Our aim was to develop a web shop that could be configured flexibly, distributed freely and is open source under the GNU-General Public Licence. What has arisen is a shopping system that offers articles with infinite options and variations.

The customer can optionally save his address. He can then access his record by simply logging in with a login name and password. This way he doesn't have to enter his address more than once. The shop provides comfortable search functions to find articles. Several types of payment are supported. Thanks to a flexible payment-interface it is possible to link the PhPepperShop to an online-payment institute with ease. PhPepperShop can optionally encrypt the credit card input via SSL.

The shop administrator can access the web shop remotely from any location via a web interface. A flexible layout allows for changing the look of the shop in detail (background images, frame size, shop logo, buttons, etc.). Articles can be created, updated and deleted. An article can even be assigned to more than one category. With the category management the shop administrator can easily create, rename, replace and delete categories and sub categories. Shipping costs can be charged by price, by weight or by an estimated amount. Small amount surcharge and pay on delivery surcharge can be added to the invoice automatically.

The web shop has been programmed using PHP and JavaScript. The HTML pages are generated by PHP, JavaScript validates customer input. The whole shop is saved in a MySQL database. The installation is unlike UNIX with an automatic installation tool written in PERL.

As mentioned in the title, the PhPepperShop is an open source project, which includes free availability of the source code and unrestricted utilisation and development. The PhPepperShop project is maintained on Source Forge. The latest news and a sample shop can be found on www.phpeppershop.com.

With increasing complexity and size of conventional websites the costs for updating content increases proportionally. The Internet is considered to be a fast medium and website visitors expect to find sites that are up to date.

Keeping a Website newsworthy is not a simple task due to certain drawbacks, that for example the hypertext markup language (HTML) brings along. The problem with HTML is, that it mixes layout and content information. Certain scripting languages such as PHP or ColdFusion generate HTML code dynamically from predefined files, as a workaround. To solve the problem of separating content from other information, the eXtensible Markup Language (XML) was invented. XML Files only contain data and no layout or presentation information. At the time of presentation the data from the XML file is run through a specially designed style sheet (XSLT), that transforms and formats the text. Using this new technology content and layout / design information is no longer mixed.

This alone however does not make a website less complex. In order to keep a site up to date efficiently, one needs to use Content Management Systems (CMS). Theses Systems allow people to administrate a website without any knowledge of the technologies needed for such tasks.

Within the scope of this thesis a CMS was to be designed and programmed. The developed program is to be completely web based, which makes it possible to change a website location independently. The only prerequisites are a computer with a browser and an Internet connection.

The developed CMS is implemented with open standard technologies such as Java and XML, along with other open source projects. It's programmed object oriented and implements an MVC structure using the struts technology.

In a project thesis a framework has been built to extract certain key values from a text file. The implemented algorithms worked only on very specific conditions. For example if you want to extract a key value from a document its approximate position had to be defined. The intention of our activity was to develop more intelligent algorithms. Furthermore did we want to develop a controller for the processing steps. We considered it to be very important that the algorithms reach informative findings without too much given specifications from the user. If the user would have to define all the details of the documents which will be searched through he would have to know the contents of the documents. If this would be that way, all the information would already exist and there would be no need for our search algorithms. We are keen that our algorithms reach the best possible results with the given specification from the user.

The daily correspondence of companies generate a huge volume of text documents. If the company wants to archive these documents in a Document Management System, the documents have to be grouped first. There are several possibilities how the documents could be arranged. A very common procedure is to assign the documents to a document class. This situation was cause for the identification of the use case "Klassifizieren". This use case tries to assign a document to a defined document class. Another case is that a user might be interested in the content of a document. Letters mostly consist of several text blocks. We defined the second use case "Blocktyp erkennen". This use case splits a document into blocks. Afterwards the blocks will be assigned to a defined blocktype.

Furthermore a controller for the processing steps has been developed. It was important for us that our controller can be reused if the software will be extended with new algorithms. So the reuse of these components was a very important criterion for the design of the controller.

