Besides the crystal preparation, the rst and crucial step in the process of protein structure determination is proper processing of the collected diraction images, as they provide the experimental observations used throughout the entire process of structure solution and renement. In the last two decades several computer programs have been developed. Among the most used and popular are: HKL2000, MOSFLM, d*TREK and XDS package. To nd out the advantages and disadvantages of the data processing programs, several very dierent data sets, including diraction data from DNA/RNA and protein crystals were tested. It has been found that all the major programs for processing and analysis of diraction data give excellent and comparable results with good quality, medium resolution data sets, but their treatment of very high resolution or imperfect data diers in terms of indexing, spot integration, scaling and the treatment of errors. If the diraction data are of good quality and the problem is relatively straightforward, the automated approach to data processing may be most appropriate. On the other hand, if one is trying to squeeze out as much information from the experimental data as possible, then only expert manual processing can be successful, regardless of the data quality.