Software Requirements analysis of quality characteristics as completeness, consistency and unambiguity assume an important role in the safety-critical software. We consider then the contributions to Natural Language (NL) Software Requirements Analysis as good practices or recommendations for producing high level quality NL requirements. From a survey of different approaches and solutions, it is possible to draw an incremental list of Good Practices (GP) for writing NL software requirements, making them understandable to their users, typically from the linguistic point of view, thus reducing the efforts for the analysis. The industrial experiences of two research/service laboratories about requirements analysis are compared with the suggestions of the literature and the technology, and the results are shown. Keywords: Natural Language, software requirements analysis, good practices