Make data persistent is a really common problem in object-oriented programming. Most solutions at this time do use Relational Database Management Systems (RDBMS) to solve that problem. In application development there has been used more and more object-oriented technologies to create software components. It is very difficult to map the objects to the relational databases, because there are two different data models. The challenge to make objects persistence is the function of our ?Persistence Container?. The status of certain objects will be saved over lifetime of processes in a relational data-base management system. Such objects are described as persistent objects. If the status of objects are changed, the Container is going to save that actual state. The goal of that persistence over lifetime is that users of our Container don?t have to restore data in case of a restart of the program. It seems to be that the program did never terminate. The effectively storage is delegated to a Persistence Service, which was the fundament of our diploma work. We got the Persistence Service from Mr. Reto Ferri. Through a simple transactional mechanism we got the ability to change objects and at the end to decide if all or nothing will be made persistent. Objects are allowed to be in multiple or in singleton references to each other. The ?Persistence Container? is also not only able to handle 1:1 references. It is also able to handle 1:n references. If there are 1:n connections between objects, the user has to create keys. The keys do allow him to access the sub objects of objects. The actual Version of ?Persistence Container? is able to handle single thread mode. Be-cause of that it is not allowed to have different processes, which access the ?Persistence Container? in parallel.

Minimally invasive surgery became more and more important over the last few years. This lead to an increasing demand for training equipment for surgeons. Virtual-Reality based simulation technology offers an enormous potential for surgical training.

Along with other universities and commercial companies, the Zurich University of Applied Sciences Winterthur takes part in an international research program with the goal of building a framework for surgical simulations. The goal of this dissertation is to gain basic knowledge of the technologies used in future projects.

Our assigned task was to develop a simple simulation program for minimally invasive surgery. The most important requirements were to implement organs with a realistic behaviour and to achieve a fluent animation on the screen. The user may move the camera and the instrument with an easy to use interface. Any organs touched by the instrument or another organ are supposed to deform in a manner so that they do not intersect with any other objects. A deformed organ slowly accelerates in the direction of the applied force and it slows down as soon as it does not touch another object anymore.

After having spent some days learning the basics of the technologies used, we gradually implemented the surgery simulator. First, we concentrated on the visual aspects of the organs and the instrument shown on the scene. Then the physical behaviours were added one after another. The result is a simulation software that fulfils all requirements.

The physical model on which the simulation is based is a mass-spring system. It is quite complex and it uses a lot of computational power. In order to achieve a fluent animation of the scene on the screen, we had to drastically simplify the model. Despite the partly incorrect physics in the model, the resulting animation looks quite natural.

Nowadays so-called "Embedded Systems" are in widespread use. The Institute of Mechatronic Systems IMS is active in this field. Usually the Java-based real-time operating system Jbed is used with sensor driven control systems. While running, such systems constantly produce large quantities of process data which is needed to supervise, configure and optimise the system.

We have developed a visualisation framework which can be used to capture this process data and visualise it on a host computer.

The visualisation framework is divided into three parts:

The data collection's task is to collect the process data and store it in a suitable data structure. The processes which collect the data are implemented as real-time tasks as well. This grants that the process data is captured within an exact pre-defined period. Furthermore, data collection processes are triggerable, meaning an event can be defined and when it occurs the data collection is started again.

The communication module is responsible for the communication between target and leading system. This is realised using remote method invocation RMI, which makes it possible to easily exchange data as well as calling remote procedures.

The last component serves to visualise the collected data. Several views are possible so process data can be visualised in different ways. How to configure the different data collection processes as well as which view is to be used, is stored in a configuration file. The different components have been arranged to make up an example host so as to demonstrate the complete visualisation framework.

Current drawing programs allow only mouse and keyboard input. For several operations,for example selecting and moving objects, this is probably the most efficient method. On the other hand, many operations are rather difficult to carry out per mouse, for example setting the stroke widths or aligning objects. Here spoken commands would be more efficient, because the user often knows very exactly what he wants but not where to find the corresponding option.

The objective of the current diploma thesis therefore is, to investigate the advantages of multimodal input for a common program such a a drawing program. The current diploma thesis has started from the concept of an existing multimodal drawing program with rather basic functionality. The aim was to better demonstrate the possibilities of multimodal input by significantly enlarging the functionality of the program. In addition, a new speech recognition software had to be integrated into the program. The result of the thesis is a new drawing program which tightly merges graphical and spoken user input into a flexible multimodal input where the user can arbitrarily switch between the two modalities: It is e.g. possible to select a menu item by starting with a voice command and then completing the selection by mouse. The voice command thereby guides the mouse to the menu item chosen and the menus popping up give the user clues to the allowed voice commands. In addition, when the user navigates through a menu by mouse the focus of the command tree is shifted accordingly.

