How to make CMB maps from huge timelines with small computers

We present in this article two different ways to make CMB maps in practice, from large timelines. One is to make a simple destriping, fitting the data and using the scan intercepts to remove the low frequency noise (stripes). The second, optimal, is to resolve linearly the map-making problem, which in case of big timelines must be simplified and changed from matrices to vectors for the calculations. Assuming few conditions on the noise, it is possible to make fast map-making tools.