This documentation-only update corrects a pair of labels in the architecture diagram introduced in v2.5.0a . Deployable components are unchanged.
Food for thought:
Code often grows to include a mix of less-than and greater-than inequalities. Should they be normalized to one direction or the other? (I prefer writing all inequalities in the less-than direction, the smaller quantity on the left.) Is one direction or the other more natural for reading and interpretation? Does it depend on the context (on what we are comparing to what)? Could we reduce the likelihood of errors by writing inequalities in one direction or the other?