Though there are many software development techniques available, SE theory is still in its infancy. This paper first elucidates the general process to build a scientific theory and then illustrates, through real instances, a general lack of methodology in SE research. One of basic requirements for scientific theory is its understanding power but not usefulness. Starting research through a clear and precise definition of concepts under investigation is extremely important but, unfortunately, not always respected in SE, especially for several fundamental concepts. Formulating hypotheses for explanation and testing afterwards is another basic requirement of science. At this regard, however, serious problems exist in our research. We point out also that the widespread use of set theory to represent computer programs is a serious mistake. From our analysis, we hope more researchers will realize that the state of the arts of SE is still quite poor and pay more attention to improve our research methodology.