The first question is : which types of data should we archive for variable stars?

Let us consider the three following types of data : (i) raw data (output of the telescope, time…); (ii) reduced data (flux, magnitude…); (iii) processed data (light curve…).

We generally need the third one and use the reduced data only when we have trouble with the processed data or when we could obtain better processed data. We never quite go back to the raw data. The situation could evolve with raw data obtained with the CCD frame where the reduction procedure is more complex than with electronic counts. Then, at first sight, we have first to archive the processed data, then make an effort on reduced data and raw data.