One of the major drawbacks of this concept at the moment is that the navigation through the menus by mouse is slowed down because of the needed focus shift in the command tree in the spoken dialogue.

E-voting by mobile phones makes it possible for every citizen to vote by mobile phone anywhere and at any time. Legal data protection is guaranteed for the complete voting process. The system is functional for both referendums and elections.

The voting server "eVoting" developed by Hans Pan and Werner Sennhauser provides a voting service accessible by internet browsers. In the course of this project we have developed a special client for mobile phones for the mobile access to this voting service. Our solution is based on Sun Microsystems Java 2 MicroEdition Platform (J2ME) - specially developed for equipment with reduced resources. Based on a detailed analysis we conclude that the voting and election procedure by mobile phone that we implemented will at least meet the security requirements of conventional voting and election processes, i.e. without any electronic means. We have applied latest technology partly even in Beta version in the implementation of our mobile phone client, e.g. kSSL (kilo Secure Sockets Layer) software for an optimized SSL-access between mobile phone and server: The server identification is verified by an X.509 certificate, the data encryption is secured by a 128 bit RC4 algorithm. The kSOAP (kilo Simple Object Access Protocol) standard allows the data transfer to the VoteServer in a standardized format. The SOAP Envelopes are transmitted by HTTP to the web server. The SOAP environment directly interfaces to the eVoting-server of Hans Pan and Werner Sennhauser.

As the link from the mobile client to the server is based on TCP/IP, the user only needs access to an internet service provider on his mobile phone.

The firm ABB extends their switches for middlevoltage (24'000V/1'600A) with internet-technology. It should be possible to control a switch from places all over the world.

In the WEBREF-Project a team of several engineers developed an embedded web server for the switchs. This small web server offers measurement, alarms and other data of a switch over standardised technology. Most of the information are available as XML-Files (eXtensible Markup Language). Some are static, others are generated dynamic and provided over CGI (Common Gateway Interface). A Webpage presents the data in a user friendly view, using JavaScript and DOM. So a browser is everything you need to control the switch.

ABB has a vision: It sould be possible to control such a switch also by a PalmPilot (Handheld, PDA). So perhaps in the future there's no need of an expensive interface for monitoring and controlling at every switch. The technician just takes his Palm and get all of the informations he need over Bluetooth. He switches thousands of volts just tabbing a button on his handheld. Unfortunately there's no browser for PalmOS available yet which can handle the dynamically generated webpages of WEBREF (lot of JavaScript code). Also the Palm has a really small screen with just 160 x 160 pixels, so there's not a lot of space to display information.

In this graduate work our team of 3 students developed a PalmOS application which takes over the jobs of the browser. This application receives the data over a serial connection using HTTP and displays it.

Besides a parser and renderer for SVG (scalable vector graphic) was developed. Some measurements and a schematic diagram of the switch are available in this open format, a W3C recommendation (world wide web consortium).

Generally speaking, a SMS server is a system that is able to receive and to send short messages. Its functionality can be increased massively by using interfaces for establishing connections to other communication channels, e.g. email. Such a system can be used for example as e-commerce application. The message length is restricted to 160 characters, for example enough to take an opinion poll.

Our task was to develop a communication platform implenting such a SMS server service. High demands have to be made on reliability and supervision of the communication channels. It's very important not to lose any messages in case of system crashes. The connection to the different communication channels shall be realised with several independent plugins that must be easy to implement. In this way the capability of extension will be guaranteed.

Our final version of the SMS server fulfils all requirements. The message filtering allows to send messages from an arbitrary source to an arbitrary destination. In other words, the functionality of our server software goes beyond the simple sending of short messages. Furthermore, it is possible to specify a group name as source or target. Therefore, the server allows not only special handling of all messages sent by members of a group, but also sending one message to all members of a group (multicast). By binding the server to a database, we can store information about messages, filter rules and groups permanently. A recovery function ensures that no messages are lost by a program crash. Additionally, this feature tries periodically to send valid messages which couldn't be delivered by the server till now. The monitoring of the SMS server is implemented by a management interface and a log file. We directed our attention to a correct functionality of all parts.

