up to 8 inches)
       Medium Screen Devices (up to 15 inches)
Discriminating devices based solely on their screen size seems at first questionable. But the screen size of the
device typically correlates with the computational resources like memory and processor and the type of wireless
connection. In addition the size of the screen determines the types of possible interactions. The reason for this
correlation is that the size of the screen (being the larges part of the device) determines the available space for the
other parts. As the screen size grows there is more space in the back of the unit for other parts. Consequently
small screens lead to less computational resources and vice versa.
In the following a brief overview of the different devices is given.
3.1 Very Small Screen Devices
Among the many mobile and wireless connected devices those with very small screen seem at first useless. The
user is limited to a typically four-line colorless LCD display. The devices offer a simple keyboard that allows a
simple interaction. Typical examples of these devices are the web enabled mobile phones (WAP), and the
popular Blackberry device of RIM.




         RIM’s Backberry Device                    WAP Phones: Present (Siemens C35i) Future


The main characteristic of these devices is that they allow the user to send and receive virtually everywhere text
messages and/or text-based web content. They offer only minimal computational resources and a set of
preinstalled applications (e-mail, text only HTTP-client). They are small, continuously connected ultra-thin
clients.
3.2 Small Screen Devices
This category of devices is characterized by displays that range between 5 and 8 inches. Typically these devices
expect the user to interact using a special pen (stylus). Clicking on the screen and recognizing the user’s
handwriting offers a simple & efficient interaction.
Typical examples of small screen devices are the Palm, Visor, and the Windows CE Pocket PCs. The small
screen devices offer between 2 and 32 MB of RAM, and an energy efficient but often slow processor. They all
can be equipped with additional programs and offer as a standard communication device a simple Infrared port.
Data exchange between the devices is simple and reliable as long as the distance between the devices is only a
few inches. To enhance the usability of these devices all of them support wireless modem-connection and/or a
wireless Ethernet connection (802.11b).




                                            Palm Computer
3.3 Medium Screen Devices
Devices with screens between 9 – 15 inches as are considered medium sized. Two types of devices are found in
this category, notebook computers and web-pads. While the notebook computer tends to be a miniaturized
version of a desktop machine, the web-pad represents a new development. Web-Pads can be described best as
notebooks without keyboard and other external devices. Using a touch sensitive display and a good wireless
Ethernet connection (802.11b) these devices are typically far lighter and thinner than notebooks. Unlike the
notebook that can be configured anyway the user pleases, the Web-pad is typically a thin HTTP-client. Offering
energy efficient processor and up to 32 MB of RAM the web-pad is not designed as a platform to run
applications. It is a truly thin web-client that depends on a good Internet connection.




