This is the Meta issue to track problems the current limitations of floating point calulation causes. Please add related Bug reports as "Depends on" with a short comment here why you want to see needs to improve the particular problem.
Please feel free to reassign:
@Regina: please feel free to leave CC or to add other Bug you found!
Nominate: Bug 32834 - bad result using mdeterm
Nominate: Bug 32862 - Improve precision of function MOD() for big numbers
(7^19 mod 13 is 6 not 7)
Bug 39293 - Big integers do not subtract correctly
do we really want to do this? this would be a huge mess and would decrease calc's performance extremely
and even worse it would force us to write our own implementation for nearly every mathematical function
Setting to WONTFIX as of Markus' comment #3.