Microsoft Switzerland
SIG Pack Systems AG, CH-8222 Beringen

This thesis has been implemented as a sub-project which is a part of an overall project of the SIG Pack Systems Ltd./ICC. The intention of the overall project, eCap2, was to automate and optimise the process sequences in the supply chain between the bottler Coca Cola Beverages in Vienna and the packaging company SIG. As a technical solution all the forecast, ship, invoice and order processes were mapped as a workflow in the BizTalk 2000 Server of Microsoft. The sub-project of the thesis addresses the subject of how the quality of service can be increased. To achieve this goal the main focus was directed at the following three parts: processes, system data and availability.

The availability check tests the state of all existing systems (SAP R/3, web server, mail server, etc.) which are part of the supply chain. The received data gets stored for immediate or future analysis.

Another important part of the thesis work is the control of the process flow. In the implemented scenario, the process flow and the maximal process transition time will be compared to controlled values which can be changed and maintained periodically. If a process transition takes longer than the maximum defined time or if a process is missing in the segment, the person in charge will be informed immediately by e-mail. Because the process flow behaves in an asynchronic manner, the controlling of the processes has to be realized in a very dynamic way.

In the system data control, the stock inventory of the packaging of the bottler will especially be checked. To achieve that goal the stock inventory will be compared with a predefined critical stock. These values are also strongly influenced by the delivery time. If the stock falls under the critical limit within the delivery time, the person in charge at the SIG will be informed.

The main focus of this thesis is the application of the J2EE Architecture. The strengths and known weaknesses of the actual version (1.1) are analyzed. In respect of version 2, which is still in draft status, other approaches and general ideas for improvements for existing issues are proposed.

Another focus is the application of Struts (Apache Software Foundation). These build a framework exploiting Java Server Pages (JSP) and Servlet technology, defined by Sun Microsystems. Struts allow for a clear separation between user interface and application logic. The approach follows the Model View Controller (MVC) design pattern.

Further investigations have taken place how the Struts logic can be combined with the visual parts of the application. An important aspect within this context is the cooperation between the web publisher and the application (logic) developer. Resulting issues are analyzed and appropriate solutions are evaluated.

As use case and for validation of the concepts, a prototype of an option rating system developed by the COMIT group has been developed. Especially, the aspects of reuse of existing and newly constructed parts have been focused.

Java has become widely accepted in the PC and enterprise area. Sun has published the Microedition with the CLD-Configuration and the MID-Profile as a further Java 2 architecture for small and mobile devices. It is an alternative to the Palm SDK for software-development on Palm OS. This study explored the possibility whether Java is suitable for Palm OS software-development and which features are supported. Two applications have been developed: an application to gather data offline with data-synchronisation and an IR-Remote-Control to control a simulated robot. An alternative technology from Esmertec called JBed has also been investigated. The performance and other properties have been compared and the results showed that JBed needs more storage space and is not necessarily faster. Sun has defined J2ME as a framework on which manufacturers should implement compatible Java-Virtual-Machines (JVM). Therefore it depends on the manufacturer of the JVM, which features are supported and which are not. Suns KVM for Palm OS offers, by contrast to JBed, no serial-port support. Both JVMs implement networking capabilities. Sun's KVM offers no floating-point support. The GUI design is restricted as well. Some features that are not supported, can be obtained from other manufacturers. I come up with the following conclusion: In principle - with a few restrictions - Java is suitable for application-development on Palm OS.

An Information Broker compiles data from various information sources and presents them in a suitable form to the customer. A possibility for companies to offer their information on the internet is the standard protocol SOAP (Simple Object Access Protocol), which however requires a considerable initial effort for many companies. An easier way for companies to provide information on the internet are HTML-sites, which are easily accessible on their webservers.

It was the object of this thesis to develop a component that is accessible on the internet and makes it possible to extract the desired information from different HTML-sites and presenting it in the way according to the customer?s demands.

We implemented a server that handles SOAP-requests for which it collects the required information from the internet and gives it back to the customer by way of SOAP-response. The component can be adapted to different HTML-sites by using a description file in XML. The system supports authentication as well as session handling (cookies, url-encoding, hidden-fields).

The main goal of this diploma thesis consisted of sketching and developing a client-server architecture which extends an existing component-framework with a distribution mechanism for log-messages. The already available framework was developed by "SunGard Trading and Risk Systems" in Zurich. It serves the purpose to simplify the integration of different systems and their data while offering the possibility to read some data out of an arbitrary source, to transform it according to user-defined rules and to finally write it into an arbitrary target. For this kind of process, the term "Dataflow" is used.

