conclusions

report.tex

@@ -665,11 +665,29 @@ additional buffer release code after.
 
 \section{TBD Conclusions}
 
-TBD 
+From the analysis of the selected project, \textbf{FasterXML/jackson-core}, 
+we can clearly see that a the patterns implemented are deviating from the standardised definitions.
+For example in the case of the singleton pattern ,
+we can see that we have public constructors in classes implementing this pattern and 
+those constructors are used during testing.
 
-\textit{a brief (possibly speculative) discussion about whether your findings
-are likely to be applicable to other projects or, conversely, they are probably
-unique to the project you selected – and why you think this to be the case.}
+To see the real benefits of the patterns implemented in this project we need to analyse the modules on top,
+because this project is a core library for a more larger project. 
+Being a often used library and a stable library, 
+we can assume that patterns implemented by this project are doing their work properly,
+despite the deviation from the standardised definitions.
+
+From the results obtained from the tool \textit{Pattern4J}, 
+we can see that we have some imprecision for the classifications of the patterns.
+The most striking found is for the decorator pattern,
+the developer stated in the documentation that the design patter used is a delegation patter.
+From a manual analysis of a subset of all the classes of the project,
+we can wee that we have slightly more false positive rate compared to the true positive values,
+so we can consider that in our case the tool have a precision lower that $50\%$.
+The result are reported in Figure \ref{fig:conf}, 
+in Sub-figure \ref{fig:conf:tab} are presented the classes analysed manually
+and in Sub-figure \ref{fig:conf:mat} is presented the confusion matrix computed using the 
+results presented in Sub-figure \ref{fig:conf:tab}.
 
 \end{document}