Every software program that interacts with a user requires a user interface. Model-View-Controller (MVC) is a common design pattern to integrate a user interface with the application domain logic. MVC separates the representation of the application domain (Model) from the display of the application's state (View) and user interaction control (Controller). However, studying the literature reveals that a variety of other related patterns exists, which we denote with Model-View-* (MV*) design patterns. This thesis discusses existing MV* patterns classified in three main families: Model-View-Controller (MVC), Model-View-View Model (MVVM), and Model-View-Presenter (MVP). We take a practitioners' point of view and emphasize the essentials of each family as well as the differences. The study shows that the selection of patterns should take into account the use cases and quality requirements at hand, and chosen technology. We illustrate the selection of a pattern with an example of our practice. The study results aim to bring more clarity in the variety of MV* design patterns and help practitioners to make better grounded decisions when selecting patterns.

This report describes an examination project made for the IT consultingcompany Sogeti. The purpose of the project was to develop and modify the translation tool Language Manager (LM) built by Sogeti to be used for translating applications. Employees at Sogeti considered some disadvantages with Language Manager, version 1.0 which among others was that language files for projects were saved at two locations. Partly in resource maps among with the source code of the applications and partly in a database. This was dual work for employees at Sogeti and it also caused redundancy inthe system. Also employees at Sogeti thought that the managing of projects and versioning did not adapt to how the system was needed to be used. The destination by the examination project was to remove the database and only use XML-files to handle languages, and also to make the new Language Manager easier to work with. New users should easily understand how to handle terms and translation in the new application and no manual should be needed to perform tasks. Language Manager version 2.0 should be written in C# .Net Framework 4.5 and the graphical user interface should be created with Windows Presentation Foundation (WPF). Sogeti wished for the Model-View-ViewModel pattern (MVVM) to be implemented. The new tool was supposed to be robust and simple with a future-safe architecture. / Den här rapporten beskriver ett examensarbete som genomfördes åt IT-konsultbolaget Sogeti med syfte till att vidareutveckla och omarbeta översättningsverktyget Language Manager (LM) som var tillverkat av Sogeti och som användes till att översätta applikationer. Anställda på Sogeti ansåg att det fanns vissa nackdelar med Language Manager version 1.0 som bland annat var att språkfiler för projekt lagrades på två platser. Dels i resursmappar tillsammans med applikationernas källkod och dels i en databas. Detta medförde dubbelt arbete för de anställda på Sogeti när de arbetade med Language Manager och det orsakade även redundans i systemet. På Sogeti ansåg man även att hanteringen av projekt och versionshanteringen av språkdata i Language Manager version 1.0 inte passade ihop med hur man arbetade med programmet. Målet med examensarbetet var att avlägsna databasen och endast arbeta med XML-filer som förvaring av språk och att Language Manager version 2.0 skulle bli enklare och mer lättarbetat. Nya användare skullemed lätthet förstå hur hantering av termer och översättning skulle göras utan hjälp av manual. Det nya översättningsverktyget skulle skrivas i C# .Net Framework 4.5 och Windows Presentation Foundation (WPF) skulle användas för att implementera det grafiska gränssnittet tillsammans med Model-View-ViewModel-mönstret (MVVM).Särskild inriktning skulle framför allt vara mot robusthet, enkelhet och med en framtidssäker arkitektur.

The goal of this thesis is to create a mobile application for the Android operating system, which makes organizing a community shopping easier for its users and their friends. Each user can build a community of their own that offers collective purchasing of products, or can offer their products for purchase. They categorize these offers to their channels and their friends choose only those channels that they find interesting. If their friends accept their offer, they see the necessary information about their individual purchases. Thus, this application enables users to offer shared purchasing to their friends e.g. at a grower from whom these people could not buy individually. They can procure quality produce together at a competitive price and they can also reduce production of plastic packaging by purchasing as a group.

Det mobila operativsystemet Android är idag ett ganska dominerande operativsystem på den mobila marknaden dels på grund av sin öppenhet men också på grund av att tillgängligheten är stor i och med både billiga och dyra telefoner finns att tillgå. Men idag har Android inget fördefinierat designmönster vilket leder till att varje utvecklare får bestämma själv vad som ska användas, vilket ibland kan leda till onödigt komplex kod i applikationerna som sen blir svårtestad och svårhanterlig. Detta arbete ämnar jämföra två designmönster, Passive Model View Controller (PMVC) och Model View View-Model (MVVM), för att se vilket designmönster som blir minst komplext med hjälp av att räkna fram mätvärden med hjälp av Cyclomatic Complexity Number (CCN). Studien är gjord utifrån arbetssättet Design & Creation och ämnar bidra med: kunskap om vilket mönster man bör välja, samt om CCN kan peka ut vilka delar i en applikation som kommer att ta mer eller mindre lång tid att testa. Under studiens gång tog vi även fram skillnader på om man anväder sig av den så kallade Single Responsibilyt Principle (SRP) eller inte. Detta för att se om separerade vyer gör någon skillnad i applikationernas komplexitet. I slutändan så visar studien på att komplexiteten i små applikationer är väldigt likvärdig, men att man även på små applikationer kan se skillnad på hur komplex koden är men också att kodkomplexitet på metodnivå kan ge riktlinjer för testfall. / Today the mobile operating system Android has a firm grip on the mobile market due to its open source code and the availability due to the variety of phones available, from very cheap models to the high-end exclusive ones. However, Android currently has no defined design pattern which causes a bit of a problem during the initial stages of development because the developer has to make that decision before the initial development starts. This sometimes causes excessive complexity in the finished application, which then becomes difficult to test and manage. In this report we compare the two design patterns Passive Model View Controller (PMVC) and Model View View-Model (MVVM) to decide which one is the least complex. For this we used so called metrics to show the complexity of the finished applications. The metrics were calculated using the existing library of functions called Cyclomatic Complexity Numbers (CCN). The study is based on the workflow Design & creation and aim to contribute with: knowledge that will make it easier to choose design pattern when developing an Android application, and if you can with CCN point out which areas in an application will take more or less time to conduct tests on. During the study we also analyzed how big of an impact the Single Responsibility Principle (SRP) had on the complexity. This was part of the study so that we could see the difference of separated views versus non-separated views. The study shows us that the complexity in small applications is very similar, no matter which design pattern you're using. However, it also shows that there are certain parts of the applications that are diverse in their complexity and that code complexity on method-level can give directions to test cases.