Critical software needs to obtain an assessment before com- missioning. This assessment is given after a long task of software analysis performed by assessors. They may be helped by tools, used interactively, to build models using information-flow analysis. Tools like SPARK-Ada exist for Ada subsets used for critical software. But some emergent lan- guages such as those of the ML family lack such adapted tools. Pro- viding similar tools for ML languages requires special attention on spe- cific features such as higher-order functions and pattern-matching. This paper presents an information-flow analysis for such a language specifi- cally designed according to the needs of assessors. This analysis can be parametrized to allow assessors getting a view of dependencies at several levels of abstraction and gives the basis for an efficient fault tolerance analysis.