We show how the LIPSOL library (Linear programming Interior-Point SOLvers) can be used in Scilab. To do this, fast Scilab-Fortran interfaces for sparse Cholesky decomposition have been developed. All other calculations are made using Scilab functions, which take into account sparsity for storing and factorizing the matrices. We make use of structural matrices algorithms from Metanet (Scilab toolbox for graphs and networks computations), in particular the Dulmage Mendelsohn algorithm.