An increasing amount of companies and organizations are starting to implement the use of cloud computing in their business. This trend results in that software, which was previously sold and distributed to the customers whom then had to install the software on their own computers, now is being replaced with Software as a Service (SaaS). SaaS makes software available through the customers’ browsers, which results in that the service providers only have to administer a single application. The process to migrate a distributed application to a service delivered as a SaaS lacks sufficient investigation; this paper will provide some guidelines for conducting such a pro- cess. During the work on this paper, a prototype of a service delivered as a SaaS has been devel- oped with the intention to test, among other things, how a distributed relational database can be converted to a key/value pair storage. A conversion of this kind enables data to be stored locally in the customers’ browsers, which relieves some pressure on the server as well as enables the application to be used in offline-mode. The paper results in three guidelines which should be considered when planning to migrate software to a service delivered as a SaaS with a local data- base; Think before you act, Don’t expose your soul and Size matters. These guidelines describe how a migration process should be planned, when an application is not deemed appropriate to migrate and when a conversion of the database is not appropriate. / Allt fler företag och organisationer börjar implementera användandet av olika molntjänster i sin verksamhet. Den här trenden medför att programvaror, som tidigare sålts och distribuerats till kunder vilka sedan själva får installera dem på sina egna datorer, nu börjar ersättas med en Soft- ware as a Service (SaaS). Det innebär att programvaran istället finns tillgänglig på Internet via kundernas webbläsare, något som medför att tjänsteleverantören enbart behöver administrera en enda programvara. Processen att migrera en distribuerad programvara till en SaaS-tjänst saknar ordentlig utredning, det här arbetet syftar därför till att ta fram några vägledande riktlinjer för en sådan process. Under arbetets gång har en prototyp av en SaaS-tjänst utvecklats med syfte att testa bland annat hur en tidigare distribuerad relationsdatabas kan konverteras till nyckel/värde- par. En sådan konvertering möjliggör lokal lagring av data i kundernas webbläsare, vilket mins- kar belastningen på servern samt erbjuder möjligheten att arbeta offline. Arbetet resulterar i tre riktlinjer att beakta då en migration till en SaaS-tjänst med lokal databas planeras; Tänk efter, före, Blotta inte din själ och Storleken har betydelse. Riktlinjerna beskriver hur en migrations- process bör planeras, när en applikation inte anses lämplig att migrera samt när en konvertering av databasen inte anses lämplig.

The venerable desktop metaphor is beginning to show signs of strain in supporting modern knowledge work. Traditional desktop systems were not designed to support the sheer number of simultaneous windows, information resources, and collaborative contexts that have become commonplace in contemporary knowledge work. Even though the desktop has been slow to evolve, knowledge workers still consistently manage multiple tasks, collaborate effectively among colleagues or clients, and manipulate information most relevant to their current task by leveraging the spatial organization of their work area. The potential exists for desktop workspaces to better support these knowledge work practices by leveraging the unifying construct of activity. Semantically-meaningful activities, conceptualized as a collection of tools (applications, documents, and other resources) within a social and organizational context, offer an alternative orientation for the desktop experience that more closely corresponds to knowledge workers' objectives and goals. In this research, I unpack some of the foundational assumptions of desktop interface design and propose an activity-centered model for organizing the desktop interface based on empirical observations of real-world knowledge work practice, theoretical understandings of cognition and activity, and my own experiences in developing two prototype systems for extending the desktop to support knowledge work. I formalize this analysis in a series of key challenges for the research and development of activity-based systems. In response to these challenges, I present the design and implementation of a third research prototype, the Giornata system, that emphasizes activity as a primary organizing principle in GUI-based interaction, information organization, and collaboration. I conclude with two evaluations of the system. First, I present findings from a longitudinal deployment of the system among a small group of representative knowledge workers; this deployment constitutes one of the first studies of how activity-based systems are adopted and appropriated in a real-world context. Second, I provide an assessment of the technologies that enable and those that pose barriers to the development of activity-based computing systems.