Recovery oriented software is built with the perspective that hardware or software failures as well as operation mistakes are facts to be coped with, since they are problems that cannot be fully solved while developing real complex applications. Consequently, any software will always have a non-zero chance of failure. Some of these failures may be caused by defects that could be removed or encapsulated. A key issue is to increase the detectability of errors, in other words, increase the self-awareness of the software s behavior. In this work, we present the results of systematically applying already well known techniques (design by contract, self-checking software, software components, debuggable software, design for testability, mock components and patterns) with the intent of creating recovery oriented software. Measuring the development of five different real-time and real world applications, we analyzed the effects of the adoption of these techniques. In particular we observed the balancing of the effort spent in different development stages and explore the reduncancy of reasoning concept that, as well as providing a higher detectability and debuggability, also leads to enhancing quality-by-construction. The results were encouraging since they were systematically better than those reported in the literature and were achieved at a feasible cost.