This is a purely historical comment. In the 1970's we were faced with the problem that if, as was then the practice, the reflection data were stored on punched cards with one reflection per card, even small molecule structures could be rather heavy to carry around. One of the innovations introduced with SHELX-76 was 'condensed data' (HKLF 1 format) that used a lossy compression to store on average 9 reflections per card (80 columns, restricted character set). Although lossy, the loss in precision was not serious in comparison with the experimental errors; the reflection intensities had usually been estimated by eye anyway. By this means I was able to pack the program (compressed as well, but not lossy!) and five test datasets into one standard box of 2000 cards for distribution purposes. This technique even proved useful in the early days of BITNET, for which the data transfer rates could be rather slow. SHELXL and some of my other programs can still read these 'condensed data'.
George Prof. George M. Sheldrick FRS Dept. Structural Chemistry, University of Goettingen, Tammannstr. 4, D37077 Goettingen, Germany Tel. +49-551-39-3021 or -3068 Fax. +49-551-39-22582