As the analysis phase revealed, the existing system posed a serious problem which had to be solved early in the design phase: most of the generated log data could not be uniquely associated with a Dataflow. Our system solves the problem with an unusual but at the same time very simple approach: It executes each Dataflow inside a separate operating system process.

While having been forced to use Java as the programming language and target platform, we were allowed to freely specify the design of our system and to choose the network technologies to be used. A comparison of the most important technologies being applicable resulted in the decision to use RMI as the communication protocol. Although at first sight, a multicast technology seemed to be appropriate, the degree of complexity of such a solution would have been out of all proportion to the gain in bandwidth.

One of the main requirements to our solution consisted of possibly keeping the coupling to the existing framework very tiny in order to allow it to be further developed as independently as possible. We succeeded in narrowing down the necessary adjustments to the existing system to 22 characters of source code!

The main goal of this work was to develop an object oriented, real-time framework based on the operating system vxWorks 1) to control the mobile robot platform Cressida. In addition, a driver framework was designed and implemented to support the various peripheral hardware components which are connected to the main processor board through a FPGA (Field Programmable Gate Array). The object oriented software design of the framework (C++) was done with the UML case tool "together".

The framework consists of four interacting software components implemented as vxWorks tasks. At the bottom of the software hierarchy is Cox, which controls the power of the engines on the Cressida chassis and receives its commands from the Navigator. The Navigator plans and controls the movements and actions of Cressida which are set up by the Manager who implements the robot application. All activities within Cox, Navigator and Manager are coordinated by the Captain who is located at the top of the whole software system and is also responsible for supervising the processor hardware.

During our work we encountered several problems especially with vxWorks since this is the first major work with this operating systems at our school. Especially running objects on different interacting tasks was not easy to accomplish.

The current state of the work is a stable framework for the platform implementing all basic tasks to run Cressida. Due to lack of time, we were not yet able to implement a complex application to fully demonstrate the capabilities of the Cressida platform.

1) by WindRiver Systems

The FELA AG is known for their product Tripon in combination with the LSVA (performance related trafic fee for the transportation of goods). This device is equiped with a GSM module which is being used for the data transmission. One of the further developped products is a universal Remote-Metering module that exchanges the data and commands with a communication center by means of GSM. The aim of this diploma thesis was to realize a center that communicates with a significant number of such microprocessor based modules through SMS-messages, acquires the data from the modules and makes them available to other applications over internet. Its main goals are:

• Secured reception of SMS with measurement data and its storage in a data base
• Remote service and administration of all configuration data of the modules
• Web access to the data base for the purpose of visualization of the recorded measurement data

The realization of the center has been seperated into two part: server and client.

The server has been entirely programmed in JAVA. It deals with the communication of the modules with the data base. This consists of the reception, transmission and preparation of the message data. The transmission and reception of the data through SMS is being done with a software module supplied by the Dolphin-company.

This software has been adjusted to our demands and communicates through a file interface with the server. The server has an RMI interface for the client interaction with the server. In this way, the client is able to trigger asynchronously a transmission on the server.

The client that has been realized as servlet that administrates the modules and its configurations. The advantage of this solution is the possibility of remote administration over the WEB from any computer connected to the internet.

In this thesis, I have implemented a virtual Chat-System with an attractive graphical Interface containig pseudo intelligent Chat-Roboters discussing topics from "women and technic" (www.frau-und-technik.ch).

The graphical user interface consist of applets embedded into an appealing Homepage. This front end communicates with a Java-Server, an Apache-Webserver and a MySQL Database hosted on a Linux system.

The user selects a discussion topic from a opening choice list and asks than the robots through the mouse clicks about their opinion. The robots representing different human characters gets their answers from a data base. Every message is visualised in a speechblow. The users can also interactively enter their own opinion into the database. It can be than re-called by clicking the fourth chatbot representing the on-line users.

After the Analysis phase, I have started the Design and implemented the server in Java. The design and Implementation of an attractive and reliable user Interface (Homepage with applets) consumed a lot of time. For this reason I had no time left to tackle the problems of artificial intelligence that were also a part of the thesis specifications. The possibility of using Flash5 for the GUI instead of an applet is discussed in the thesis report.

This work is well expandable and offers interesting possibilities within the area of experimentation with artificial intelligence.