This is an old revision of the document!
This is a list of nice to have features or changes in CP2K that nobody got around to do, yet
cp_output_handling.Fsuch that it does not require the input_section. Instead there should be a routine to parse the input section once and store that information into a novel
WRITEstatements were converted into functions calls, one could annotate the output in a standardized way. Such annotations would then allow for parsing the output in a generic fashion.
requiredflag from input keywords and sections. The flag states that there is no default value available, but the same effect can be achieved by simply not supplying a default value. To fully remove the flag one also has to adopt the generation of the input reference manual.
cp_error_typeand the 'pseudo exception handling' code like
IF (.NOT.failure). These constructs clutter the code without providing any useful functionality. The side-feature of the error_type to carry a reference to a
cp_logger_typeis sparsely used. It can be completely replaced by the
workmatrices, the assignment to threads should be static.
arnoldishould be independent from data storage format.
dbcsr_api. There should only be one API.
dbcsr_adddoes not check for symmetry.
dbcsr_adddoes not check the “transpose-state”.
dbcsr_copyhas confusing order of arguments.
dbcsr_tracedoes not work with matrices of different symmetries.
dbcsr_adddoes not work with matrices of different symmetries.