You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Compared to "Clearing" data, which has semantics of "this data, which was valid, has now expired" sometimes it is useful to instead log that a component itself represents invalid data.
For example, when an algorithm doesn't converge or outputs a bad value. We know a concrete timestamp when the invalid data was computed, and it's helpful to be able to represent this in the timeline explicitly.
We already have two ways of representing non-data:
[ ] -- the data batch itself is empty
[ Null ] the data itself is unset
We need to think through if one of these can be unambiguously used in the case of "invalid" vs "unset" data, or if we would need to introduce an alternative parallel mechanism.
The text was updated successfully, but these errors were encountered:
Compared to "Clearing" data, which has semantics of "this data, which was valid, has now expired" sometimes it is useful to instead log that a component itself represents invalid data.
For example, when an algorithm doesn't converge or outputs a bad value. We know a concrete timestamp when the invalid data was computed, and it's helpful to be able to represent this in the timeline explicitly.
We already have two ways of representing non-data:
[ ]
-- the data batch itself is empty[ Null ]
the data itself is unsetWe need to think through if one of these can be unambiguously used in the case of "invalid" vs "unset" data, or if we would need to introduce an alternative parallel mechanism.
The text was updated successfully, but these errors were encountered: