Tentatives to detect chaos in variable stars are presented. Constraint
s on the accuracy on individual measurements and length of the records
, necessary, to obtain defenitive conclusions, are discussed. Some dir
ections for future improvements are sketched.