Definition
The Network Weight Nw(X) of a space X is the least cardinality of a network for X: a family N of subsets of X such that for every open set U and every point x∈U there is some N∈N with x∈N⊆U (equivalently every open set is the union of members of N contained in it).
Principle
Principle
Network Weight measures how small a family of regions can be while still locally refining every open set; a network need not consist of open sets nor be closed under unions, but must witness local neighborhoods.
Demonstration
Demonstration
In any second-countable space Nw(X)=ℵ0 because a countable base is also a countable network. For the discrete space of cardinal κ the family of singletons is a network so Nw(X)=κ.
Misapplication
Misapplication
Using a family that only intersects each open set without the required inclusion property confuses network with π-network; treating networks as necessarily open or bases leads to overconstraining the invariant.
Consequence
Consequence
A small network weight implies control over continuous images and some covering properties; it often yields separability or Lindelöf-type consequences when combined with other hypotheses.
Reversal
Reversal
Requiring that every open set be union of base elements (not merely witnessed locally) yields the ordinary weight w(X); removing the local inclusion requirement weakens to π-weight, changing the invariant's strength.
Boundary
Boundary
Applies to topological spaces; networks may consist of arbitrary subsets (not necessarily open), so Nw(X) excludes invariants that demand open bases or algebraic structure unless they serve as networks.
Semantic Tension
Semantic Tension
Network Weight sits between Π-Weight and ordinary weight: networks provide stronger local control than π-networks but are weaker than bases. Confusion among these gives different cardinal measures.
Synthesis
Synthesis
Network Weight is the minimal size of a family of subsets witnessing local neighborhoods for every open set; it captures a middle strength of covering information between π-networks and bases.