Home

Welcome to the fcs wiki!

Known minor issues and limitations - ideas for improvements

1. For integer data, fcs uses $P1B to determine the size (16 or 32 bits) for all parameters. If some formats mix 16 and 32 bit integer data fcs will misread the data. Anybody know of a cytometer that mixes 16 and 32 bit integers to save space? Also, we don't support 64-bit integer data should any exist.
2. For integer data, fcs does not mask off unused bits. For example, if $P1B is 16, but $P1R is 1024, "proper" behavior is to mask off the high bits (i.e. value = value & 0x3FF), because some SW stores stuff in those "spare bits". (bad idea). This masking should probably be optional.
3. fcs only supports standard big-endian, i.e. $BYTEORD = 4,3,2,1 and little-endian $BYTEORD = 1,2,3,4 data. I know that some old PDPs used an unusual 3,4,1,2 format and their data may be unreadable. At the very least the code should be more rigorous and throw an error on these formats.