                                            Honeywell’s WebPad


The most interesting feature that distinguishes the web-pad from the notebook computer is that it targets users
with no or only minimal computer experiences. Web-pads are in a way a kind of hardware version of a standard
http-browser. The advantage of limiting a machine to being only a browser is that such a device can be switched
on and off without any harm. Unlike a PC that needs typically minutes to boot up, the web-pad is instantly ready
and can be switched off safely at any moment in time. Web pages are light, compact and relatively robust. The
touch-screen allows for interactivity, and the size of the screen allows it to be used as a convenient, nearly normal
size keyboard. Thus the web-pad represents the long promised web-pc. Due the simplicity of its design the costs
per unit can be kept below 1000 $ (US) which will ensure a successful penetration of the mass market.
4. A Few Words about Bluetooth
After a long period of waiting the first Bluetooth cards are available. Like wireless Ethernet Bluetooth uses radio
frequency based data transmission. The major difference between Bluetooth and wireless Ethernet is the concept
of the Pico-Net. Bluetooth is based on the idea that devices that are within close proximity should establish a
small network, allowing them to access each other’s resources. In contrast to wireless Ethernet, which is based on
more or less static IP addresses, Bluetooth enables a less restrictive low-level communication.
While Bluetooth seems to be an ideal solution for ubiquitous computing the usage of the same frequency band as
wireless Ethernet (802.11b) makes the use of both in one room impossible. We therefore abandon for the time
being research into using Bluetooth in favor of wireless Ethernet.
5. Client-Server - Software Model
Every mobile device allows the uploading of additional applications. Unfortunately our experiences in building
applications for the various mobile devices were not very encouraging. The differences and incompatibilities
between devices (even if they are in the same group) make it practically impossible to develop applications in
reasonable time. Instead of developing applications for each device it seems to be better to view them all as
clients of a XML based markup language.
Viewing them as XML clients allows the use of the three-tier architecture, consisting of XML-client (e.g.
WebPad), web server and the applications, which are running on various network nodes. But unlike the classical
web model the use of the mobile devices introduces the problem of adapting the content to the screen size.
6. Using Agents to Adapt
The diversity of the devices leads to serious hardware adaptation problems. One way to simplify the task is to
keep the content in a XML based form and convert the XML documents to the required type (e.g. HTML,
WML). In addition to transforming the XML document into the required mime type it is necessary to decide what
to display and what to ignore. This content adaptation is highly user specific. Depending on the knowledge of the
user, her current goals and intensions, the same XML document might be displayed in different ways for different
users.
In our current work we assume that all XML documents contain clearly marked information entities. We also
assume that for each information entity we have three different ways to present it, one for very small screens, one
for small screens and one for medium screens.
Personal agents residing on agent servers keep user models, which allow them to perform ranking of the
information entities contained in the XML document. In the first step the agent ranks the information entities
according to the expected utility for the user. In the next step the most important entities are selected and the
agent determines the appropriate presentation, which is sent to the user device e.g. a web-pad.
To simplify the task for the personal agents, we currently use a set of specialized converter agents, which are able
to extract and convert the content from the XML documents into the appropriate MIME type documents.
The use of agents has many advantages – unfortunately execution speed is not one of them. To enable a timely
execution of our agents we were forced to design a proprietary agent framework called DICE that allows us to
cluster various machines. By clustering the machines and distributing the load of the agents we managed to
minimize the delays between user request and page delivery to 1.5 seconds.
7. An overview of our current Projects
Currently we focus on the development of information and collaboration tools for
       Internet enabled mobile phones (WAP Phones)
       Use of Infrared for short range communication
       Web-Pads using wireless Ethernet (802.11b)


7.1 WAP Phones
The Wireless Application Protocol allows the users of WAP enabled phones (nearly all modern mobile phones
are WAP ready) to access the web. WAP is based on two basic concepts, the use of a markup language designed
to support small screens called Wireless Markup Language (WML) and a gateway that connects the Internet to
the WAP devices.
Research is focusing on testing the usability of WAP devices with limited displays as a means for supporting
simple search and notification tasks. We are currently using the WAP infrastructure of our local telephone
company to build a system that allows the students to schedule appointments and to send and request help
requests to their peers. In the future we will link this service to I- Help, a knowledge-based matchmaking peer
help system, which is currently in use (on desk-top computers). The small device will allow ubiquitous access to
helpers and active notification. WAP enabled phones allow the students to communicate via sending of simple
text blocks (~500 bytes) and speech.


7.2 Ubiquitous Computing & Infrared
Nearly all-mobile devices are equipped with an infrared port. Using infrared as a means for wireless
communications has two main advantages, price and no interference with radio frequency based transmissions.
Since the implementation of infrared communication is relatively straightforward we are working on testing the
use of large screen computer devices that can only be accessed by using a palm. The palm would serve as a smart
remote control that would allow the student to identify herself and access appropriate web class resources. This
means that the student can interrupt her work at any time and when she comes back or goes to another computer,
her state will be recreated based on the information stored on the palm. This will allow students to move freely
around campus and to use campus resources independently of where they are.


7.3 Web-Pads
Our current research with web-pads is performed using standard http-browsers on notebook computers that are
connected using wireless Ethernet (802.11b). In many ways the use of web-pads seem to add no additional
problems to the design of purely web-based applications. This will allow students to have access to the Internet
and web-based class resources anywhere, also out of the buildings.


8. Summary
In this paper we presented an overview of current mobile and wireless computing devices. After classifying and
discussing the individual types the client server model was presented as the most likely way to deliver content.
Due to the diversity of the small screen clients there is a need for content adaptation according to hardware and
user interests. We see user modeling and personal agents as a way to achieve such adaptation.
9. References

[1] Blackberry: httptp://www.blackberry.net/
[2] WAP: www.wapforum.com
[3] Web-Pad: http://content.honeywell.com/yourhome/webpad/WebPad.pdf

								
To top